Welcome to AIOHTTP

Asynchronous HTTP Client/Server for asyncio and Python.

Current version is 3.7.4.

Key Features

Library Installation

$ pip install aiohttp

You may want to install optional cchardet library as faster replacement for chardet:

$ pip install cchardet

For speeding up DNS resolving by client API you may install aiodns as well. This option is highly recommended:

$ pip install aiodns

Installing speedups altogether

The following will get you aiohttp along with chardet, aiodns and brotlipy in one bundle. No need to type separate commands anymore!

$ pip install aiohttp[speedups]

Getting Started

Client example

import aiohttp
import asyncio

async def main():

    async with aiohttp.ClientSession() as session:
        async with session.get('http://python.org') as response:

            print("Status:", response.status)
            print("Content-type:", response.headers['content-type'])

            html = await response.text()
            print("Body:", html[:15], "...")

loop = asyncio.get_event_loop()
loop.run_until_complete(main())

This prints:

Status: 200
Content-type: text/html; charset=utf-8
Body: <!doctype html> ...

Coming from requests ? Read why we need so many lines.

Server example:

from aiohttp import web

async def handle(request):
    name = request.match_info.get('name', "Anonymous")
    text = "Hello, " + name
    return web.Response(text=text)

app = web.Application()
app.add_routes([web.get('/', handle),
                web.get('/{name}', handle)])

if __name__ == '__main__':
    web.run_app(app)

For more information please visit Client and Server pages.

What’s new in aiohttp 3?

Go to What’s new in aiohttp 3.0 page for aiohttp 3.0 major release changes.

Tutorial

Polls tutorial

Source code

The project is hosted on GitHub

Please feel free to file an issue on the bug tracker if you have found a bug or have some suggestion in order to improve the library.

The library uses Azure Pipelines for Continuous Integration.

Dependencies

  • Python 3.6+

  • async_timeout

  • attrs

  • chardet

  • multidict

  • yarl

  • Optional cchardet as faster replacement for chardet.

    Install it explicitly via:

    $ pip install cchardet

  • Optional aiodns for fast DNS resolving. The library is highly recommended.

    $ pip install aiodns

Communication channels

aio-libs discourse group: https://aio-libs.discourse.group

Feel free to post your questions and ideas here.

gitter chat https://gitter.im/aio-libs/Lobby

We support Stack Overflow. Please add aiohttp tag to your question there.

Contributing

Please read the instructions for contributors before making a Pull Request.

Authors and License

The aiohttp package is written mostly by Nikolay Kim and Andrew Svetlov.

It’s Apache 2 licensed and freely available.

Feel free to improve this package and send a pull request to GitHub.

Policy for Backward Incompatible Changes

aiohttp keeps backward compatibility.

After deprecating some Public API (method, class, function argument, etc.) the library guaranties the usage of deprecated API is still allowed at least for a year and half after publishing new release with deprecation.

All deprecations are reflected in documentation and raises DeprecationWarning.

Sometimes we are forced to break the own rule for sake of very strong reason. Most likely the reason is a critical bug which cannot be solved without major API change, but we are working hard for keeping these changes as rare as possible.

Table Of Contents

Client

The page contains all information about aiohttp Client API:

Client Quickstart

Eager to get started? This page gives a good introduction in how to get started with aiohttp client API.

First, make sure that aiohttp is installed and up-to-date

Let’s get started with some simple examples.

Make a Request

Begin by importing the aiohttp module, and asyncio:

import aiohttp
import asyncio

Now, let’s try to get a web-page. For example let’s query http://httpbin.org/get:

async def main():
    async with aiohttp.ClientSession() as session:
        async with session.get('http://httpbin.org/get') as resp:
            print(resp.status)
            print(await resp.text())


loop = asyncio.get_event_loop()
loop.run_until_complete(main())

Now, we have a ClientSession called session and a ClientResponse object called resp. We can get all the information we need from the response. The mandatory parameter of ClientSession.get() coroutine is an HTTP url (str or class:yarl.URL instance).

In order to make an HTTP POST request use ClientSession.post() coroutine:

session.post('http://httpbin.org/post', data=b'data')

Other HTTP methods are available as well:

session.put('http://httpbin.org/put', data=b'data')
session.delete('http://httpbin.org/delete')
session.head('http://httpbin.org/get')
session.options('http://httpbin.org/get')
session.patch('http://httpbin.org/patch', data=b'data')

Note

Don’t create a session per request. Most likely you need a session per application which performs all requests altogether.

More complex cases may require a session per site, e.g. one for Github and other one for Facebook APIs. Anyway making a session for every request is a very bad idea.

A session contains a connection pool inside. Connection reusage and keep-alives (both are on by default) may speed up total performance.

A session context manager usage is not mandatory but await session.close() method should be called in this case, e.g.:

session = aiohttp.ClientSession()
async with session.get('...'):
    # ...
await session.close()
Passing Parameters In URLs

You often want to send some sort of data in the URL’s query string. If you were constructing the URL by hand, this data would be given as key/value pairs in the URL after a question mark, e.g. httpbin.org/get?key=val. Requests allows you to provide these arguments as a dict, using the params keyword argument. As an example, if you wanted to pass key1=value1 and key2=value2 to httpbin.org/get, you would use the following code:

params = {'key1': 'value1', 'key2': 'value2'}
async with session.get('http://httpbin.org/get',
                       params=params) as resp:
    expect = 'http://httpbin.org/get?key1=value1&key2=value2'
    assert str(resp.url) == expect

You can see that the URL has been correctly encoded by printing the URL.

For sending data with multiple values for the same key MultiDict may be used; the library support nested lists ({'key': ['value1', 'value2']}) alternative as well.

It is also possible to pass a list of 2 item tuples as parameters, in that case you can specify multiple values for each key:

params = [('key', 'value1'), ('key', 'value2')]
async with session.get('http://httpbin.org/get',
                       params=params) as r:
    expect = 'http://httpbin.org/get?key=value2&key=value1'
    assert str(r.url) == expect

You can also pass str content as param, but beware – content is not encoded by library. Note that + is not encoded:

async with session.get('http://httpbin.org/get',
                       params='key=value+1') as r:
        assert str(r.url) == 'http://httpbin.org/get?key=value+1'

Note

aiohttp internally performs URL canonicalization before sending request.

Canonicalization encodes host part by IDNA codec and applies requoting to path and query parts.

For example URL('http://example.com/путь/%30?a=%31') is converted to URL('http://example.com/%D0%BF%D1%83%D1%82%D1%8C/0?a=1').

Sometimes canonicalization is not desirable if server accepts exact representation and does not requote URL itself.

To disable canonicalization use encoded=True parameter for URL construction:

await session.get(
    URL('http://example.com/%30', encoded=True))

Warning

Passing params overrides encoded=True, never use both options.

Response Content and Status Code

We can read the content of the server’s response and its status code. Consider the GitHub time-line again:

async with session.get('https://api.github.com/events') as resp:
    print(resp.status)
    print(await resp.text())

prints out something like:

200
'[{"created_at":"2015-06-12T14:06:22Z","public":true,"actor":{...

aiohttp automatically decodes the content from the server. You can specify custom encoding for the text() method:

await resp.text(encoding='windows-1251')
Binary Response Content

You can also access the response body as bytes, for non-text requests:

print(await resp.read())

b'[{"created_at":"2015-06-12T14:06:22Z","public":true,"actor":{...

The gzip and deflate transfer-encodings are automatically decoded for you.

You can enable brotli transfer-encodings support, just install brotlipy.

JSON Request

Any of session’s request methods like request(), ClientSession.get(), ClientSesssion.post() etc. accept json parameter:

async with aiohttp.ClientSession() as session:
    async with session.post(url, json={'test': 'object'})

By default session uses python’s standard json module for serialization. But it is possible to use different serializer. ClientSession accepts json_serialize parameter:

import ujson

async with aiohttp.ClientSession(
        json_serialize=ujson.dumps) as session:
    await session.post(url, json={'test': 'object'})

Note

ujson library is faster than standard json but slightly incompatible.

JSON Response Content

There’s also a built-in JSON decoder, in case you’re dealing with JSON data:

async with session.get('https://api.github.com/events') as resp:
    print(await resp.json())

In case that JSON decoding fails, json() will raise an exception. It is possible to specify custom encoding and decoder functions for the json() call.

Note

The methods above reads the whole response body into memory. If you are planning on reading lots of data, consider using the streaming response method documented below.

Streaming Response Content

While methods read(), json() and text() are very convenient you should use them carefully. All these methods load the whole response in memory. For example if you want to download several gigabyte sized files, these methods will load all the data in memory. Instead you can use the content attribute. It is an instance of the aiohttp.StreamReader class. The gzip and deflate transfer-encodings are automatically decoded for you:

async with session.get('https://api.github.com/events') as resp:
    await resp.content.read(10)

In general, however, you should use a pattern like this to save what is being streamed to a file:

with open(filename, 'wb') as fd:
    while True:
        chunk = await resp.content.read(chunk_size)
        if not chunk:
            break
        fd.write(chunk)

It is not possible to use read(), json() and text() after explicit reading from content.

More complicated POST requests

Typically, you want to send some form-encoded data – much like an HTML form. To do this, simply pass a dictionary to the data argument. Your dictionary of data will automatically be form-encoded when the request is made:

payload = {'key1': 'value1', 'key2': 'value2'}
async with session.post('http://httpbin.org/post',
                        data=payload) as resp:
    print(await resp.text())

{
  ...
  "form": {
    "key2": "value2",
    "key1": "value1"
  },
  ...
}

If you want to send data that is not form-encoded you can do it by passing a bytes instead of a dict. This data will be posted directly and content-type set to ‘application/octet-stream’ by default:

async with session.post(url, data=b'\x00Binary-data\x00') as resp:
    ...

If you want to send JSON data:

async with session.post(url, json={'example': 'test'}) as resp:
    ...

To send text with appropriate content-type just use data argument:

async with session.post(url, data='Тест') as resp:
    ...
POST a Multipart-Encoded File

To upload Multipart-encoded files:

url = 'http://httpbin.org/post'
files = {'file': open('report.xls', 'rb')}

await session.post(url, data=files)

You can set the filename and content_type explicitly:

url = 'http://httpbin.org/post'
data = FormData()
data.add_field('file',
               open('report.xls', 'rb'),
               filename='report.xls',
               content_type='application/vnd.ms-excel')

await session.post(url, data=data)

If you pass a file object as data parameter, aiohttp will stream it to the server automatically. Check StreamReader for supported format information.

See also

Working with Multipart

Streaming uploads

aiohttp supports multiple types of streaming uploads, which allows you to send large files without reading them into memory.

As a simple case, simply provide a file-like object for your body:

with open('massive-body', 'rb') as f:
   await session.post('http://httpbin.org/post', data=f)

Or you can use asynchronous generator:

async def file_sender(file_name=None):
    async with aiofiles.open(file_name, 'rb') as f:
        chunk = await f.read(64*1024)
        while chunk:
            yield chunk
            chunk = await f.read(64*1024)

# Then you can use file_sender as a data provider:

async with session.post('http://httpbin.org/post',
                        data=file_sender(file_name='huge_file')) as resp:
    print(await resp.text())

Because the content attribute is a StreamReader (provides async iterator protocol), you can chain get and post requests together:

resp = await session.get('http://python.org')
await session.post('http://httpbin.org/post',
                   data=resp.content)

Note

Python 3.5 has no native support for asynchronous generators, use async_generator library as workaround.

Deprecated since version 3.1: aiohttp still supports aiohttp.streamer decorator but this approach is deprecated in favor of asynchronous generators as shown above.

WebSockets

aiohttp works with client websockets out-of-the-box.

You have to use the aiohttp.ClientSession.ws_connect() coroutine for client websocket connection. It accepts a url as a first parameter and returns ClientWebSocketResponse, with that object you can communicate with websocket server using response’s methods:

async with session.ws_connect('http://example.org/ws') as ws:
    async for msg in ws:
        if msg.type == aiohttp.WSMsgType.TEXT:
            if msg.data == 'close cmd':
                await ws.close()
                break
            else:
                await ws.send_str(msg.data + '/answer')
        elif msg.type == aiohttp.WSMsgType.ERROR:
            break

You must use the only websocket task for both reading (e.g. await ws.receive() or async for msg in ws:) and writing but may have multiple writer tasks which can only send data asynchronously (by await ws.send_str('data') for example).

Timeouts

Timeout settings are stored in ClientTimeout data structure.

By default aiohttp uses a total 300 seconds (5min) timeout, it means that the whole operation should finish in 5 minutes.

The value could be overridden by timeout parameter for the session (specified in seconds):

timeout = aiohttp.ClientTimeout(total=60)
async with aiohttp.ClientSession(timeout=timeout) as session:
    ...

Timeout could be overridden for a request like ClientSession.get():

async with session.get(url, timeout=timeout) as resp:
    ...

Supported ClientTimeout fields are:

total

The maximal number of seconds for the whole operation including connection establishment, request sending and response reading.

connect

The maximal number of seconds for connection establishment of a new connection or for waiting for a free connection from a pool if pool connection limits are exceeded.

sock_connect

The maximal number of seconds for connecting to a peer for a new connection, not given from a pool.

sock_read

The maximal number of seconds allowed for period between reading a new data portion from a peer.

All fields are floats, None or 0 disables a particular timeout check, see the ClientTimeout reference for defaults and additional details.

Thus the default timeout is:

aiohttp.ClientTimeout(total=5*60, connect=None,
                      sock_connect=None, sock_read=None)

Note

aiohttp ceils timeout if the value is equal or greater than 5 seconds. The timeout expires at the next integer second greater than current_time + timeout.

The ceiling is done for the sake of optimization, when many concurrent tasks are scheduled to wake-up at the almost same but different absolute times. It leads to very many event loop wakeups, which kills performance.

The optimization shifts absolute wakeup times by scheduling them to exactly the same time as other neighbors, the loop wakes up once-per-second for timeout expiration.

Smaller timeouts are not rounded to help testing; in the real life network timeouts usually greater than tens of seconds.

Advanced Client Usage

Client Session

ClientSession is the heart and the main entry point for all client API operations.

Create the session first, use the instance for performing HTTP requests and initiating WebSocket connections.

The session contains a cookie storage and connection pool, thus cookies and connections are shared between HTTP requests sent by the same session.

Custom Request Headers

If you need to add HTTP headers to a request, pass them in a dict to the headers parameter.

For example, if you want to specify the content-type directly:

url = 'http://example.com/image'
payload = b'GIF89a\x01\x00\x01\x00\x00\xff\x00,\x00\x00'
          b'\x00\x00\x01\x00\x01\x00\x00\x02\x00;'
headers = {'content-type': 'image/gif'}

await session.post(url,
                   data=payload,
                   headers=headers)

You also can set default headers for all session requests:

headers={"Authorization": "Basic bG9naW46cGFzcw=="}
async with aiohttp.ClientSession(headers=headers) as session:
    async with session.get("http://httpbin.org/headers") as r:
        json_body = await r.json()
        assert json_body['headers']['Authorization'] == \
            'Basic bG9naW46cGFzcw=='

Typical use case is sending JSON body. You can specify content type directly as shown above, but it is more convenient to use special keyword json:

await session.post(url, json={'example': 'text'})

For text/plain

await session.post(url, data='Привет, Мир!')
Custom Cookies

To send your own cookies to the server, you can use the cookies parameter of ClientSession constructor:

url = 'http://httpbin.org/cookies'
cookies = {'cookies_are': 'working'}
async with ClientSession(cookies=cookies) as session:
    async with session.get(url) as resp:
        assert await resp.json() == {
           "cookies": {"cookies_are": "working"}}

Note

httpbin.org/cookies endpoint returns request cookies in JSON-encoded body. To access session cookies see ClientSession.cookie_jar.

ClientSession may be used for sharing cookies between multiple requests:

async with aiohttp.ClientSession() as session:
    await session.get(
        'http://httpbin.org/cookies/set?my_cookie=my_value')
    filtered = session.cookie_jar.filter_cookies(
        'http://httpbin.org')
    assert filtered['my_cookie'].value == 'my_value'
    async with session.get('http://httpbin.org/cookies') as r:
        json_body = await r.json()
        assert json_body['cookies']['my_cookie'] == 'my_value'
Response Headers and Cookies

We can view the server’s response ClientResponse.headers using a CIMultiDictProxy:

assert resp.headers == {
    'ACCESS-CONTROL-ALLOW-ORIGIN': '*',
    'CONTENT-TYPE': 'application/json',
    'DATE': 'Tue, 15 Jul 2014 16:49:51 GMT',
    'SERVER': 'gunicorn/18.0',
    'CONTENT-LENGTH': '331',
    'CONNECTION': 'keep-alive'}

The dictionary is special, though: it’s made just for HTTP headers. According to RFC 7230, HTTP Header names are case-insensitive. It also supports multiple values for the same key as HTTP protocol does.

So, we can access the headers using any capitalization we want:

assert resp.headers['Content-Type'] == 'application/json'

assert resp.headers.get('content-type') == 'application/json'

All headers are converted from binary data using UTF-8 with surrogateescape option. That works fine on most cases but sometimes unconverted data is needed if a server uses nonstandard encoding. While these headers are malformed from RFC 7230 perspective they may be retrieved by using ClientResponse.raw_headers property:

assert resp.raw_headers == (
    (b'SERVER', b'nginx'),
    (b'DATE', b'Sat, 09 Jan 2016 20:28:40 GMT'),
    (b'CONTENT-TYPE', b'text/html; charset=utf-8'),
    (b'CONTENT-LENGTH', b'12150'),
    (b'CONNECTION', b'keep-alive'))

If a response contains some HTTP Cookies, you can quickly access them:

url = 'http://example.com/some/cookie/setting/url'
async with session.get(url) as resp:
    print(resp.cookies['example_cookie_name'])

Note

Response cookies contain only values, that were in Set-Cookie headers of the last request in redirection chain. To gather cookies between all redirection requests please use aiohttp.ClientSession object.

Redirection History

If a request was redirected, it is possible to view previous responses using the history attribute:

resp = await session.get('http://example.com/some/redirect/')
assert resp.status == 200
assert resp.url = URL('http://example.com/some/other/url/')
assert len(resp.history) == 1
assert resp.history[0].status == 301
assert resp.history[0].url = URL(
    'http://example.com/some/redirect/')

If no redirects occurred or allow_redirects is set to False, history will be an empty sequence.

Uploading pre-compressed data

To upload data that is already compressed before passing it to aiohttp, call the request function with the used compression algorithm name (usually deflate or gzip) as the value of the Content-Encoding header:

async def my_coroutine(session, headers, my_data):
    data = zlib.compress(my_data)
    headers = {'Content-Encoding': 'deflate'}
    async with session.post('http://httpbin.org/post',
                            data=data,
                            headers=headers)
        pass
Disabling content type validation for JSON responses

The standard explicitly restricts JSON Content-Type HTTP header to application/json or any extended form, e.g. application/vnd.custom-type+json. Unfortunately, some servers send a wrong type, like text/html.

This can be worked around in two ways:

  1. Pass the expected type explicitly (in this case checking will be strict, without the extended form support, so custom/xxx+type won’t be accepted):

    await resp.json(content_type='custom/type').

  2. Disable the check entirely:

    await resp.json(content_type=None).

Client Tracing

The execution flow of a specific request can be followed attaching listeners coroutines to the signals provided by the TraceConfig instance, this instance will be used as a parameter for the ClientSession constructor having as a result a client that triggers the different signals supported by the TraceConfig. By default any instance of ClientSession class comes with the signals ability disabled. The following snippet shows how the start and the end signals of a request flow can be followed:

async def on_request_start(
        session, trace_config_ctx, params):
    print("Starting request")

async def on_request_end(session, trace_config_ctx, params):
    print("Ending request")

trace_config = aiohttp.TraceConfig()
trace_config.on_request_start.append(on_request_start)
trace_config.on_request_end.append(on_request_end)
async with aiohttp.ClientSession(
        trace_configs=[trace_config]) as client:
    client.get('http://example.com/some/redirect/')

The trace_configs is a list that can contain instances of TraceConfig class that allow run the signals handlers coming from different TraceConfig instances. The following example shows how two different TraceConfig that have a different nature are installed to perform their job in each signal handle:

from mylib.traceconfig import AuditRequest
from mylib.traceconfig import XRay

async with aiohttp.ClientSession(
        trace_configs=[AuditRequest(), XRay()]) as client:
    client.get('http://example.com/some/redirect/')

All signals take as a parameters first, the ClientSession instance used by the specific request related to that signals and second, a SimpleNamespace instance called trace_config_ctx. The trace_config_ctx object can be used to share the state through to the different signals that belong to the same request and to the same TraceConfig class, perhaps:

async def on_request_start(
        session, trace_config_ctx, params):
    trace_config_ctx.start = asyncio.get_event_loop().time()

async def on_request_end(session, trace_config_ctx, params):
    elapsed = asyncio.get_event_loop().time() - trace_config_ctx.start
    print("Request took {}".format(elapsed))

The trace_config_ctx param is by default a SimpleNampespace that is initialized at the beginning of the request flow. However, the factory used to create this object can be overwritten using the trace_config_ctx_factory constructor param of the TraceConfig class.

The trace_request_ctx param can given at the beginning of the request execution, accepted by all of the HTTP verbs, and will be passed as a keyword argument for the trace_config_ctx_factory factory. This param is useful to pass data that is only available at request time, perhaps:

async def on_request_start(
        session, trace_config_ctx, params):
    print(trace_config_ctx.trace_request_ctx)


session.get('http://example.com/some/redirect/',
            trace_request_ctx={'foo': 'bar'})

See also

Tracing Reference section for more information about the different signals supported.

Connectors

To tweak or change transport layer of requests you can pass a custom connector to ClientSession and family. For example:

conn = aiohttp.TCPConnector()
session = aiohttp.ClientSession(connector=conn)

Note

By default session object takes the ownership of the connector, among other things closing the connections once the session is closed. If you are keen on share the same connector through different session instances you must give the connector_owner parameter as False for each session instance.

See also

Connectors section for more information about different connector types and configuration options.

Limiting connection pool size

To limit amount of simultaneously opened connections you can pass limit parameter to connector:

conn = aiohttp.TCPConnector(limit=30)

The example limits total amount of parallel connections to 30.

The default is 100.

If you explicitly want not to have limits, pass 0. For example:

conn = aiohttp.TCPConnector(limit=0)

To limit amount of simultaneously opened connection to the same endpoint ((host, port, is_ssl) triple) you can pass limit_per_host parameter to connector:

conn = aiohttp.TCPConnector(limit_per_host=30)

The example limits amount of parallel connections to the same to 30.

The default is 0 (no limit on per host bases).

Tuning the DNS cache

By default TCPConnector comes with the DNS cache table enabled, and resolutions will be cached by default for 10 seconds. This behavior can be changed either to change of the TTL for a resolution, as can be seen in the following example:

conn = aiohttp.TCPConnector(ttl_dns_cache=300)

or disabling the use of the DNS cache table, meaning that all requests will end up making a DNS resolution, as the following example shows:

conn = aiohttp.TCPConnector(use_dns_cache=False)
Resolving using custom nameservers

In order to specify the nameservers to when resolving the hostnames, aiodns is required:

from aiohttp.resolver import AsyncResolver

resolver = AsyncResolver(nameservers=["8.8.8.8", "8.8.4.4"])
conn = aiohttp.TCPConnector(resolver=resolver)
Unix domain sockets

If your HTTP server uses UNIX domain sockets you can use UnixConnector:

conn = aiohttp.UnixConnector(path='/path/to/socket')
session = aiohttp.ClientSession(connector=conn)
Named pipes in Windows

If your HTTP server uses Named pipes you can use NamedPipeConnector:

conn = aiohttp.NamedPipeConnector(path=r'\\.\pipe\<name-of-pipe>')
session = aiohttp.ClientSession(connector=conn)

It will only work with the ProactorEventLoop

SSL control for TCP sockets

By default aiohttp uses strict checks for HTTPS protocol. Certification checks can be relaxed by setting ssl to False:

r = await session.get('https://example.com', ssl=False)

If you need to setup custom ssl parameters (use own certification files for example) you can create a ssl.SSLContext instance and pass it into the proper ClientSession method:

sslcontext = ssl.create_default_context(
   cafile='/path/to/ca-bundle.crt')
r = await session.get('https://example.com', ssl=sslcontext)

If you need to verify self-signed certificates, you can do the same thing as the previous example, but add another call to ssl.SSLContext.load_cert_chain() with the key pair:

sslcontext = ssl.create_default_context(
   cafile='/path/to/ca-bundle.crt')
sslcontext.load_cert_chain('/path/to/client/public/device.pem',
                           '/path/to/client/private/device.key')
r = await session.get('https://example.com', ssl=sslcontext)

There is explicit errors when ssl verification fails

aiohttp.ClientConnectorSSLError:

try:
    await session.get('https://expired.badssl.com/')
except aiohttp.ClientConnectorSSLError as e:
    assert isinstance(e, ssl.SSLError)

aiohttp.ClientConnectorCertificateError:

try:
    await session.get('https://wrong.host.badssl.com/')
except aiohttp.ClientConnectorCertificateError as e:
    assert isinstance(e, ssl.CertificateError)

If you need to skip both ssl related errors

aiohttp.ClientSSLError:

try:
    await session.get('https://expired.badssl.com/')
except aiohttp.ClientSSLError as e:
    assert isinstance(e, ssl.SSLError)

try:
    await session.get('https://wrong.host.badssl.com/')
except aiohttp.ClientSSLError as e:
    assert isinstance(e, ssl.CertificateError)

You may also verify certificates via SHA256 fingerprint:

# Attempt to connect to https://www.python.org
# with a pin to a bogus certificate:
bad_fp = b'0'*64
exc = None
try:
    r = await session.get('https://www.python.org',
                          ssl=aiohttp.Fingerprint(bad_fp))
except aiohttp.FingerprintMismatch as e:
    exc = e
assert exc is not None
assert exc.expected == bad_fp

# www.python.org cert's actual fingerprint
assert exc.got == b'...'

Note that this is the fingerprint of the DER-encoded certificate. If you have the certificate in PEM format, you can convert it to DER with e.g:

openssl x509 -in crt.pem -inform PEM -outform DER > crt.der

Note

Tip: to convert from a hexadecimal digest to a binary byte-string, you can use binascii.unhexlify().

ssl parameter could be passed to TCPConnector as default, the value from ClientSession.get() and others override default.

Proxy support

aiohttp supports plain HTTP proxies and HTTP proxies that can be upgraded to HTTPS via the HTTP CONNECT method. aiohttp does not support proxies that must be connected to via https://. To connect, use the proxy parameter:

async with aiohttp.ClientSession() as session:
    async with session.get("http://python.org",
                           proxy="http://proxy.com") as resp:
        print(resp.status)

It also supports proxy authorization:

async with aiohttp.ClientSession() as session:
    proxy_auth = aiohttp.BasicAuth('user', 'pass')
    async with session.get("http://python.org",
                           proxy="http://proxy.com",
                           proxy_auth=proxy_auth) as resp:
        print(resp.status)

Authentication credentials can be passed in proxy URL:

session.get("http://python.org",
            proxy="http://user:pass@some.proxy.com")

Contrary to the requests library, it won’t read environment variables by default. But you can do so by passing trust_env=True into aiohttp.ClientSession constructor for extracting proxy configuration from HTTP_PROXY or HTTPS_PROXY environment variables (both are case insensitive):

async with aiohttp.ClientSession(trust_env=True) as session:
    async with session.get("http://python.org") as resp:
        print(resp.status)

Proxy credentials are given from ~/.netrc file if present (see aiohttp.ClientSession for more details).

Graceful Shutdown

When ClientSession closes at the end of an async with block (or through a direct ClientSession.close() call), the underlying connection remains open due to asyncio internal details. In practice, the underlying connection will close after a short while. However, if the event loop is stopped before the underlying connection is closed, a ResourceWarning: unclosed transport warning is emitted (when warnings are enabled).

To avoid this situation, a small delay must be added before closing the event loop to allow any open underlying connections to close.

For a ClientSession without SSL, a simple zero-sleep (await asyncio.sleep(0)) will suffice:

async def read_website():
    async with aiohttp.ClientSession() as session:
        async with session.get('http://example.org/') as resp:
            await resp.read()

loop = asyncio.get_event_loop()
loop.run_until_complete(read_website())
# Zero-sleep to allow underlying connections to close
loop.run_until_complete(asyncio.sleep(0))
loop.close()

For a ClientSession with SSL, the application must wait a short duration before closing:

...
# Wait 250 ms for the underlying SSL connections to close
loop.run_until_complete(asyncio.sleep(0.250))
loop.close()

Note that the appropriate amount of time to wait will vary from application to application.

All if this will eventually become obsolete when the asyncio internals are changed so that aiohttp itself can wait on the underlying connection to close. Please follow issue #1925 for the progress on this.

Client Reference

Client Session

Client session is the recommended interface for making HTTP requests.

Session encapsulates a connection pool (connector instance) and supports keepalives by default. Unless you are connecting to a large, unknown number of different servers over the lifetime of your application, it is suggested you use a single session for the lifetime of your application to benefit from connection pooling.

Usage example:

import aiohttp
import asyncio

async def fetch(client):
    async with client.get('http://python.org') as resp:
        assert resp.status == 200
        return await resp.text()

async def main():
    async with aiohttp.ClientSession() as client:
        html = await fetch(client)
        print(html)

loop = asyncio.get_event_loop()
loop.run_until_complete(main())

The client session supports the context manager protocol for self closing.

class aiohttp.ClientSession(*, connector=None, loop=None, cookies=None, headers=None, skip_auto_headers=None, auth=None, json_serialize=json.dumps, version=aiohttp.HttpVersion11, cookie_jar=None, read_timeout=None, conn_timeout=None, timeout=sentinel, raise_for_status=False, connector_owner=True, auto_decompress=True, read_bufsize=2 ** 16, requote_redirect_url=False, trust_env=False, trace_configs=None)

The class for creating client sessions and making requests.

Parameters

  • connector (aiohttp.BaseConnector) – BaseConnector sub-class instance to support connection pooling.

  • loop –

    event loop used for processing HTTP requests.

    If loop is None the constructor borrows it from connector if specified.

    asyncio.get_event_loop() is used for getting default event loop otherwise.

    Deprecated since version 2.0.

  • cookies (dict) – Cookies to send with the request (optional)

  • headers –

    HTTP Headers to send with every request (optional).

    May be either iterable of key-value pairs or Mapping (e.g. dict, CIMultiDict).

  • skip_auto_headers –

    set of headers for which autogeneration should be skipped.

    aiohttp autogenerates headers like User-Agent or Content-Type if these headers are not explicitly passed. Using skip_auto_headers parameter allows to skip that generation. Note that Content-Length autogeneration can’t be skipped.

    Iterable of str or istr (optional)

  • auth (aiohttp.BasicAuth) – an object that represents HTTP Basic Authorization (optional)

  • version – supported HTTP version, HTTP 1.1 by default.

  • cookie_jar –

    Cookie Jar, AbstractCookieJar instance.

    By default every session instance has own private cookie jar for automatic cookies processing but user may redefine this behavior by providing own jar implementation.

    One example is not processing cookies at all when working in proxy mode.

    If no cookie processing is needed, a aiohttp.DummyCookieJar instance can be provided.

  • json_serialize (callable) –

    Json serializer callable.

    By default json.dumps() function.

  • raise_for_status (bool) –

    Automatically call ClientResponse.raise_for_status() for each response, False by default.

    This parameter can be overridden when you making a request, e.g.:

    client_session = aiohttp.ClientSession(raise_for_status=True)
resp = await client_session.get(url, raise_for_status=False)
async with resp:
    assert resp.status == 200

    Set the parameter to True if you need raise_for_status for most of cases but override raise_for_status for those requests where you need to handle responses with status 400 or higher.

  • timeout –

    a ClientTimeout settings structure, 300 seconds (5min)

    total timeout by default.

    New in version 3.3.

  • read_timeout (float) –

    Request operations timeout. read_timeout is cumulative for all request operations (request, redirects, responses, data consuming). By default, the read timeout is 5*60 seconds. Use None or 0 to disable timeout checks.

    Deprecated since version 3.3: Use timeout parameter instead.

  • conn_timeout (float) –

    timeout for connection establishing (optional). Values 0 or None mean no timeout.

    Deprecated since version 3.3: Use timeout parameter instead.

  • connector_owner (bool) –

    Close connector instance on session closing.

    Setting the parameter to False allows to share connection pool between sessions without sharing session state: cookies etc.

  • auto_decompress (bool) –

    Automatically decompress response body,

    True by default

    New in version 2.3.

  • read_bufsize (int) –

    Size of the read buffer (ClientResponse.content).

    64 KiB by default.

    New in version 3.7.

  • trust_env (bool) –

    Get proxies information from HTTP_PROXY / HTTPS_PROXY environment variables if the parameter is True (False by default).

    Get proxy credentials from ~/.netrc file if present.

    See also

    .netrc documentation: https://www.gnu.org/software/inetutils/manual/html_node/The-_002enetrc-file.html

    New in version 2.3.

    Changed in version 3.0: Added support for ~/.netrc file.

  • requote_redirect_url (bool) –

    Apply URL requoting for redirection URLs if

    automatic redirection is enabled (True by default).

    New in version 3.5.

  • trace_configs – A list of TraceConfig instances used for client tracing. None (default) is used for request tracing disabling. See Tracing Reference for more information.

closed

True if the session has been closed, False otherwise.

A read-only property.

connector

aiohttp.BaseConnector derived instance used for the session.

A read-only property.

cookie_jar

The session cookies, AbstractCookieJar instance.

Gives access to cookie jar’s content and modifiers.

A read-only property.

requote_redirect_url

aiohttp re quote’s redirect urls by default, but some servers require exact url from location header. To disable re-quote system set requote_redirect_url attribute to False.

New in version 2.1.

Note

This parameter affects all subsequent requests.

Deprecated since version 3.5: The attribute modification is deprecated.

loop

A loop instance used for session creation.

A read-only property.

Deprecated since version 3.5.

timeout

Default client timeouts, ClientTimeout instance. The value can be tuned by passing timeout parameter to ClientSession constructor.

New in version 3.7.

headers

HTTP Headers that sent with every request

May be either iterable of key-value pairs or Mapping (e.g. dict, CIMultiDict).

New in version 3.7.

skip_auto_headers

Set of headers for which autogeneration skipped.

frozenset of str or istr (optional)

New in version 3.7.

auth

An object that represents HTTP Basic Authorization.

BasicAuth (optional)

New in version 3.7.

json_serialize

Json serializer callable.

By default json.dumps() function.

New in version 3.7.

connector_owner

Should connector be closed on session closing

bool (optional)

New in version 3.7.

raise_for_status

Should ClientResponse.raise_for_status() be called for each response

Either bool or callable

New in version 3.7.

auto_decompress

Should the body response be automatically decompressed

bool default is True

New in version 3.7.

trust_env

Should get proxies information from HTTP_PROXY / HTTPS_PROXY environment variables or ~/.netrc file if present

bool default is False

New in version 3.7.

trace_config

A list of TraceConfig instances used for client tracing. None (default) is used for request tracing disabling. See Tracing Reference for more information.

New in version 3.7.

coroutine async-with request(method, url, *, params=None, data=None, json=None, cookies=None, headers=None, skip_auto_headers=None, auth=None, allow_redirects=True, max_redirects=10, compress=None, chunked=None, expect100=False, raise_for_status=None, read_until_eof=True, read_bufsize=None, proxy=None, proxy_auth=None, timeout=sentinel, ssl=None, verify_ssl=None, fingerprint=None, ssl_context=None, proxy_headers=None)

Performs an asynchronous HTTP request. Returns a response object.

Parameters

  • method (str) – HTTP method

  • url – Request URL, str or URL.

  • params –

    Mapping, iterable of tuple of key/value pairs or string to be sent as parameters in the query string of the new request. Ignored for subsequent redirected requests (optional)

    Allowed values are:

  • data – The data to send in the body of the request. This can be a FormData object or anything that can be passed into FormData, e.g. a dictionary, bytes, or file-like object. (optional)

  • json – Any json compatible python object (optional). json and data parameters could not be used at the same time.

  • cookies (dict) –

    HTTP Cookies to send with

    the request (optional)

    Global session cookies and the explicitly set cookies will be merged when sending the request.

    New in version 3.5.

  • headers (dict) – HTTP Headers to send with the request (optional)

  • skip_auto_headers –

    set of headers for which autogeneration should be skipped.

    aiohttp autogenerates headers like User-Agent or Content-Type if these headers are not explicitly passed. Using skip_auto_headers parameter allows to skip that generation.

    Iterable of str or istr (optional)

  • auth (aiohttp.BasicAuth) – an object that represents HTTP Basic Authorization (optional)

  • allow_redirects (bool) – If set to False, do not follow redirects. True by default (optional).

  • max_redirects (int) – Maximum number of redirects to follow. 10 by default.

  • compress (bool) – Set to True if request has to be compressed with deflate encoding. If compress can not be combined with a Content-Encoding and Content-Length headers. None by default (optional).

  • chunked (int) – Enable chunked transfer encoding. It is up to the developer to decide how to chunk data streams. If chunking is enabled, aiohttp encodes the provided chunks in the “Transfer-encoding: chunked” format. If chunked is set, then the Transfer-encoding and content-length headers are disallowed. None by default (optional).

  • expect100 (bool) – Expect 100-continue response from server. False by default (optional).

  • raise_for_status (bool) –

    Automatically call ClientResponse.raise_for_status() for

    response if set to True. If set to None value from ClientSession will be used. None by default (optional).

    New in version 3.4.

  • read_until_eof (bool) – Read response until EOF if response does not have Content-Length header. True by default (optional).

  • read_bufsize (int) –

    Size of the read buffer (ClientResponse.content).

    None by default, it means that the session global value is used.

    New in version 3.7.

  • proxy – Proxy URL, str or URL (optional)

  • proxy_auth (aiohttp.BasicAuth) – an object that represents proxy HTTP Basic Authorization (optional)

  • timeout (int) –

    override the session’s timeout.

    Changed in version 3.3: The parameter is ClientTimeout instance, float is still supported for sake of backward compatibility.

    If float is passed it is a total timeout (in seconds).

  • ssl –

    SSL validation mode. None for default SSL check

    (ssl.create_default_context() is used), False for skip SSL certificate validation, aiohttp.Fingerprint for fingerprint validation, ssl.SSLContext for custom SSL certificate validation.

    Supersedes verify_ssl, ssl_context and fingerprint parameters.

    New in version 3.0.

  • verify_ssl (bool) –

    Perform SSL certificate validation for HTTPS requests (enabled by default). May be disabled to skip validation for sites with invalid certificates.

    New in version 2.3.

    Deprecated since version 3.0: Use ssl=False

  • fingerprint (bytes) –

    Pass the SHA256 digest of the expected certificate in DER format to verify that the certificate the server presents matches. Useful for certificate pinning.

    Warning: use of MD5 or SHA1 digests is insecure and removed.

    New in version 2.3.

    Deprecated since version 3.0: Use ssl=aiohttp.Fingerprint(digest)

  • ssl_context (ssl.SSLContext) –

    ssl context used for processing HTTPS requests (optional).

    ssl_context may be used for configuring certification authority channel, supported SSL options etc.

    New in version 2.3.

    Deprecated since version 3.0: Use ssl=ssl_context

  • proxy_headers (abc.Mapping) –

    HTTP headers to send to the proxy if the parameter proxy has been provided.

    New in version 2.3.

  • trace_request_ctx –

    Object used to give as a kw param for each new TraceConfig object instantiated, used to give information to the tracers that is only available at request time.

    New in version 3.0.

Return ClientResponse

a client response object.

coroutine async-with get(url, *, allow_redirects=True, **kwargs)

Perform a GET request.

In order to modify inner request parameters, provide kwargs.

Parameters

  • url – Request URL, str or URL

  • allow_redirects (bool) – If set to False, do not follow redirects. True by default (optional).

Return ClientResponse

a client response object.

coroutine async-with post(url, *, data=None, **kwargs)

Perform a POST request.

In order to modify inner request parameters, provide kwargs.

Parameters

  • url – Request URL, str or URL

  • data – Data to send in the body of the request; see request for details (optional)

Return ClientResponse

a client response object.

coroutine async-with put(url, *, data=None, **kwargs)

Perform a PUT request.

In order to modify inner request parameters, provide kwargs.

Parameters

  • url – Request URL, str or URL

  • data – Data to send in the body of the request; see request for details (optional)

Return ClientResponse

a client response object.

coroutine async-with delete(url, **kwargs)

Perform a DELETE request.

In order to modify inner request parameters, provide kwargs.

Parameters

url – Request URL, str or URL

Return ClientResponse

a client response object.

coroutine async-with head(url, *, allow_redirects=False, **kwargs)

Perform a HEAD request.

In order to modify inner request parameters, provide kwargs.

Parameters

  • url – Request URL, str or URL

  • allow_redirects (bool) – If set to False, do not follow redirects. False by default (optional).

Return ClientResponse

a client response object.

coroutine async-with options(url, *, allow_redirects=True, **kwargs)

Perform an OPTIONS request.

In order to modify inner request parameters, provide kwargs.

Parameters

  • url – Request URL, str or URL

  • allow_redirects (bool) – If set to False, do not follow redirects. True by default (optional).

Return ClientResponse

a client response object.

coroutine async-with patch(url, *, data=None, **kwargs)

Perform a PATCH request.

In order to modify inner request parameters, provide kwargs.

Parameters

  • url – Request URL, str or URL

  • data – Data to send in the body of the request; see request for details (optional)

Return ClientResponse

a client response object.

coroutine async-with ws_connect(url, *, method='GET', protocols=(), timeout=10.0, receive_timeout=None, auth=None, autoclose=True, autoping=True, heartbeat=None, origin=None, headers=None, proxy=None, proxy_auth=None, ssl=None, verify_ssl=None, fingerprint=None, ssl_context=None, proxy_headers=None, compress=0, max_msg_size=4194304)

Create a websocket connection. Returns a ClientWebSocketResponse object.

Parameters

  • url – Websocket server url, str or URL

  • protocols (tuple) – Websocket protocols

  • timeout (float) – Timeout for websocket to close. 10 seconds by default

  • receive_timeout (float) – Timeout for websocket to receive complete message. None (unlimited) seconds by default

  • auth (aiohttp.BasicAuth) – an object that represents HTTP Basic Authorization (optional)

  • autoclose (bool) – Automatically close websocket connection on close message from server. If autoclose is False then close procedure has to be handled manually. True by default

  • autoping (bool) – automatically send pong on ping message from server. True by default

  • heartbeat (float) – Send ping message every heartbeat seconds and wait pong response, if pong response is not received then close connection. The timer is reset on any data reception.(optional)

  • origin (str) – Origin header to send to server(optional)

  • headers (dict) – HTTP Headers to send with the request (optional)

  • proxy (str) – Proxy URL, str or URL (optional)

  • proxy_auth (aiohttp.BasicAuth) – an object that represents proxy HTTP Basic Authorization (optional)

  • ssl –

    SSL validation mode. None for default SSL check

    (ssl.create_default_context() is used), False for skip SSL certificate validation, aiohttp.Fingerprint for fingerprint validation, ssl.SSLContext for custom SSL certificate validation.

    Supersedes verify_ssl, ssl_context and fingerprint parameters.

    New in version 3.0.

  • verify_ssl (bool) –

    Perform SSL certificate validation for HTTPS requests (enabled by default). May be disabled to skip validation for sites with invalid certificates.

    New in version 2.3.

    Deprecated since version 3.0: Use ssl=False

  • fingerprint (bytes) –

    Pass the SHA256 digest of the expected certificate in DER format to verify that the certificate the server presents matches. Useful for certificate pinning.

    Note: use of MD5 or SHA1 digests is insecure and deprecated.

    New in version 2.3.

    Deprecated since version 3.0: Use ssl=aiohttp.Fingerprint(digest)

  • ssl_context (ssl.SSLContext) –

    ssl context used for processing HTTPS requests (optional).

    ssl_context may be used for configuring certification authority channel, supported SSL options etc.

    New in version 2.3.

    Deprecated since version 3.0: Use ssl=ssl_context

  • proxy_headers (dict) –

    HTTP headers to send to the proxy if the parameter proxy has been provided.

    New in version 2.3.

  • compress (int) –

    Enable Per-Message Compress Extension support.

    0 for disable, 9 to 15 for window bit support. Default value is 0.

    New in version 2.3.

  • max_msg_size (int) –

    maximum size of read websocket message,

    4 MB by default. To disable the size limit use 0.

    New in version 3.3.

  • method (str) –

    HTTP method to establish WebSocket connection,

    'GET' by default.

    New in version 3.5.

coroutine close()

Close underlying connector.

Release all acquired resources.

detach()

Detach connector from session without closing the former.

Session is switched to closed state anyway.

Basic API

While we encourage ClientSession usage we also provide simple coroutines for making HTTP requests.

Basic API is good for performing simple HTTP requests without keepaliving, cookies and complex connection stuff like properly configured SSL certification chaining.

async-with aiohttp.request(method, url, *, params=None, data=None, json=None, headers=None, cookies=None, auth=None, allow_redirects=True, max_redirects=10, encoding='utf-8', version=HttpVersion(major=1, minor=1), compress=None, chunked=None, expect100=False, raise_for_status=False, read_bufsize=None, connector=None, loop=None, read_until_eof=True, timeout=sentinel)

Asynchronous context manager for performing an asynchronous HTTP request. Returns a ClientResponse response object.

Parameters

  • method (str) – HTTP method

  • url – Requested URL, str or URL

  • params (dict) – Parameters to be sent in the query string of the new request (optional)

  • data – The data to send in the body of the request. This can be a FormData object or anything that can be passed into FormData, e.g. a dictionary, bytes, or file-like object. (optional)

  • json – Any json compatible python object (optional). json and data parameters could not be used at the same time.

  • headers (dict) – HTTP Headers to send with the request (optional)

  • cookies (dict) – Cookies to send with the request (optional)

  • auth (aiohttp.BasicAuth) – an object that represents HTTP Basic Authorization (optional)

  • allow_redirects (bool) – If set to False, do not follow redirects. True by default (optional).

  • version (aiohttp.protocol.HttpVersion) – Request HTTP version (optional)

  • compress (bool) – Set to True if request has to be compressed with deflate encoding. False instructs aiohttp to not compress data. None by default (optional).

  • chunked (int) – Enables chunked transfer encoding. None by default (optional).

  • expect100 (bool) – Expect 100-continue response from server. False by default (optional).

  • raise_for_status (bool) –

    Automatically call

    ClientResponse.raise_for_status() for response if set to True. If set to None value from ClientSession will be used. None by default (optional).

    New in version 3.4.

  • connector (aiohttp.BaseConnector) – BaseConnector sub-class instance to support connection pooling.

  • read_until_eof (bool) – Read response until EOF if response does not have Content-Length header. True by default (optional).

  • read_bufsize (int) –

    Size of the read buffer (ClientResponse.content).

    None by default, it means that the session global value is used.

    New in version 3.7.

  • timeout – a ClientTimeout settings structure, 300 seconds (5min) total timeout by default.

  • loop –

    event loop

    used for processing HTTP requests. If param is None, asyncio.get_event_loop() is used for getting default event loop.

    Deprecated since version 2.0.

Return ClientResponse

a client response object.

Usage:

import aiohttp

async def fetch():
    async with aiohttp.request('GET',
            'http://python.org/') as resp:
        assert resp.status == 200
        print(await resp.text())
Connectors

Connectors are transports for aiohttp client API.

There are standard connectors:

  1. TCPConnector for regular TCP sockets (both HTTP and HTTPS schemes supported).

  2. UnixConnector for connecting via UNIX socket (it’s used mostly for testing purposes).

All connector classes should be derived from BaseConnector.

By default all connectors support keep-alive connections (behavior is controlled by force_close constructor’s parameter).

BaseConnector
class aiohttp.BaseConnector(*, keepalive_timeout=15, force_close=False, limit=100, limit_per_host=0, enable_cleanup_closed=False, loop=None)

Base class for all connectors.

Parameters

  • keepalive_timeout (float) – timeout for connection reusing after releasing (optional). Values 0. For disabling keep-alive feature use force_close=True flag.

  • limit (int) – total number simultaneous connections. If limit is None the connector has no limit (default: 100).

  • limit_per_host (int) – limit simultaneous connections to the same endpoint. Endpoints are the same if they are have equal (host, port, is_ssl) triple. If limit is 0 the connector has no limit (default: 0).

  • force_close (bool) – close underlying sockets after connection releasing (optional).

  • enable_cleanup_closed (bool) – some SSL servers do not properly complete SSL shutdown process, in that case asyncio leaks ssl connections. If this parameter is set to True, aiohttp additionally aborts underlining transport after 2 seconds. It is off by default.

  • loop –

    event loop used for handling connections. If param is None, asyncio.get_event_loop() is used for getting default event loop.

    Deprecated since version 2.0.

closed

Read-only property, True if connector is closed.

force_close

Read-only property, True if connector should ultimately close connections on releasing.

limit

The total number for simultaneous connections. If limit is 0 the connector has no limit. The default limit size is 100.

limit_per_host

The limit for simultaneous connections to the same endpoint.

Endpoints are the same if they are have equal (host, port, is_ssl) triple.

If limit_per_host is None the connector has no limit per host.

Read-only property.

coroutine close()

Close all opened connections.

coroutine connect(request)

Get a free connection from pool or create new one if connection is absent in the pool.

The call may be paused if limit is exhausted until used connections returns to pool.

Parameters

request (aiohttp.ClientRequest) – request object which is connection initiator.

Returns

Connection object.

coroutine _create_connection(req)

Abstract method for actual connection establishing, should be overridden in subclasses.

TCPConnector
class aiohttp.TCPConnector(*, ssl=None, verify_ssl=True, fingerprint=None, use_dns_cache=True, ttl_dns_cache=10, family=0, ssl_context=None, local_addr=None, resolver=None, keepalive_timeout=sentinel, force_close=False, limit=100, limit_per_host=0, enable_cleanup_closed=False, loop=None)

Connector for working with HTTP and HTTPS via TCP sockets.

The most common transport. When you don’t know what connector type to use, use a TCPConnector instance.

TCPConnector inherits from BaseConnector.

Constructor accepts all parameters suitable for BaseConnector plus several TCP-specific ones:

param ssl
SSL validation mode. None for default SSL check

(ssl.create_default_context() is used), False for skip SSL certificate validation, aiohttp.Fingerprint for fingerprint validation, ssl.SSLContext for custom SSL certificate validation.

Supersedes verify_ssl, ssl_context and fingerprint parameters.

New in version 3.0.

Parameters

  • verify_ssl (bool) –

    perform SSL certificate validation for HTTPS requests (enabled by default). May be disabled to skip validation for sites with invalid certificates.

    Deprecated since version 2.3: Pass verify_ssl to ClientSession.get() etc.

  • fingerprint (bytes) –

    pass the SHA256 digest of the expected certificate in DER format to verify that the certificate the server presents matches. Useful for certificate pinning.

    Note: use of MD5 or SHA1 digests is insecure and deprecated.

    Deprecated since version 2.3: Pass verify_ssl to ClientSession.get() etc.

  • use_dns_cache (bool) –

    use internal cache for DNS lookups, True by default.

    Enabling an option may speedup connection establishing a bit but may introduce some side effects also.

  • ttl_dns_cache (int) –

    expire after some seconds the DNS entries, None means cached forever. By default 10 seconds (optional).

    In some environments the IP addresses related to a specific HOST can change after a specific time. Use this option to keep the DNS cache updated refreshing each entry after N seconds.

  • limit (int) – total number simultaneous connections. If limit is None the connector has no limit (default: 100).

  • limit_per_host (int) – limit simultaneous connections to the same endpoint. Endpoints are the same if they are have equal (host, port, is_ssl) triple. If limit is 0 the connector has no limit (default: 0).

  • resolver (aiohttp.abc.AbstractResolver) –

    custom resolver instance to use. aiohttp.DefaultResolver by default (asynchronous if aiodns>=1.1 is installed).

    Custom resolvers allow to resolve hostnames differently than the way the host is configured.

    The resolver is aiohttp.ThreadedResolver by default, asynchronous version is pretty robust but might fail in very rare cases.

  • family (int) –

    TCP socket family, both IPv4 and IPv6 by default. For IPv4 only use socket.AF_INET, for IPv6 only – socket.AF_INET6.

    family is 0 by default, that means both IPv4 and IPv6 are accepted. To specify only concrete version please pass socket.AF_INET or socket.AF_INET6 explicitly.

  • ssl_context (ssl.SSLContext) –

    SSL context used for processing HTTPS requests (optional).

    ssl_context may be used for configuring certification authority channel, supported SSL options etc.

  • local_addr (tuple) – tuple of (local_host, local_port) used to bind socket locally if specified.

  • force_close (bool) – close underlying sockets after connection releasing (optional).

  • enable_cleanup_closed (bool) – Some ssl servers do not properly complete SSL shutdown process, in that case asyncio leaks SSL connections. If this parameter is set to True, aiohttp additionally aborts underlining transport after 2 seconds. It is off by default.

family

TCP socket family e.g. socket.AF_INET or socket.AF_INET6

Read-only property.

dns_cache

Use quick lookup in internal DNS cache for host names if True.

Read-only bool property.

cached_hosts

The cache of resolved hosts if dns_cache is enabled.

Read-only types.MappingProxyType property.

clear_dns_cache(self, host=None, port=None)

Clear internal DNS cache.

Remove specific entry if both host and port are specified, clear all cache otherwise.

UnixConnector
class aiohttp.UnixConnector(path, *, conn_timeout=None, keepalive_timeout=30, limit=100, force_close=False, loop=None)

Unix socket connector.

Use UnixConnector for sending HTTP/HTTPS requests through UNIX Sockets as underlying transport.

UNIX sockets are handy for writing tests and making very fast connections between processes on the same host.

UnixConnector is inherited from BaseConnector.

Usage:

conn = UnixConnector(path='/path/to/socket')
session = ClientSession(connector=conn)
async with session.get('http://python.org') as resp:
    ...

Constructor accepts all parameters suitable for BaseConnector plus UNIX-specific one:

Parameters

path (str) – Unix socket path

path

Path to UNIX socket, read-only str property.

Connection
class aiohttp.Connection

Encapsulates single connection in connector object.

End user should never create Connection instances manually but get it by BaseConnector.connect() coroutine.

closed

bool read-only property, True if connection was closed, released or detached.

loop

Event loop used for connection

Deprecated since version 3.5.

transport

Connection transport

close()

Close connection with forcibly closing underlying socket.

release()

Release connection back to connector.

Underlying socket is not closed, the connection may be reused later if timeout (30 seconds by default) for connection was not expired.

Response object
class aiohttp.ClientResponse

Client response returned by ClientSession.request() and family.

User never creates the instance of ClientResponse class but gets it from API calls.

ClientResponse supports async context manager protocol, e.g.:

resp = await client_session.get(url)
async with resp:
    assert resp.status == 200

After exiting from async with block response object will be released (see release() coroutine).

version

Response’s version, HttpVersion instance.

status

HTTP status code of response (int), e.g. 200.

reason

HTTP status reason of response (str), e.g. "OK".

ok

Boolean representation of HTTP status code (bool). True if status is less than 400; otherwise, False.

method

Request’s method (str).

url

URL of request (URL).

real_url

Unmodified URL of request with URL fragment unstripped (URL).

New in version 3.2.

connection

Connection used for handling response.

content

Payload stream, which contains response’s BODY (StreamReader). It supports various reading methods depending on the expected format. When chunked transfer encoding is used by the server, allows retrieving the actual http chunks.

Reading from the stream may raise aiohttp.ClientPayloadError if the response object is closed before response receives all data or in case if any transfer encoding related errors like misformed chunked encoding of broken compression data.

cookies

HTTP cookies of response (Set-Cookie HTTP header, SimpleCookie).

headers

A case-insensitive multidict proxy with HTTP headers of response, CIMultiDictProxy.

raw_headers

Unmodified HTTP headers of response as unconverted bytes, a sequence of (key, value) pairs.

Link HTTP header parsed into a MultiDictProxy.

For each link, key is link param rel when it exists, or link url as str otherwise, and value is MultiDictProxy of link params and url at key url as URL instance.

New in version 3.2.

content_type

Read-only property with content part of Content-Type header.

Note

Returns value is 'application/octet-stream' if no Content-Type header present in HTTP headers according to RFC 2616. To make sure Content-Type header is not present in the server reply, use headers or raw_headers, e.g. 'CONTENT-TYPE' not in resp.headers.

charset

Read-only property that specifies the encoding for the request’s BODY.

The value is parsed from the Content-Type HTTP header.

Returns str like 'utf-8' or None if no Content-Type header present in HTTP headers or it has no charset information.

content_disposition

Read-only property that specified the Content-Disposition HTTP header.

Instance of ContentDisposition or None if no Content-Disposition header present in HTTP headers.

history

A Sequence of ClientResponse objects of preceding requests (earliest request first) if there were redirects, an empty sequence otherwise.

close()

Close response and underlying connection.

For keep-alive support see release().

coroutine read()

Read the whole response’s body as bytes.

Close underlying connection if data reading gets an error, release connection otherwise.

Raise an aiohttp.ClientResponseError if the data can’t be read.

Return bytes

read BODY.

See also

close(), release().

coroutine release()

It is not required to call release on the response object. When the client fully receives the payload, the underlying connection automatically returns back to pool. If the payload is not fully read, the connection is closed

raise_for_status()

Raise an aiohttp.ClientResponseError if the response status is 400 or higher.

Do nothing for success responses (less than 400).

coroutine text(encoding=None)

Read response’s body and return decoded str using specified encoding parameter.

If encoding is None content encoding is autocalculated using Content-Type HTTP header and chardet tool if the header is not provided by server.

cchardet is used with fallback to chardet if cchardet is not available.

Close underlying connection if data reading gets an error, release connection otherwise.

Parameters

encoding (str) – text encoding used for BODY decoding, or None for encoding autodetection (default).

Return str

decoded BODY

Raises

LookupError – if the encoding detected by chardet or cchardet is unknown by Python (e.g. VISCII).

Note

If response has no charset info in Content-Type HTTP header cchardet / chardet is used for content encoding autodetection.

It may hurt performance. If page encoding is known passing explicit encoding parameter might help:

await resp.text('ISO-8859-1')
coroutine json(*, encoding=None, loads=json.loads, content_type='application/json')

Read response’s body as JSON, return dict using specified encoding and loader. If data is not still available a read call will be done,

If encoding is None content encoding is autocalculated using cchardet or chardet as fallback if cchardet is not available.

if response’s content-type does not match content_type parameter aiohttp.ContentTypeError get raised. To disable content type check pass None value.

Parameters

  • encoding (str) –

    text encoding used for BODY decoding, or None for encoding autodetection (default).

    By the standard JSON encoding should be UTF-8 but practice beats purity: some servers return non-UTF responses. Autodetection works pretty fine anyway.

  • loads (callable) – callable() used for loading JSON data, json.loads() by default.

  • content_type (str) – specify response’s content-type, if content type does not match raise aiohttp.ClientResponseError. To disable content-type check, pass None as value. (default: application/json).

Returns

BODY as JSON data parsed by loads parameter or None if BODY is empty or contains white-spaces only.

request_info

A namedtuple with request URL and headers from ClientRequest object, aiohttp.RequestInfo instance.

get_encoding()

Automatically detect content encoding using charset info in Content-Type HTTP header. If this info is not exists or there are no appropriate codecs for encoding then cchardet / chardet is used.

Beware that it is not always safe to use the result of this function to decode a response. Some encodings detected by cchardet are not known by Python (e.g. VISCII).

Raises

RuntimeError – if called before the body has been read, for cchardet usage

New in version 3.0.

ClientWebSocketResponse

To connect to a websocket server aiohttp.ws_connect() or aiohttp.ClientSession.ws_connect() coroutines should be used, do not create an instance of class ClientWebSocketResponse manually.

class aiohttp.ClientWebSocketResponse

Class for handling client-side websockets.

closed

Read-only property, True if close() has been called or CLOSE message has been received from peer.

protocol

Websocket subprotocol chosen after start() call.

May be None if server and client protocols are not overlapping.

get_extra_info(name, default=None)

Reads extra info from connection’s transport

exception()

Returns exception if any occurs or returns None.

coroutine ping(message=b'')

Send PING to peer.

Parameters

message – optional payload of ping message, str (converted to UTF-8 encoded bytes) or bytes.

Changed in version 3.0: The method is converted into coroutine

coroutine pong(message=b'')

Send PONG to peer.

Parameters

message – optional payload of pong message, str (converted to UTF-8 encoded bytes) or bytes.

Changed in version 3.0: The method is converted into coroutine

coroutine send_str(data, compress=None)

Send data to peer as TEXT message.

Parameters

  • data (str) – data to send.

  • compress (int) – sets specific level of compression for single message, None for not overriding per-socket setting.

Raises

TypeError – if data is not str

Changed in version 3.0: The method is converted into coroutine, compress parameter added.

coroutine send_bytes(data, compress=None)

Send data to peer as BINARY message.

Parameters

  • data – data to send.

  • compress (int) – sets specific level of compression for single message, None for not overriding per-socket setting.

Raises

TypeError – if data is not bytes, bytearray or memoryview.

Changed in version 3.0: The method is converted into coroutine, compress parameter added.

coroutine send_json(data, compress=None, *, dumps=json.dumps)

Send data to peer as JSON string.

Parameters

  • data – data to send.

  • compress (int) – sets specific level of compression for single message, None for not overriding per-socket setting.

  • dumps (callable) – any callable that accepts an object and returns a JSON string (json.dumps() by default).

Raises

Changed in version 3.0: The method is converted into coroutine, compress parameter added.

coroutine close(*, code=1000, message=b'')

A coroutine that initiates closing handshake by sending CLOSE message. It waits for close response from server. To add a timeout to close() call just wrap the call with asyncio.wait() or asyncio.wait_for().

Parameters

  • code (int) – closing code

  • message – optional payload of close message, str (converted to UTF-8 encoded bytes) or bytes.

coroutine receive()

A coroutine that waits upcoming data message from peer and returns it.

The coroutine implicitly handles PING, PONG and CLOSE without returning the message.

It process ping-pong game and performs closing handshake internally.

Returns

WSMessage

coroutine receive_str()

A coroutine that calls receive() but also asserts the message type is TEXT.

Return str

peer’s message content.

Raises

TypeError – if message is BINARY.

coroutine receive_bytes()

A coroutine that calls receive() but also asserts the message type is BINARY.

Return bytes

peer’s message content.

Raises

TypeError – if message is TEXT.

coroutine receive_json(*, loads=json.loads)

A coroutine that calls receive_str() and loads the JSON string to a Python dict.

Parameters

loads (callable) – any callable that accepts str and returns dict with parsed JSON (json.loads() by default).

Return dict

loaded JSON content

Raises
Utilities
ClientTimeout
class aiohttp.ClientTimeout(*, total=None, connect=None, sock_connect=None, sock_read=None)

A data class for client timeout settings.

See Timeouts for usage examples.

total

Total number of seconds for the whole request.

float, None by default.

connect

Maximal number of seconds for acquiring a connection from pool. The time consists connection establishment for a new connection or waiting for a free connection from a pool if pool connection limits are exceeded.

For pure socket connection establishment time use sock_connect.

float, None by default.

sock_connect

Maximal number of seconds for connecting to a peer for a new connection, not given from a pool. See also connect.

float, None by default.

sock_read

Maximal number of seconds for reading a portion of data from a peer.

float, None by default.

New in version 3.3.

RequestInfo
class aiohttp.RequestInfo

A data class with request URL and headers from ClientRequest object, available as ClientResponse.request_info attribute.

url

Requested url, yarl.URL instance.

method

Request HTTP method like 'GET' or 'POST', str.

headers

HTTP headers for request, multidict.CIMultiDict instance.

real_url

Requested url with URL fragment unstripped, yarl.URL instance.

New in version 3.2.

BasicAuth
class aiohttp.BasicAuth(login, password='', encoding='latin1')

HTTP basic authentication helper.

Parameters

  • login (str) – login

  • password (str) – password

  • encoding (str) – encoding ('latin1' by default)

Should be used for specifying authorization data in client API, e.g. auth parameter for ClientSession.request().

classmethod decode(auth_header, encoding='latin1')

Decode HTTP basic authentication credentials.

Parameters

  • auth_header (str) – The Authorization header to decode.

  • encoding (str) – (optional) encoding (‘latin1’ by default)

Returns

decoded authentication data, BasicAuth.

classmethod from_url(url)

Constructed credentials info from url’s user and password parts.

Returns

credentials data, BasicAuth or None is credentials are not provided.

New in version 2.3.

encode()

Encode credentials into string suitable for Authorization header etc.

Returns

encoded authentication data, str.

CookieJar
class aiohttp.CookieJar(*, unsafe=False, quote_cookie=True, loop=None)

The cookie jar instance is available as ClientSession.cookie_jar.

The jar contains Morsel items for storing internal cookie data.

API provides a count of saved cookies:

len(session.cookie_jar)

These cookies may be iterated over:

for cookie in session.cookie_jar:
    print(cookie.key)
    print(cookie["domain"])

The class implements collections.abc.Iterable, collections.abc.Sized and aiohttp.AbstractCookieJar interfaces.

Implements cookie storage adhering to RFC 6265.

Parameters

  • unsafe (bool) – (optional) Whether to accept cookies from IPs.

  • quote_cookie (bool) –

    (optional) Whether to quote cookies according to

    RFC 2109. Some backend systems (not compatible with RFC mentioned above) does not support quoted cookies.

    New in version 3.7.

  • loop (bool) –

    an event loop instance. See aiohttp.abc.AbstractCookieJar

    Deprecated since version 2.0.

update_cookies(cookies, response_url=None)

Update cookies returned by server in Set-Cookie header.

Parameters

  • cookies – a collections.abc.Mapping (e.g. dict, SimpleCookie) or iterable of pairs with cookies returned by server’s response.

  • response_url (str) – URL of response, None for shared cookies. Regular cookies are coupled with server’s URL and are sent only to this server, shared ones are sent in every client request.

filter_cookies(request_url)

Return jar’s cookies acceptable for URL and available in Cookie header for sending client requests for given URL.

Parameters

response_url (str) – request’s URL for which cookies are asked.

Returns

http.cookies.SimpleCookie with filtered cookies for given URL.

save(file_path)

Write a pickled representation of cookies into the file at provided path.

Parameters

file_path – Path to file where cookies will be serialized, str or pathlib.Path instance.

load(file_path)

Load a pickled representation of cookies from the file at provided path.

Parameters

file_path – Path to file from where cookies will be imported, str or pathlib.Path instance.

class aiohttp.DummyCookieJar(*, loop=None)

Dummy cookie jar which does not store cookies but ignores them.

Could be useful e.g. for web crawlers to iterate over Internet without blowing up with saved cookies information.

To install dummy cookie jar pass it into session instance:

jar = aiohttp.DummyCookieJar()
session = aiohttp.ClientSession(cookie_jar=DummyCookieJar())
class aiohttp.Fingerprint(digest)

Fingerprint helper for checking SSL certificates by SHA256 digest.

Parameters

digest (bytes) – SHA256 digest for certificate in DER-encoded binary form (see ssl.SSLSocket.getpeercert()).

To check fingerprint pass the object into ClientSession.get() call, e.g.:

import hashlib

with open(path_to_cert, 'rb') as f:
    digest = hashlib.sha256(f.read()).digest()

await session.get(url, ssl=aiohttp.Fingerprint(digest))

New in version 3.0.

FormData

A FormData object contains the form data and also handles encoding it into a body that is either multipart/form-data or application/x-www-form-urlencoded. multipart/form-data is used if at least one field is an io.IOBase object or was added with at least one optional argument to add_field (content_type, filename, or content_transfer_encoding). Otherwise, application/x-www-form-urlencoded is used.

FormData instances are callable and return a Payload on being called.

class aiohttp.FormData(fields, quote_fields=True, charset=None)

Helper class for multipart/form-data and application/x-www-form-urlencoded body generation.

Parameters

fields –

A container for the key/value pairs of this form.

Possible types are:

If it is a tuple or list, it must be a valid argument for add_fields.

For dict, multidict.MultiDict, and multidict.MultiDictProxy, the keys and values must be valid name and value arguments to add_field, respectively.

add_field(name, value, content_type=None, filename=None, content_transfer_encoding=None)

Add a field to the form.

Parameters

  • name (str) – Name of the field

  • value –

    Value of the field

    Possible types are:

  • content_type (str) – The field’s content-type header (optional)

  • filename (str) –

    The field’s filename (optional)

    If this is not set and value is a bytes, bytesarray, or memoryview object, the name argument is used as the filename unless content_transfer_encoding is specified.

    If filename is not set and value is an io.IOBase object, the filename is extracted from the object if possible.

  • content_transfer_encoding (str) – The field’s content-transfer-encoding header (optional)

add_fields(fields)

Add one or more fields to the form.

Parameters

fields –

An iterable containing:

Client exceptions

Exception hierarchy has been significantly modified in version 2.0. aiohttp defines only exceptions that covers connection handling and server response misbehaviors. For developer specific mistakes, aiohttp uses python standard exceptions like ValueError or TypeError.

Reading a response content may raise a ClientPayloadError exception. This exception indicates errors specific to the payload encoding. Such as invalid compressed data, malformed chunked-encoded chunks or not enough data that satisfy the content-length header.

All exceptions are available as members of aiohttp module.

exception aiohttp.ClientError

Base class for all client specific exceptions.

Derived from Exception

class aiohttp.ClientPayloadError

This exception can only be raised while reading the response payload if one of these errors occurs:

  1. invalid compression

  2. malformed chunked encoding

  3. not enough data that satisfy Content-Length HTTP header.

Derived from ClientError

exception aiohttp.InvalidURL

URL used for fetching is malformed, e.g. it does not contain host part.

Derived from ClientError and ValueError

url

Invalid URL, yarl.URL instance.

class aiohttp.ContentDisposition

Represent Content-Disposition header

value

A str instance. Value of Content-Disposition header itself, e.g. attachment.

filename

A str instance. Content filename extracted from parameters. May be None.

parameters

Read-only mapping contains all parameters.

Response errors
exception aiohttp.ClientResponseError

These exceptions could happen after we get response from server.

Derived from ClientError

request_info

Instance of RequestInfo object, contains information about request.

status

HTTP status code of response (int), e.g. 400.

message

Message of response (str), e.g. "OK".

headers

Headers in response, a list of pairs.

history

History from failed response, if available, else empty tuple.

A tuple of ClientResponse objects used for handle redirection responses.

code

HTTP status code of response (int), e.g. 400.

Deprecated since version 3.1.

class aiohttp.WSServerHandshakeError

Web socket server response error.

Derived from ClientResponseError

class aiohttp.ContentTypeError

Invalid content type.

Derived from ClientResponseError

New in version 2.3.

class aiohttp.TooManyRedirects

Client was redirected too many times.

Maximum number of redirects can be configured by using parameter max_redirects in request.

Derived from ClientResponseError

New in version 3.2.

Connection errors
class aiohttp.ClientConnectionError

These exceptions related to low-level connection problems.

Derived from ClientError

class aiohttp.ClientOSError

Subset of connection errors that are initiated by an OSError exception.

Derived from ClientConnectionError and OSError

class aiohttp.ClientConnectorError

Connector related exceptions.

Derived from ClientOSError

class aiohttp.ClientProxyConnectionError

Derived from ClientConnectorError

class aiohttp.ServerConnectionError

Derived from ClientConnectionError

class aiohttp.ClientSSLError

Derived from ClientConnectorError

class aiohttp.ClientConnectorSSLError

Response ssl error.

Derived from ClientSSLError and ssl.SSLError

class aiohttp.ClientConnectorCertificateError

Response certificate error.

Derived from ClientSSLError and ssl.CertificateError

class aiohttp.ServerDisconnectedError

Server disconnected.

Derived from ServerDisconnectionError

message

Partially parsed HTTP message (optional).

class aiohttp.ServerTimeoutError

Server operation timeout: read timeout, etc.

Derived from ServerConnectionError and asyncio.TimeoutError

class aiohttp.ServerFingerprintMismatch

Server fingerprint mismatch.

Derived from ServerConnectionError

Tracing Reference

New in version 3.0.

A reference for client tracing API.

See also

Client Tracing for tracing usage instructions.

Request life cycle

A request goes through the following stages and corresponding fallbacks.

Overview

Name

Description

start

on_request_start

redirect

on_request_redirect

acquire_connection

Connection acquiring

headers_received

exception

on_request_exception

end

on_request_end

headers_sent

chunk_sent

on_request_chunk_sent

chunk_received

on_response_chunk_received
Connection acquiring

Name

Description

begin

end

queued_start

on_connection_queued_start

create_start

on_connection_create_start

reuseconn

on_connection_reuseconn

queued_end

on_connection_queued_end

create_end

on_connection_create_end

exception

Exception raised

resolve_dns

DNS resolving

sock_connect

Connection establishment
DNS resolving

Name

Description

begin

end

exception

Exception raised

resolve_end

on_dns_resolvehost_end

resolve_start

on_dns_resolvehost_start

cache_hit

on_dns_cache_hit

cache_miss

on_dns_cache_miss
TraceConfig
class aiohttp.TraceConfig(trace_config_ctx_factory=SimpleNamespace)

Trace config is the configuration object used to trace requests launched by a ClientSession object using different events related to different parts of the request flow.

Parameters

trace_config_ctx_factory – factory used to create trace contexts, default class used types.SimpleNamespace

trace_config_ctx(trace_request_ctx=None)
Parameters

trace_request_ctx – Will be used to pass as a kw for the trace_config_ctx_factory.

Build a new trace context from the config.

Every signal handler should have the following signature:

async def on_signal(session, context, params): ...

where session is ClientSession instance, context is an object returned by trace_config_ctx() call and params is a data class with signal parameters. The type of params depends on subscribed signal and described below.

on_request_start

Property that gives access to the signals that will be executed when a request starts.

params is aiohttp.TraceRequestStartParams instance.

on_request_chunk_sent

Property that gives access to the signals that will be executed when a chunk of request body is sent.

params is aiohttp.TraceRequestChunkSentParams instance.

New in version 3.1.

on_response_chunk_received

Property that gives access to the signals that will be executed when a chunk of response body is received.

params is aiohttp.TraceResponseChunkReceivedParams instance.

New in version 3.1.

on_request_redirect

Property that gives access to the signals that will be executed when a redirect happens during a request flow.

params is aiohttp.TraceRequestRedirectParams instance.

on_request_end

Property that gives access to the signals that will be executed when a request ends.

params is aiohttp.TraceRequestEndParams instance.

on_request_exception

Property that gives access to the signals that will be executed when a request finishes with an exception.

params is aiohttp.TraceRequestExceptionParams instance.

on_connection_queued_start

Property that gives access to the signals that will be executed when a request has been queued waiting for an available connection.

params is aiohttp.TraceConnectionQueuedStartParams instance.

on_connection_queued_end

Property that gives access to the signals that will be executed when a request that was queued already has an available connection.

params is aiohttp.TraceConnectionQueuedEndParams instance.

on_connection_create_start

Property that gives access to the signals that will be executed when a request creates a new connection.

params is aiohttp.TraceConnectionCreateStartParams instance.

on_connection_create_end

Property that gives access to the signals that will be executed when a request that created a new connection finishes its creation.

params is aiohttp.TraceConnectionCreateEndParams instance.

on_connection_reuseconn

Property that gives access to the signals that will be executed when a request reuses a connection.

params is aiohttp.TraceConnectionReuseconnParams instance.

on_dns_resolvehost_start

Property that gives access to the signals that will be executed when a request starts to resolve the domain related with the request.

params is aiohttp.TraceDnsResolveHostStartParams instance.

on_dns_resolvehost_end

Property that gives access to the signals that will be executed when a request finishes to resolve the domain related with the request.

params is aiohttp.TraceDnsResolveHostEndParams instance.

on_dns_cache_hit

Property that gives access to the signals that will be executed when a request was able to use a cached DNS resolution for the domain related with the request.

params is aiohttp.TraceDnsCacheHitParams instance.

on_dns_cache_miss

Property that gives access to the signals that will be executed when a request was not able to use a cached DNS resolution for the domain related with the request.

params is aiohttp.TraceDnsCacheMissParams instance.

TraceRequestStartParams
class aiohttp.TraceRequestStartParams

See TraceConfig.on_request_start for details.

method

Method that will be used to make the request.

url

URL that will be used for the request.

headers

Headers that will be used for the request, can be mutated.

TraceRequestChunkSentParams
class aiohttp.TraceRequestChunkSentParams

New in version 3.1.

See TraceConfig.on_request_chunk_sent for details.

method

Method that will be used to make the request.

url

URL that will be used for the request.

chunk

Bytes of chunk sent

TraceResponseChunkReceivedParams
class aiohttp.TraceResponseChunkReceivedParams

New in version 3.1.

See TraceConfig.on_response_chunk_received for details.

method

Method that will be used to make the request.

url

URL that will be used for the request.

chunk

Bytes of chunk received

TraceRequestEndParams
class aiohttp.TraceRequestEndParams

See TraceConfig.on_request_end for details.

method

Method used to make the request.

url

URL used for the request.

headers

Headers used for the request.

response

Response ClientResponse.

TraceRequestExceptionParams
class aiohttp.TraceRequestExceptionParams

See TraceConfig.on_request_exception for details.

method

Method used to make the request.

url

URL used for the request.

headers

Headers used for the request.

exception

Exception raised during the request.

TraceRequestRedirectParams
class aiohttp.TraceRequestRedirectParams

See TraceConfig.on_request_redirect for details.

method

Method used to get this redirect request.

url

URL used for this redirect request.

headers

Headers used for this redirect.

response

Response ClientResponse got from the redirect.

TraceConnectionQueuedStartParams
class aiohttp.TraceConnectionQueuedStartParams

See TraceConfig.on_connection_queued_start for details.

There are no attributes right now.

TraceConnectionQueuedEndParams
class aiohttp.TraceConnectionQueuedEndParams

See TraceConfig.on_connection_queued_end for details.

There are no attributes right now.

TraceConnectionCreateStartParams
class aiohttp.TraceConnectionCreateStartParams

See TraceConfig.on_connection_create_start for details.

There are no attributes right now.

TraceConnectionCreateEndParams
class aiohttp.TraceConnectionCreateEndParams

See TraceConfig.on_connection_create_end for details.

There are no attributes right now.

TraceConnectionReuseconnParams
class aiohttp.TraceConnectionReuseconnParams

See TraceConfig.on_connection_reuseconn for details.

There are no attributes right now.

TraceDnsResolveHostStartParams
class aiohttp.TraceDnsResolveHostStartParams

See TraceConfig.on_dns_resolvehost_start for details.

host

Host that will be resolved.

TraceDnsResolveHostEndParams
class aiohttp.TraceDnsResolveHostEndParams

See TraceConfig.on_dns_resolvehost_end for details.

host

Host that has been resolved.

TraceDnsCacheHitParams
class aiohttp.TraceDnsCacheHitParams

See TraceConfig.on_dns_cache_hit for details.

host

Host found in the cache.

TraceDnsCacheMissParams
class aiohttp.TraceDnsCacheMissParams

See TraceConfig.on_dns_cache_miss for details.

host

Host didn’t find the cache.

The aiohttp Request Lifecycle

Why is aiohttp client API that way?

The first time you use aiohttp, you’ll notice that a simple HTTP request is performed not with one, but with up to three steps:

async with aiohttp.ClientSession() as session:
    async with session.get('http://python.org') as response:
        print(await response.text())

It’s especially unexpected when coming from other libraries such as the very popular requests, where the “hello world” looks like this:

response = requests.get('http://python.org')
print(response.text)

So why is the aiohttp snippet so verbose?

Because aiohttp is asynchronous, its API is designed to make the most out of non-blocking network operations. In code like this, requests will block three times, and does it transparently, while aiohttp gives the event loop three opportunities to switch context:

  • When doing the .get(), both libraries send a GET request to the remote server. For aiohttp, this means asynchronous I/O, which is marked here with an async with that gives you the guarantee that not only it doesn’t block, but that it’s cleanly finalized.

  • When doing response.text in requests, you just read an attribute. The call to .get() already preloaded and decoded the entire response payload, in a blocking manner. aiohttp loads only the headers when .get() is executed, letting you decide to pay the cost of loading the body afterward, in a second asynchronous operation. Hence the await response.text().

  • async with aiohttp.ClientSession() does not perform I/O when entering the block, but at the end of it, it will ensure all remaining resources are closed correctly. Again, this is done asynchronously and must be marked as such. The session is also a performance tool, as it manages a pool of connections for you, allowing you to reuse them instead of opening and closing a new one at each request. You can even manage the pool size by passing a connector object.

Using a session as a best practice

The requests library does in fact also provides a session system. Indeed, it lets you do:

with requests.Session() as session:
    response = session.get('http://python.org')
    print(response.text)

It’s just not the default behavior, nor is it advertised early in the documentation. Because of this, most users take a hit in performance, but can quickly start hacking. And for requests, it’s an understandable trade-off, since its goal is to be “HTTP for humans” and simplicity has always been more important than performance in this context.

However, if one uses aiohttp, one chooses asynchronous programming, a paradigm that makes the opposite trade-off: more verbosity for better performance. And so the library default behavior reflects this, encouraging you to use performant best practices from the start.

How to use the ClientSession ?

By default the aiohttp.ClientSession object will hold a connector with a maximum of 100 connections, putting the rest in a queue. This is quite a big number, this means you must be connected to a hundred different servers (not pages!) concurrently before even having to consider if your task needs resource adjustment.

In fact, you can picture the session object as a user starting and closing a browser: it wouldn’t make sense to do that every time you want to load a new tab.

So you are expected to reuse a session object and make many requests from it. For most scripts and average-sized software, this means you can create a single session, and reuse it for the entire execution of the program. You can even pass the session around as a parameter in functions. For example, the typical “hello world”:

import aiohttp
import asyncio

async def main():
    async with aiohttp.ClientSession() as session:
        async with session.get('http://python.org') as response:
            html = await response.text()
            print(html)

loop = asyncio.get_event_loop()
loop.run_until_complete(main())

Can become this:

import aiohttp
import asyncio

async def fetch(session, url):
    async with session.get(url) as response:
        return await response.text()

async def main():
    async with aiohttp.ClientSession() as session:
        html = await fetch(session, 'http://python.org')
        print(html)

loop = asyncio.get_event_loop()
loop.run_until_complete(main())

On more complex code bases, you can even create a central registry to hold the session object from anywhere in the code, or a higher level Client class that holds a reference to it.

When to create more than one session object then? It arises when you want more granularity with your resources management:

  • you want to group connections by a common configuration. e.g: sessions can set cookies, headers, timeout values, etc. that are shared for all connections they hold.

  • you need several threads and want to avoid sharing a mutable object between them.

  • you want several connection pools to benefit from different queues and assign priorities. e.g: one session never uses the queue and is for high priority requests, the other one has a small concurrency limit and a very long queue, for non important requests.

Server

The page contains all information about aiohttp Server API:

Web Server Quickstart

Run a Simple Web Server

In order to implement a web server, first create a request handler.

A request handler must be a coroutine that accepts a Request instance as its only parameter and returns a Response instance:

from aiohttp import web

async def hello(request):
    return web.Response(text="Hello, world")

Next, create an Application instance and register the request handler on a particular HTTP method and path:

app = web.Application()
app.add_routes([web.get('/', hello)])

After that, run the application by run_app() call:

web.run_app(app)

That’s it. Now, head over to http://localhost:8080/ to see the results.

Alternatively if you prefer route decorators create a route table and register a web-handler:

routes = web.RouteTableDef()

@routes.get('/')
async def hello(request):
    return web.Response(text="Hello, world")

app = web.Application()
app.add_routes(routes)
web.run_app(app)

Both ways essentially do the same work, the difference is only in your taste: do you prefer Django style with famous urls.py or Flask with shiny route decorators.

aiohttp server documentation uses both ways in code snippets to emphasize their equality, switching from one style to another is very trivial.

See also

Graceful shutdown section explains what run_app() does and how to implement complex server initialization/finalization from scratch.

Application runners for more handling more complex cases like asynchronous web application serving and multiple hosts support.

Command Line Interface (CLI)

aiohttp.web implements a basic CLI for quickly serving an Application in development over TCP/IP:

$ python -m aiohttp.web -H localhost -P 8080 package.module:init_func

package.module:init_func should be an importable callable that accepts a list of any non-parsed command-line arguments and returns an Application instance after setting it up:

def init_func(argv):
    app = web.Application()
    app.router.add_get("/", index_handler)
    return app
Handler

A request handler must be a coroutine that accepts a Request instance as its only argument and returns a StreamResponse derived (e.g. Response) instance:

async def handler(request):
    return web.Response()

Handlers are setup to handle requests by registering them with the Application.add_routes() on a particular route (HTTP method and path pair) using helpers like get() and post():

app.add_routes([web.get('/', handler),
                web.post('/post', post_handler),
                web.put('/put', put_handler)])

Or use route decorators:

routes = web.RouteTableDef()

@routes.get('/')
async def get_handler(request):
    ...

@routes.post('/post')
async def post_handler(request):
    ...

@routes.put('/put')
async def put_handler(request):
    ...

app.add_routes(routes)

Wildcard HTTP method is also supported by route() or RouteTableDef.route(), allowing a handler to serve incoming requests on a path having any HTTP method:

app.add_routes([web.route('*', '/path', all_handler)])

The HTTP method can be queried later in the request handler using the Request.method property.

By default endpoints added with GET method will accept HEAD requests and return the same response headers as they would for a GET request. You can also deny HEAD requests on a route:

web.get('/', handler, allow_head=False)

Here handler won’t be called on HEAD request and the server will respond with 405: Method Not Allowed.

Resources and Routes

Internally routes are served by Application.router (UrlDispatcher instance).

The router is a list of resources.

Resource is an entry in route table which corresponds to requested URL.

Resource in turn has at least one route.

Route corresponds to handling HTTP method by calling web handler.

Thus when you add a route the resouce object is created under the hood.

The library implementation merges all subsequent route additions for the same path adding the only resource for all HTTP methods.

Consider two examples:

app.add_routes([web.get('/path1', get_1),
                web.post('/path1', post_1),
                web.get('/path2', get_2),
                web.post('/path2', post_2)]

and:

app.add_routes([web.get('/path1', get_1),
                web.get('/path2', get_2),
                web.post('/path2', post_2),
                web.post('/path1', post_1)]

First one is optimized. You have got the idea.

Variable Resources

Resource may have variable path also. For instance, a resource with the path '/a/{name}/c' would match all incoming requests with paths such as '/a/b/c', '/a/1/c', and '/a/etc/c'.

A variable part is specified in the form {identifier}, where the identifier can be used later in a request handler to access the matched value for that part. This is done by looking up the identifier in the Request.match_info mapping:

@routes.get('/{name}')
async def variable_handler(request):
    return web.Response(
        text="Hello, {}".format(request.match_info['name']))

By default, each part matches the regular expression [^{}/]+.

You can also specify a custom regex in the form {identifier:regex}:

web.get(r'/{name:\d+}', handler)
Reverse URL Constructing using Named Resources

Routes can also be given a name:

@routes.get('/root', name='root')
async def handler(request):
    ...

Which can then be used to access and build a URL for that resource later (e.g. in a request handler):

url = request.app.router['root'].url_for().with_query({"a": "b", "c": "d"})
assert url == URL('/root?a=b&c=d')

A more interesting example is building URLs for variable resources:

app.router.add_resource(r'/{user}/info', name='user-info')

In this case you can also pass in the parts of the route:

url = request.app.router['user-info'].url_for(user='john_doe')
url_with_qs = url.with_query("a=b")
assert url_with_qs == '/john_doe/info?a=b'
Organizing Handlers in Classes

As discussed above, handlers can be first-class coroutines:

async def hello(request):
    return web.Response(text="Hello, world")

app.router.add_get('/', hello)

But sometimes it’s convenient to group logically similar handlers into a Python class.

Since aiohttp.web does not dictate any implementation details, application developers can organize handlers in classes if they so wish:

class Handler:

    def __init__(self):
        pass

    async def handle_intro(self, request):
        return web.Response(text="Hello, world")

    async def handle_greeting(self, request):
        name = request.match_info.get('name', "Anonymous")
        txt = "Hello, {}".format(name)
        return web.Response(text=txt)

handler = Handler()
app.add_routes([web.get('/intro', handler.handle_intro),
                web.get('/greet/{name}', handler.handle_greeting)])
Class Based Views

aiohttp.web has support for class based views.

You can derive from View and define methods for handling http requests:

class MyView(web.View):
    async def get(self):
        return await get_resp(self.request)

    async def post(self):
        return await post_resp(self.request)

Handlers should be coroutines accepting self only and returning response object as regular web-handler. Request object can be retrieved by View.request property.

After implementing the view (MyView from example above) should be registered in application’s router:

web.view('/path/to', MyView)

or:

@routes.view('/path/to')
class MyView(web.View):
    ...

Example will process GET and POST requests for /path/to but raise 405 Method not allowed exception for unimplemented HTTP methods.

Resource Views

All registered resources in a router can be viewed using the UrlDispatcher.resources() method:

for resource in app.router.resources():
    print(resource)

A subset of the resources that were registered with a name can be viewed using the UrlDispatcher.named_resources() method:

for name, resource in app.router.named_resources().items():
    print(name, resource)
Alternative ways for registering routes

Code examples shown above use imperative style for adding new routes: they call app.router.add_get(...) etc.

There are two alternatives: route tables and route decorators.

Route tables look like Django way:

async def handle_get(request):
    ...


async def handle_post(request):
    ...

app.router.add_routes([web.get('/get', handle_get),
                       web.post('/post', handle_post),

The snippet calls add_routes() to register a list of route definitions (aiohttp.web.RouteDef instances) created by aiohttp.web.get() or aiohttp.web.post() functions.

See also

RouteDef and StaticDef reference.

Route decorators are closer to Flask approach:

routes = web.RouteTableDef()

@routes.get('/get')
async def handle_get(request):
    ...


@routes.post('/post')
async def handle_post(request):
    ...

app.router.add_routes(routes)

It is also possible to use decorators with class-based views:

routes = web.RouteTableDef()

@routes.view("/view")
class MyView(web.View):
    async def get(self):
        ...

    async def post(self):
        ...

app.router.add_routes(routes)

The example creates a aiohttp.web.RouteTableDef container first.

The container is a list-like object with additional decorators aiohttp.web.RouteTableDef.get(), aiohttp.web.RouteTableDef.post() etc. for registering new routes.

After filling the container add_routes() is used for adding registered route definitions into application’s router.

See also

RouteTableDef reference.

All tree ways (imperative calls, route tables and decorators) are equivalent, you could use what do you prefer or even mix them on your own.

New in version 2.3.

JSON Response

It is a common case to return JSON data in response, aiohttp.web provides a shortcut for returning JSON – aiohttp.web.json_response():

async def handler(request):
    data = {'some': 'data'}
    return web.json_response(data)

The shortcut method returns aiohttp.web.Response instance so you can for example set cookies before returning it from handler.

User Sessions

Often you need a container for storing user data across requests. The concept is usually called a session.

aiohttp.web has no built-in concept of a session, however, there is a third-party library, aiohttp_session, that adds session support:

import asyncio
import time
import base64
from cryptography import fernet
from aiohttp import web
from aiohttp_session import setup, get_session, session_middleware
from aiohttp_session.cookie_storage import EncryptedCookieStorage

async def handler(request):
    session = await get_session(request)
    last_visit = session['last_visit'] if 'last_visit' in session else None
    text = 'Last visited: {}'.format(last_visit)
    return web.Response(text=text)

async def make_app():
    app = web.Application()
    # secret_key must be 32 url-safe base64-encoded bytes
    fernet_key = fernet.Fernet.generate_key()
    secret_key = base64.urlsafe_b64decode(fernet_key)
    setup(app, EncryptedCookieStorage(secret_key))
    app.add_routes([web.get('/', handler)])
    return app

web.run_app(make_app())
HTTP Forms

HTTP Forms are supported out of the box.

If form’s method is "GET" (<form method="get">) use Request.query for getting form data.

To access form data with "POST" method use Request.post() or Request.multipart().

Request.post() accepts both 'application/x-www-form-urlencoded' and 'multipart/form-data' form’s data encoding (e.g. <form enctype="multipart/form-data">). It stores files data in temporary directory. If client_max_size is specified post raises ValueError exception. For efficiency use Request.multipart(), It is especially effective for uploading large files (File Uploads).

Values submitted by the following form:

<form action="/login" method="post" accept-charset="utf-8"
      enctype="application/x-www-form-urlencoded">

    <label for="login">Login</label>
    <input id="login" name="login" type="text" value="" autofocus/>
    <label for="password">Password</label>
    <input id="password" name="password" type="password" value=""/>

    <input type="submit" value="login"/>
</form>

could be accessed as:

async def do_login(request):
    data = await request.post()
    login = data['login']
    password = data['password']
File Uploads

aiohttp.web has built-in support for handling files uploaded from the browser.

First, make sure that the HTML <form> element has its enctype attribute set to enctype="multipart/form-data". As an example, here is a form that accepts an MP3 file:

<form action="/store/mp3" method="post" accept-charset="utf-8"
      enctype="multipart/form-data">

    <label for="mp3">Mp3</label>
    <input id="mp3" name="mp3" type="file" value=""/>

    <input type="submit" value="submit"/>
</form>

Then, in the request handler you can access the file input field as a FileField instance. FileField is simply a container for the file as well as some of its metadata:

async def store_mp3_handler(request):

    # WARNING: don't do that if you plan to receive large files!
    data = await request.post()

    mp3 = data['mp3']

    # .filename contains the name of the file in string format.
    filename = mp3.filename

    # .file contains the actual file data that needs to be stored somewhere.
    mp3_file = data['mp3'].file

    content = mp3_file.read()

    return web.Response(body=content,
                        headers=MultiDict(
                            {'CONTENT-DISPOSITION': mp3_file}))

You might have noticed a big warning in the example above. The general issue is that Request.post() reads the whole payload in memory, resulting in possible OOM errors. To avoid this, for multipart uploads, you should use Request.multipart() which returns a multipart reader:

async def store_mp3_handler(request):

    reader = await request.multipart()

    # /!\ Don't forget to validate your inputs /!\

    # reader.next() will `yield` the fields of your form

    field = await reader.next()
    assert field.name == 'name'
    name = await field.read(decode=True)

    field = await reader.next()
    assert field.name == 'mp3'
    filename = field.filename
    # You cannot rely on Content-Length if transfer is chunked.
    size = 0
    with open(os.path.join('/spool/yarrr-media/mp3/', filename), 'wb') as f:
        while True:
            chunk = await field.read_chunk()  # 8192 bytes by default.
            if not chunk:
                break
            size += len(chunk)
            f.write(chunk)

    return web.Response(text='{} sized of {} successfully stored'
                             ''.format(filename, size))
WebSockets

aiohttp.web supports WebSockets out-of-the-box.

To setup a WebSocket, create a WebSocketResponse in a request handler and then use it to communicate with the peer:

async def websocket_handler(request):

    ws = web.WebSocketResponse()
    await ws.prepare(request)

    async for msg in ws:
        if msg.type == aiohttp.WSMsgType.TEXT:
            if msg.data == 'close':
                await ws.close()
            else:
                await ws.send_str(msg.data + '/answer')
        elif msg.type == aiohttp.WSMsgType.ERROR:
            print('ws connection closed with exception %s' %
                  ws.exception())

    print('websocket connection closed')

    return ws

The handler should be registered as HTTP GET processor:

app.add_routes([web.get('/ws', websocket_handler)])
Redirects

To redirect user to another endpoint - raise HTTPFound with an absolute URL, relative URL or view name (the argument from router):

raise web.HTTPFound('/redirect')

The following example shows redirect to view named ‘login’ in routes:

async def handler(request):
    location = request.app.router['login'].url_for()
    raise web.HTTPFound(location=location)

router.add_get('/handler', handler)
router.add_get('/login', login_handler, name='login')

Example with login validation:

@aiohttp_jinja2.template('login.html')
async def login(request):

    if request.method == 'POST':
        form = await request.post()
        error = validate_login(form)
        if error:
            return {'error': error}
        else:
            # login form is valid
            location = request.app.router['index'].url_for()
            raise web.HTTPFound(location=location)

    return {}

app.router.add_get('/', index, name='index')
app.router.add_get('/login', login, name='login')
app.router.add_post('/login', login, name='login')
Exceptions

aiohttp.web defines a set of exceptions for every HTTP status code.

Each exception is a subclass of HTTPException and relates to a single HTTP status code:

async def handler(request):
    raise aiohttp.web.HTTPFound('/redirect')

Warning

Returning HTTPException or its subclasses is deprecated and will be removed in subsequent aiohttp versions.

Each exception class has a status code according to RFC 2068: codes with 100-300 are not really errors; 400s are client errors, and 500s are server errors.

HTTP Exception hierarchy chart:

Exception
  HTTPException
    HTTPSuccessful
      * 200 - HTTPOk
      * 201 - HTTPCreated
      * 202 - HTTPAccepted
      * 203 - HTTPNonAuthoritativeInformation
      * 204 - HTTPNoContent
      * 205 - HTTPResetContent
      * 206 - HTTPPartialContent
    HTTPRedirection
      * 300 - HTTPMultipleChoices
      * 301 - HTTPMovedPermanently
      * 302 - HTTPFound
      * 303 - HTTPSeeOther
      * 304 - HTTPNotModified
      * 305 - HTTPUseProxy
      * 307 - HTTPTemporaryRedirect
      * 308 - HTTPPermanentRedirect
    HTTPError
      HTTPClientError
        * 400 - HTTPBadRequest
        * 401 - HTTPUnauthorized
        * 402 - HTTPPaymentRequired
        * 403 - HTTPForbidden
        * 404 - HTTPNotFound
        * 405 - HTTPMethodNotAllowed
        * 406 - HTTPNotAcceptable
        * 407 - HTTPProxyAuthenticationRequired
        * 408 - HTTPRequestTimeout
        * 409 - HTTPConflict
        * 410 - HTTPGone
        * 411 - HTTPLengthRequired
        * 412 - HTTPPreconditionFailed
        * 413 - HTTPRequestEntityTooLarge
        * 414 - HTTPRequestURITooLong
        * 415 - HTTPUnsupportedMediaType
        * 416 - HTTPRequestRangeNotSatisfiable
        * 417 - HTTPExpectationFailed
        * 421 - HTTPMisdirectedRequest
        * 422 - HTTPUnprocessableEntity
        * 424 - HTTPFailedDependency
        * 426 - HTTPUpgradeRequired
        * 428 - HTTPPreconditionRequired
        * 429 - HTTPTooManyRequests
        * 431 - HTTPRequestHeaderFieldsTooLarge
        * 451 - HTTPUnavailableForLegalReasons
      HTTPServerError
        * 500 - HTTPInternalServerError
        * 501 - HTTPNotImplemented
        * 502 - HTTPBadGateway
        * 503 - HTTPServiceUnavailable
        * 504 - HTTPGatewayTimeout
        * 505 - HTTPVersionNotSupported
        * 506 - HTTPVariantAlsoNegotiates
        * 507 - HTTPInsufficientStorage
        * 510 - HTTPNotExtended
        * 511 - HTTPNetworkAuthenticationRequired

All HTTP exceptions have the same constructor signature:

HTTPNotFound(*, headers=None, reason=None,
             body=None, text=None, content_type=None)

If not directly specified, headers will be added to the default response headers.

Classes HTTPMultipleChoices, HTTPMovedPermanently, HTTPFound, HTTPSeeOther, HTTPUseProxy, HTTPTemporaryRedirect have the following constructor signature:

HTTPFound(location, *, headers=None, reason=None,
          body=None, text=None, content_type=None)

where location is value for Location HTTP header.

HTTPMethodNotAllowed is constructed by providing the incoming unsupported method and list of allowed methods:

HTTPMethodNotAllowed(method, allowed_methods, *,
                     headers=None, reason=None,
                     body=None, text=None, content_type=None)

Web Server Advanced

Unicode support

aiohttp does requoting of incoming request path.

Unicode (non-ASCII) symbols are processed transparently on both route adding and resolving (internally everything is converted to percent-encoding form by yarl library).

But in case of custom regular expressions for Variable Resources please take care that URL is percent encoded: if you pass Unicode patterns they don’t match to requoted path.

Peer disconnection

When a client peer is gone a subsequent reading or writing raises OSError or more specific exception like ConnectionResetError.

The reason for disconnection is vary; it can be a network issue or explicit socket closing on the peer side without reading the whole server response.

aiohttp handles disconnection properly but you can handle it explicitly, e.g.:

async def handler(request):
    try:
        text = await request.text()
    except OSError:
        # disconnected
Passing a coroutine into run_app and Gunicorn

run_app() accepts either application instance or a coroutine for making an application. The coroutine based approach allows to perform async IO before making an app:

async def app_factory():
    await pre_init()
    app = web.Application()
    app.router.add_get(...)
    return app

web.run_app(app_factory())

Gunicorn worker supports a factory as well. For Gunicorn the factory should accept zero parameters:

async def my_web_app():
    app = web.Application()
    app.router.add_get(...)
    return app

Start gunicorn:

$ gunicorn my_app_module:my_web_app --bind localhost:8080 --worker-class aiohttp.GunicornWebWorker

New in version 3.1.

Custom Routing Criteria

Sometimes you need to register handlers on more complex criteria than simply a HTTP method and path pair.

Although UrlDispatcher does not support any extra criteria, routing based on custom conditions can be accomplished by implementing a second layer of routing in your application.

The following example shows custom routing based on the HTTP Accept header:

class AcceptChooser:

    def __init__(self):
        self._accepts = {}

    async def do_route(self, request):
        for accept in request.headers.getall('ACCEPT', []):
            acceptor = self._accepts.get(accept)
            if acceptor is not None:
                return (await acceptor(request))
        raise HTTPNotAcceptable()

    def reg_acceptor(self, accept, handler):
        self._accepts[accept] = handler


async def handle_json(request):
    # do json handling

async def handle_xml(request):
    # do xml handling

chooser = AcceptChooser()
app.add_routes([web.get('/', chooser.do_route)])

chooser.reg_acceptor('application/json', handle_json)
chooser.reg_acceptor('application/xml', handle_xml)
Static file handling

The best way to handle static files (images, JavaScripts, CSS files etc.) is using Reverse Proxy like nginx or CDN services.

But for development it’s very convenient to handle static files by aiohttp server itself.

To do it just register a new static route by RouteTableDef.static() or static() calls:

app.add_routes([web.static('/prefix', path_to_static_folder)])

routes.static('/prefix', path_to_static_folder)

When a directory is accessed within a static route then the server responses to client with HTTP/403 Forbidden by default. Displaying folder index instead could be enabled with show_index parameter set to True:

web.static('/prefix', path_to_static_folder, show_index=True)

When a symlink from the static directory is accessed, the server responses to client with HTTP/404 Not Found by default. To allow the server to follow symlinks, parameter follow_symlinks should be set to True:

web.static('/prefix', path_to_static_folder, follow_symlinks=True)

When you want to enable cache busting, parameter append_version can be set to True

Cache busting is the process of appending some form of file version hash to the filename of resources like JavaScript and CSS files. The performance advantage of doing this is that we can tell the browser to cache these files indefinitely without worrying about the client not getting the latest version when the file changes:

web.static('/prefix', path_to_static_folder, append_version=True)
Template Rendering

aiohttp.web does not support template rendering out-of-the-box.

However, there is a third-party library, aiohttp_jinja2, which is supported by the aiohttp authors.

Using it is rather simple. First, setup a jinja2 environment with a call to aiohttp_jinja2.setup():

app = web.Application()
aiohttp_jinja2.setup(app,
    loader=jinja2.FileSystemLoader('/path/to/templates/folder'))

After that you may use the template engine in your handlers. The most convenient way is to simply wrap your handlers with the aiohttp_jinja2.template() decorator:

@aiohttp_jinja2.template('tmpl.jinja2')
async def handler(request):
    return {'name': 'Andrew', 'surname': 'Svetlov'}

If you prefer the Mako template engine, please take a look at the aiohttp_mako library.

Warning

aiohttp_jinja2.template() should be applied before RouteTableDef.get() decorator and family, e.g. it must be the first (most down decorator in the chain):

@routes.get('/path')
@aiohttp_jinja2.template('tmpl.jinja2')
async def handler(request):
    return {'name': 'Andrew', 'surname': 'Svetlov'}
Reading from the same task in WebSockets

Reading from the WebSocket (await ws.receive()) must only be done inside the request handler task; however, writing (ws.send_str(...)) to the WebSocket, closing (await ws.close()) and canceling the handler task may be delegated to other tasks. See also FAQ section.

aiohttp.web creates an implicit asyncio.Task for handling every incoming request.

Note

While aiohttp.web itself only supports WebSockets without downgrading to LONG-POLLING, etc., our team supports SockJS, an aiohttp-based library for implementing SockJS-compatible server code.

Warning

Parallel reads from websocket are forbidden, there is no possibility to call WebSocketResponse.receive() from two tasks.

See FAQ section for instructions how to solve the problem.

Data Sharing aka No Singletons Please

aiohttp.web discourages the use of global variables, aka singletons. Every variable should have its own context that is not global.

So, Application and Request support a collections.abc.MutableMapping interface (i.e. they are dict-like objects), allowing them to be used as data stores.

Application’s config

For storing global-like variables, feel free to save them in an Application instance:

app['my_private_key'] = data

and get it back in the web-handler:

async def handler(request):
    data = request.app['my_private_key']

In case of nested applications the desired lookup strategy could be the following:

  1. Search the key in the current nested application.

  2. If the key is not found continue searching in the parent application(s).

For this please use Request.config_dict read-only property:

async def handler(request):
    data = request.config_dict['my_private_key']
Request’s storage

Variables that are only needed for the lifetime of a Request, can be stored in a Request:

async def handler(request):
  request['my_private_key'] = "data"
  ...

This is mostly useful for Middlewares and Signals handlers to store data for further processing by the next handlers in the chain.

Response’s storage

StreamResponse and Response objects also support collections.abc.MutableMapping interface. This is useful when you want to share data with signals and middlewares once all the work in the handler is done:

async def handler(request):
  [ do all the work ]
  response['my_metric'] = 123
  return response
Naming hint

To avoid clashing with other aiohttp users and third-party libraries, please choose a unique key name for storing data.

If your code is published on PyPI, then the project name is most likely unique and safe to use as the key. Otherwise, something based on your company name/url would be satisfactory (i.e. org.company.app).

ContextVars support

Starting from Python 3.7 asyncio has Context Variables as a context-local storage (a generalization of thread-local concept that works with asyncio tasks also).

aiohttp server supports it in the following way:

  • A server inherits the current task’s context used when creating it. aiohttp.web.run_app() runs a task for handling all underlying jobs running the app, but alternatively Application runners can be used.

  • Application initialization / finalization events (Application.cleanup_ctx, Application.on_startup and Application.on_shutdown, Application.on_cleanup) are executed inside the same context.

    E.g. all context modifications made on application startup are visible on teardown.

  • On every request handling aiohttp creates a context copy. web-handler has all variables installed on initialization stage. But the context modification made by a handler or middleware is invisible to another HTTP request handling call.

An example of context vars usage:

from contextvars import ContextVar

from aiohttp import web

VAR = ContextVar('VAR', default='default')


async def coro():
    return VAR.get()


async def handler(request):
    var = VAR.get()
    VAR.set('handler')
    ret = await coro()
    return web.Response(text='\n'.join([var,
                                        ret]))


async def on_startup(app):
    print('on_startup', VAR.get())
    VAR.set('on_startup')


async def on_cleanup(app):
    print('on_cleanup', VAR.get())
    VAR.set('on_cleanup')


async def init():
    print('init', VAR.get())
    VAR.set('init')
    app = web.Application()
    app.router.add_get('/', handler)

    app.on_startup.append(on_startup)
    app.on_cleanup.append(on_cleanup)
    return app


web.run_app(init())
print('done', VAR.get())

New in version 3.5.

Middlewares

aiohttp.web provides a powerful mechanism for customizing request handlers via middlewares.

A middleware is a coroutine that can modify either the request or response. For example, here’s a simple middleware which appends ' wink' to the response:

from aiohttp.web import middleware

@middleware
async def middleware(request, handler):
    resp = await handler(request)
    resp.text = resp.text + ' wink'
    return resp

Note

The example won’t work with streamed responses or websockets

Every middleware should accept two parameters, a request instance and a handler, and return the response or raise an exception. If the exception is not an instance of HTTPException it is converted to 500 HTTPInternalServerError after processing the middlewares chain.

Warning

Second argument should be named handler exactly.

When creating an Application, these middlewares are passed to the keyword-only middlewares parameter:

app = web.Application(middlewares=[middleware_1,
                                   middleware_2])

Internally, a single request handler is constructed by applying the middleware chain to the original handler in reverse order, and is called by the RequestHandler as a regular handler.

Since middlewares are themselves coroutines, they may perform extra await calls when creating a new handler, e.g. call database etc.

Middlewares usually call the handler, but they may choose to ignore it, e.g. displaying 403 Forbidden page or raising HTTPForbidden exception if the user does not have permissions to access the underlying resource. They may also render errors raised by the handler, perform some pre- or post-processing like handling CORS and so on.

The following code demonstrates middlewares execution order:

from aiohttp import web

async def test(request):
    print('Handler function called')
    return web.Response(text="Hello")

@web.middleware
async def middleware1(request, handler):
    print('Middleware 1 called')
    response = await handler(request)
    print('Middleware 1 finished')
    return response

@web.middleware
async def middleware2(request, handler):
    print('Middleware 2 called')
    response = await handler(request)
    print('Middleware 2 finished')
    return response


app = web.Application(middlewares=[middleware1, middleware2])
app.router.add_get('/', test)
web.run_app(app)

Produced output:

Middleware 1 called
Middleware 2 called
Handler function called
Middleware 2 finished
Middleware 1 finished
Example

A common use of middlewares is to implement custom error pages. The following example will render 404 errors using a JSON response, as might be appropriate a JSON REST service:

from aiohttp import web

@web.middleware
async def error_middleware(request, handler):
    try:
        response = await handler(request)
        if response.status != 404:
            return response
        message = response.message
    except web.HTTPException as ex:
        if ex.status != 404:
            raise
        message = ex.reason
    return web.json_response({'error': message})

app = web.Application(middlewares=[error_middleware])
Middleware Factory

A middleware factory is a function that creates a middleware with passed arguments. For example, here’s a trivial middleware factory:

def middleware_factory(text):
    @middleware
    async def sample_middleware(request, handler):
        resp = await handler(request)
        resp.text = resp.text + text
        return resp
    return sample_middleware

Remember that contrary to regular middlewares you need the result of a middleware factory not the function itself. So when passing a middleware factory to an app you actually need to call it:

app = web.Application(middlewares=[middleware_factory(' wink')])
Signals

Although middlewares can customize request handlers before or after a Response has been prepared, they can’t customize a Response while it’s being prepared. For this aiohttp.web provides signals.

For example, a middleware can only change HTTP headers for unprepared responses (see StreamResponse.prepare()), but sometimes we need a hook for changing HTTP headers for streamed responses and WebSockets. This can be accomplished by subscribing to the Application.on_response_prepare signal, which is called after default headers have been computed and directly before headers are sent:

async def on_prepare(request, response):
    response.headers['My-Header'] = 'value'

app.on_response_prepare.append(on_prepare)

Additionally, the Application.on_startup and Application.on_cleanup signals can be subscribed to for application component setup and tear down accordingly.

The following example will properly initialize and dispose an aiopg connection engine:

from aiopg.sa import create_engine

async def create_aiopg(app):
    app['pg_engine'] = await create_engine(
        user='postgre',
        database='postgre',
        host='localhost',
        port=5432,
        password=''
    )

async def dispose_aiopg(app):
    app['pg_engine'].close()
    await app['pg_engine'].wait_closed()

app.on_startup.append(create_aiopg)
app.on_cleanup.append(dispose_aiopg)

Signal handlers should not return a value but may modify incoming mutable parameters.

Signal handlers will be run sequentially, in order they were added. All handlers must be asynchronous since aiohttp 3.0.

Cleanup Context

Bare Application.on_startup / Application.on_cleanup pair still has a pitfall: signals handlers are independent on each other.

E.g. we have [create_pg, create_redis] in startup signal and [dispose_pg, dispose_redis] in cleanup.

If, for example, create_pg(app) call fails create_redis(app) is not called. But on application cleanup both dispose_pg(app) and dispose_redis(app) are still called: cleanup signal has no knowledge about startup/cleanup pairs and their execution state.

The solution is Application.cleanup_ctx usage:

async def pg_engine(app):
    app['pg_engine'] = await create_engine(
        user='postgre',
        database='postgre',
        host='localhost',
        port=5432,
        password=''
    )
    yield
    app['pg_engine'].close()
    await app['pg_engine'].wait_closed()

app.cleanup_ctx.append(pg_engine)

The attribute is a list of asynchronous generators, a code before yield is an initialization stage (called on startup), a code after yield is executed on cleanup. The generator must have only one yield.

aiohttp guarantees that cleanup code is called if and only if startup code was successfully finished.

Asynchronous generators are supported by Python 3.6+, on Python 3.5 please use async_generator library.

New in version 3.1.

Nested applications

Sub applications are designed for solving the problem of the big monolithic code base. Let’s assume we have a project with own business logic and tools like administration panel and debug toolbar.

Administration panel is a separate application by its own nature but all toolbar URLs are served by prefix like /admin.

Thus we’ll create a totally separate application named admin and connect it to main app with prefix by Application.add_subapp():

admin = web.Application()
# setup admin routes, signals and middlewares

app.add_subapp('/admin/', admin)

Middlewares and signals from app and admin are chained.

It means that if URL is '/admin/something' middlewares from app are applied first and admin.middlewares are the next in the call chain.

The same is going for Application.on_response_prepare signal – the signal is delivered to both top level app and admin if processing URL is routed to admin sub-application.

Common signals like Application.on_startup, Application.on_shutdown and Application.on_cleanup are delivered to all registered sub-applications. The passed parameter is sub-application instance, not top-level application.

Third level sub-applications can be nested into second level ones – there are no limitation for nesting level.

Url reversing for sub-applications should generate urls with proper prefix.

But for getting URL sub-application’s router should be used:

admin = web.Application()
admin.add_routes([web.get('/resource', handler, name='name')])

app.add_subapp('/admin/', admin)

url = admin.router['name'].url_for()

The generated url from example will have a value URL('/admin/resource').

If main application should do URL reversing for sub-application it could use the following explicit technique:

admin = web.Application()
admin.add_routes([web.get('/resource', handler, name='name')])

app.add_subapp('/admin/', admin)
app['admin'] = admin

async def handler(request):  # main application's handler
    admin = request.app['admin']
    url = admin.router['name'].url_for()
Expect Header

aiohttp.web supports Expect header. By default it sends HTTP/1.1 100 Continue line to client, or raises HTTPExpectationFailed if header value is not equal to “100-continue”. It is possible to specify custom Expect header handler on per route basis. This handler gets called if Expect header exist in request after receiving all headers and before processing application’s Middlewares and route handler. Handler can return None, in that case the request processing continues as usual. If handler returns an instance of class StreamResponse, request handler uses it as response. Also handler can raise a subclass of HTTPException. In this case all further processing will not happen and client will receive appropriate http response.

Note

A server that does not understand or is unable to comply with any of the expectation values in the Expect field of a request MUST respond with appropriate error status. The server MUST respond with a 417 (Expectation Failed) status if any of the expectations cannot be met or, if there are other problems with the request, some other 4xx status.

http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.20

If all checks pass, the custom handler must write a HTTP/1.1 100 Continue status code before returning.

The following example shows how to setup a custom handler for the Expect header:

async def check_auth(request):
    if request.version != aiohttp.HttpVersion11:
        return

    if request.headers.get('EXPECT') != '100-continue':
        raise HTTPExpectationFailed(text="Unknown Expect: %s" % expect)

    if request.headers.get('AUTHORIZATION') is None:
        raise HTTPForbidden()

    request.transport.write(b"HTTP/1.1 100 Continue\r\n\r\n")

async def hello(request):
    return web.Response(body=b"Hello, world")

app = web.Application()
app.add_routes([web.add_get('/', hello, expect_handler=check_auth)])
Custom resource implementation

To register custom resource use UrlDispatcher.register_resource(). Resource instance must implement AbstractResource interface.

Application runners

run_app() provides a simple blocking API for running an Application.

For starting the application asynchronously or serving on multiple HOST/PORT AppRunner exists.

The simple startup code for serving HTTP site on 'localhost', port 8080 looks like:

runner = web.AppRunner(app)
await runner.setup()
site = web.TCPSite(runner, 'localhost', 8080)
await site.start()

while True:
    await asyncio.sleep(3600)  # sleep forever

To stop serving call AppRunner.cleanup():

await runner.cleanup()

New in version 3.0.

Graceful shutdown

Stopping aiohttp web server by just closing all connections is not always satisfactory.

The problem is: if application supports websockets or data streaming it most likely has open connections at server shutdown time.

The library has no knowledge how to close them gracefully but developer can help by registering Application.on_shutdown signal handler and call the signal on web server closing.

Developer should keep a list of opened connections (Application is a good candidate).

The following websocket snippet shows an example for websocket handler:

from aiohttp import web
import weakref

app = web.Application()
app['websockets'] = weakref.WeakSet()

async def websocket_handler(request):
    ws = web.WebSocketResponse()
    await ws.prepare(request)

    request.app['websockets'].add(ws)
    try:
        async for msg in ws:
            ...
    finally:
        request.app['websockets'].discard(ws)

    return ws

Signal handler may look like:

from aiohttp import WSCloseCode

async def on_shutdown(app):
    for ws in set(app['websockets']):
        await ws.close(code=WSCloseCode.GOING_AWAY,
                       message='Server shutdown')

app.on_shutdown.append(on_shutdown)

Both run_app() and AppRunner.cleanup() call shutdown signal handlers.

Background tasks

Sometimes there’s a need to perform some asynchronous operations just after application start-up.

Even more, in some sophisticated systems there could be a need to run some background tasks in the event loop along with the application’s request handler. Such as listening to message queue or other network message/event sources (e.g. ZeroMQ, Redis Pub/Sub, AMQP, etc.) to react to received messages within the application.

For example the background task could listen to ZeroMQ on zmq.SUB socket, process and forward retrieved messages to clients connected via WebSocket that are stored somewhere in the application (e.g. in the application['websockets'] list).

To run such short and long running background tasks aiohttp provides an ability to register Application.on_startup signal handler(s) that will run along with the application’s request handler.

For example there’s a need to run one quick task and two long running tasks that will live till the application is alive. The appropriate background tasks could be registered as an Application.on_startup signal handlers as shown in the example below:

async def listen_to_redis(app):
    try:
        sub = await aioredis.create_redis(('localhost', 6379))
        ch, *_ = await sub.subscribe('news')
        async for msg in ch.iter(encoding='utf-8'):
            # Forward message to all connected websockets:
            for ws in app['websockets']:
                ws.send_str('{}: {}'.format(ch.name, msg))
    except asyncio.CancelledError:
        pass
    finally:
        await sub.unsubscribe(ch.name)
        await sub.quit()


async def start_background_tasks(app):
    app['redis_listener'] = asyncio.create_task(listen_to_redis(app))


async def cleanup_background_tasks(app):
    app['redis_listener'].cancel()
    await app['redis_listener']


app = web.Application()
app.on_startup.append(start_background_tasks)
app.on_cleanup.append(cleanup_background_tasks)
web.run_app(app)

The task listen_to_redis() will run forever. To shut it down correctly Application.on_cleanup signal handler may be used to send a cancellation to it.

Handling error pages

Pages like 404 Not Found and 500 Internal Error could be handled by custom middleware, see polls demo for example.

Deploying behind a Proxy

As discussed in Server Deployment the preferable way is deploying aiohttp web server behind a Reverse Proxy Server like nginx for production usage.

In this way properties like BaseRequest.scheme BaseRequest.host and BaseRequest.remote are incorrect.

Real values should be given from proxy server, usually either Forwarded or old-fashion X-Forwarded-For, X-Forwarded-Host, X-Forwarded-Proto HTTP headers are used.

aiohttp does not take forwarded headers into account by default because it produces security issue: HTTP client might add these headers too, pushing non-trusted data values.

That’s why aiohttp server should setup forwarded headers in custom middleware in tight conjunction with reverse proxy configuration.

For changing BaseRequest.scheme BaseRequest.host and BaseRequest.remote the middleware might use BaseRequest.clone().

See also

https://github.com/aio-libs/aiohttp-remotes provides secure helpers for modifying scheme, host and remote attributes according to Forwarded and X-Forwarded-* HTTP headers.

Swagger support

aiohttp-swagger is a library that allow to add Swagger documentation and embed the Swagger-UI into your aiohttp.web project.

CORS support

aiohttp.web itself does not support Cross-Origin Resource Sharing, but there is an aiohttp plugin for it: aiohttp_cors.

Debug Toolbar

aiohttp-debugtoolbar is a very useful library that provides a debugging toolbar while you’re developing an aiohttp.web application.

Install it with pip:

$ pip install aiohttp_debugtoolbar

Just call aiohttp_debugtoolbar.setup():

import aiohttp_debugtoolbar
from aiohttp_debugtoolbar import toolbar_middleware_factory

app = web.Application()
aiohttp_debugtoolbar.setup(app)

The toolbar is ready to use. Enjoy!!!

Dev Tools

aiohttp-devtools provides a couple of tools to simplify development of aiohttp.web applications.

Install with pip:

$ pip install aiohttp-devtools

  • runserver provides a development server with auto-reload, live-reload, static file serving and aiohttp-debugtoolbar integration.

  • start is a cookiecutter command which does the donkey work of creating new :mod:`aiohttp.web Applications.

Documentation and a complete tutorial of creating and running an app locally are available at aiohttp-devtools.

Low Level Server

This topic describes aiohttp.web based low level API.

Abstract

Sometimes user don’t need high-level concepts introduced in Server: applications, routers, middlewares and signals.

All what is needed is supporting asynchronous callable which accepts a request and returns a response object.

This is done by introducing aiohttp.web.Server class which serves a protocol factory role for asyncio.AbstractEventLoop.create_server() and bridges data stream to web handler and sends result back.

Low level web handler should accept the single BaseRequest parameter and performs one of the following actions:

  1. Return a Response with the whole HTTP body stored in memory.

  2. Create a StreamResponse, send headers by StreamResponse.prepare() call, send data chunks by StreamResponse.write() and return finished response.

  3. Raise HTTPException derived exception (see Exceptions section).

    All other exceptions not derived from HTTPException leads to 500 Internal Server Error response.

  4. Initiate and process Web-Socket connection by WebSocketResponse using (see WebSockets).

Run a Basic Low-Level Server

The following code demonstrates very trivial usage example:

import asyncio
from aiohttp import web


async def handler(request):
    return web.Response(text="OK")


async def main():
    server = web.Server(handler)
    runner = web.ServerRunner(server)
    await runner.setup()
    site = web.TCPSite(runner, 'localhost', 8080)
    await site.start()

    print("======= Serving on http://127.0.0.1:8080/ ======")

    # pause here for very long time by serving HTTP requests and
    # waiting for keyboard interruption
    await asyncio.sleep(100*3600)


loop = asyncio.get_event_loop()

try:
    loop.run_until_complete(main())
except KeyboardInterrupt:
    pass
loop.close()

In the snippet we have handler which returns a regular Response with "OK" in BODY.

This handler is processed by server (Server which acts as protocol factory). Network communication is created by runners API to serve http://127.0.0.1:8080/.

The handler should process every request for every path, e.g. GET, POST, Web-Socket.

The example is very basic: it always return 200 OK response, real life code is much more complex usually.

Server Reference

Request and Base Request

The Request object contains all the information about an incoming HTTP request.

BaseRequest is used for Low-Level Servers (which have no applications, routers, signals and middlewares). Request has an Request.app and Request.match_info attributes.

A BaseRequest / Request are dict like objects, allowing them to be used for sharing data among Middlewares and Signals handlers.

class aiohttp.web.BaseRequest
version

HTTP version of request, Read-only property.

Returns aiohttp.protocol.HttpVersion instance.

method

HTTP method, read-only property.

The value is upper-cased str like "GET", "POST", "PUT" etc.

url

A URL instance with absolute URL to resource (scheme, host and port are included).

Note

In case of malformed request (e.g. without "HOST" HTTP header) the absolute url may be unavailable.

rel_url

A URL instance with relative URL to resource (contains path, query and fragment parts only, scheme, host and port are excluded).

The property is equal to .url.relative() but is always present.

See also

A note from url.

scheme

A string representing the scheme of the request.

The scheme is 'https' if transport for request handling is SSL, 'http' otherwise.

The value could be overridden by clone().

Read-only str property.

Changed in version 2.3: Forwarded and X-Forwarded-Proto are not used anymore.

Call .clone(scheme=new_scheme) for setting up the value explicitly.

See also

Deploying behind a Proxy

secure

Shorthand for request.url.scheme == 'https'

Read-only bool property.

See also

scheme

forwarded

A tuple containing all parsed Forwarded header(s).

Makes an effort to parse Forwarded headers as specified by RFC 7239:

  • It adds one (immutable) dictionary per Forwarded field-value, i.e. per proxy. The element corresponds to the data in the Forwarded field-value added by the first proxy encountered by the client. Each subsequent item corresponds to those added by later proxies.

  • It checks that every value has valid syntax in general as specified in RFC 7239#section-4: either a token or a quoted-string.

  • It un-escapes quoted-pairs.

  • It does NOT validate ‘by’ and ‘for’ contents as specified in RFC 7239#section-6.

  • It does NOT validate host contents (Host ABNF).

  • It does NOT validate proto contents for valid URI scheme names.

Returns a tuple containing one or more MappingProxy objects

See also

scheme

See also

host

host

Host name of the request, resolved in this order:

  • Overridden value by clone() call.

  • Host HTTP header

  • socket.gtfqdn()

Read-only str property.

Changed in version 2.3: Forwarded and X-Forwarded-Host are not used anymore.

Call .clone(host=new_host) for setting up the value explicitly.

See also

Deploying behind a Proxy

remote

Originating IP address of a client initiated HTTP request.

The IP is resolved through the following headers, in this order:

  • Overridden value by clone() call.

  • Peer name of opened socket.

Read-only str property.

Call .clone(remote=new_remote) for setting up the value explicitly.

New in version 2.3.

See also

Deploying behind a Proxy

path_qs

The URL including PATH_INFO and the query string. e.g., /app/blog?id=10

Read-only str property.

path

The URL including PATH INFO without the host or scheme. e.g., /app/blog. The path is URL-decoded. For raw path info see raw_path.

Read-only str property.

raw_path

The URL including raw PATH INFO without the host or scheme. Warning, the path may be URL-encoded and may contain invalid URL characters, e.g. /my%2Fpath%7Cwith%21some%25strange%24characters.

For URL-decoded version please take a look on path.

Read-only str property.

query

A multidict with all the variables in the query string.

Read-only MultiDictProxy lazy property.

query_string

The query string in the URL, e.g., id=10

Read-only str property.

headers

A case-insensitive multidict proxy with all headers.

Read-only CIMultiDictProxy property.

raw_headers

HTTP headers of response as unconverted bytes, a sequence of (key, value) pairs.

keep_alive

True if keep-alive connection enabled by HTTP client and protocol version supports it, otherwise False.

Read-only bool property.

transport

A transport used to process request. Read-only property.

The property can be used, for example, for getting IP address of client’s peer:

peername = request.transport.get_extra_info('peername')
if peername is not None:
    host, port = peername
loop

An event loop instance used by HTTP request handling.

Read-only asyncio.AbstractEventLoop property.

Deprecated since version 3.5.

cookies

A multidict of all request’s cookies.

Read-only MultiDictProxy lazy property.

content

A StreamReader instance, input stream for reading request’s BODY.

Read-only property.

body_exists

Return True if request has HTTP BODY, False otherwise.

Read-only bool property.

New in version 2.3.

can_read_body

Return True if request’s HTTP BODY can be read, False otherwise.

Read-only bool property.

New in version 2.3.

has_body

Return True if request’s HTTP BODY can be read, False otherwise.

Read-only bool property.

Deprecated since version 2.3: Use can_read_body() instead.

content_type

Read-only property with content part of Content-Type header.

Returns str like 'text/html'

Note

Returns value is 'application/octet-stream' if no Content-Type header present in HTTP headers according to RFC 2616

charset

Read-only property that specifies the encoding for the request’s BODY.

The value is parsed from the Content-Type HTTP header.

Returns str like 'utf-8' or None if Content-Type has no charset information.

content_length

Read-only property that returns length of the request’s BODY.

The value is parsed from the Content-Length HTTP header.

Returns int or None if Content-Length is absent.

http_range

Read-only property that returns information about Range HTTP header.

Returns a slice where .start is left inclusive bound, .stop is right exclusive bound and .step is 1.

The property might be used in two manners:

  1. Attribute-access style (example assumes that both left and right borders are set, the real logic for case of open bounds is more complex):

    rng = request.http_range
with open(filename, 'rb') as f:
    f.seek(rng.start)
    return f.read(rng.stop-rng.start)

  2. Slice-style:

    return buffer[request.http_range]
if_modified_since

Read-only property that returns the date specified in the If-Modified-Since header.

Returns datetime.datetime or None if If-Modified-Since header is absent or is not a valid HTTP date.

if_unmodified_since

Read-only property that returns the date specified in the If-Unmodified-Since header.

Returns datetime.datetime or None if If-Unmodified-Since header is absent or is not a valid HTTP date.

New in version 3.1.

if_range

Read-only property that returns the date specified in the If-Range header.

Returns datetime.datetime or None if If-Range header is absent or is not a valid HTTP date.

New in version 3.1.

clone(*, method=..., rel_url=..., headers=...)

Clone itself with replacement some attributes.

Creates and returns a new instance of Request object. If no parameters are given, an exact copy is returned. If a parameter is not passed, it will reuse the one from the current request object.

Parameters

  • method (str) – http method

  • rel_url – url to use, str or URL

  • headers – CIMultiDict or compatible headers container.

Returns

a cloned Request instance.

get_extra_info(name, default=None)

Reads extra information from the protocol’s transport. If no value associated with name is found, default is returned.

Parameters

  • name (str) – The key to look up in the transport extra information.

  • default – Default value to be used when no value for name is found (default is None).

New in version 3.7.

coroutine read()

Read request body, returns bytes object with body content.

Note

The method does store read data internally, subsequent read() call will return the same value.

coroutine text()

Read request body, decode it using charset encoding or UTF-8 if no encoding was specified in MIME-type.

Returns str with body content.

Note

The method does store read data internally, subsequent text() call will return the same value.

coroutine json(*, loads=json.loads)

Read request body decoded as json.

The method is just a boilerplate coroutine implemented as:

async def json(self, *, loads=json.loads):
    body = await self.text()
    return loads(body)
Parameters

loads (callable) – any callable that accepts str and returns dict with parsed JSON (json.loads() by default).

Note

The method does store read data internally, subsequent json() call will return the same value.

coroutine multipart()

Returns aiohttp.multipart.MultipartReader which processes incoming multipart request.

The method is just a boilerplate coroutine implemented as:

async def multipart(self, *, reader=aiohttp.multipart.MultipartReader):
    return reader(self.headers, self._payload)

This method is a coroutine for consistency with the else reader methods.

Warning

The method does not store read data internally. That means once you exhausts multipart reader, you cannot get the request payload one more time.

See also

Working with Multipart

Changed in version 3.4: Dropped reader parameter.

coroutine post()

A coroutine that reads POST parameters from request body.

Returns MultiDictProxy instance filled with parsed data.

If method is not POST, PUT, PATCH, TRACE or DELETE or content_type is not empty or application/x-www-form-urlencoded or multipart/form-data returns empty multidict.

Note

The method does store read data internally, subsequent post() call will return the same value.

coroutine release()

Release request.

Eat unread part of HTTP BODY if present.

Note

User code may never call release(), all required work will be processed by aiohttp.web internal machinery.

class aiohttp.web.Request

A request used for receiving request’s information by web handler.

Every handler accepts a request instance as the first positional parameter.

The class in derived from BaseRequest, shares all parent’s attributes and methods but has a couple of additional properties:

match_info

Read-only property with AbstractMatchInfo instance for result of route resolving.

Note

Exact type of property depends on used router. If app.router is UrlDispatcher the property contains UrlMappingMatchInfo instance.

app

An Application instance used to call request handler, Read-only property.

config_dict

A aiohttp.ChainMapProxy instance for mapping all properties from the current application returned by app property and all its parents.

See also

Application’s config

New in version 3.2.

Note

You should never create the Request instance manually – aiohttp.web does it for you. But clone() may be used for cloning modified request copy with changed path, method etc.

Response classes

For now, aiohttp.web has three classes for the HTTP response: StreamResponse, Response and FileResponse.

Usually you need to use the second one. StreamResponse is intended for streaming data, while Response contains HTTP BODY as an attribute and sends own content as single piece with the correct Content-Length HTTP header.

For sake of design decisions Response is derived from StreamResponse parent class.

The response supports keep-alive handling out-of-the-box if request supports it.

You can disable keep-alive by force_close() though.

The common case for sending an answer from web-handler is returning a Response instance:

async def handler(request):
    return Response(text="All right!")

Response classes are dict like objects, allowing them to be used for sharing data among Middlewares and Signals handlers:

resp['key'] = value

New in version 3.0: Dict-like interface support.

StreamResponse
class aiohttp.web.StreamResponse(*, status=200, reason=None)

The base class for the HTTP response handling.

Contains methods for setting HTTP response headers, cookies, response status code, writing HTTP response BODY and so on.

The most important thing you should know about response — it is Finite State Machine.

That means you can do any manipulations with headers, cookies and status code only before prepare() coroutine is called.

Once you call prepare() any change of the HTTP header part will raise RuntimeError exception.

Any write() call after write_eof() is also forbidden.

Parameters

  • status (int) – HTTP status code, 200 by default.

  • reason (str) – HTTP reason. If param is None reason will be calculated basing on status parameter. Otherwise pass str with arbitrary status explanation..

prepared

Read-only bool property, True if prepare() has been called, False otherwise.

task

A task that serves HTTP request handling.

May be useful for graceful shutdown of long-running requests (streaming, long polling or web-socket).

status

Read-only property for HTTP response status code, int.

200 (OK) by default.

reason

Read-only property for HTTP response reason, str.

set_status(status, reason=None)

Set status and reason.

reason value is auto calculated if not specified (None).

keep_alive

Read-only property, copy of Request.keep_alive by default.

Can be switched to False by force_close() call.

force_close()

Disable keep_alive for connection. There are no ways to enable it back.

compression

Read-only bool property, True if compression is enabled.

False by default.

See also

enable_compression()

enable_compression(force=None)

Enable compression.

When force is unset compression encoding is selected based on the request’s Accept-Encoding header.

Accept-Encoding is not checked if force is set to a ContentCoding.

See also

compression

chunked

Read-only property, indicates if chunked encoding is on.

Can be enabled by enable_chunked_encoding() call.

See also

enable_chunked_encoding

enable_chunked_encoding()

Enables chunked encoding for response. There are no ways to disable it back. With enabled chunked encoding each write() operation encoded in separate chunk.

Warning

chunked encoding can be enabled for HTTP/1.1 only.

Setting up both content_length and chunked encoding is mutually exclusive.

See also

chunked

headers

CIMultiDict instance for outgoing HTTP headers.

cookies

An instance of http.cookies.SimpleCookie for outgoing cookies.

Warning

Direct setting up Set-Cookie header may be overwritten by explicit calls to cookie manipulation.

We are encourage using of cookies and set_cookie(), del_cookie() for cookie manipulations.

Convenient way for setting cookies, allows to specify some additional properties like max_age in a single call.

Parameters

  • name (str) – cookie name

  • value (str) – cookie value (will be converted to str if value has another type).

  • expires – expiration date (optional)

  • domain (str) – cookie domain (optional)

  • max_age (int) – defines the lifetime of the cookie, in seconds. The delta-seconds value is a decimal non- negative integer. After delta-seconds seconds elapse, the client should discard the cookie. A value of zero means the cookie should be discarded immediately. (optional)

  • path (str) – specifies the subset of URLs to which this cookie applies. (optional, '/' by default)

  • secure (bool) – attribute (with no value) directs the user agent to use only (unspecified) secure means to contact the origin server whenever it sends back this cookie. The user agent (possibly under the user’s control) may determine what level of security it considers appropriate for “secure” cookies. The secure should be considered security advice from the server to the user agent, indicating that it is in the session’s interest to protect the cookie contents. (optional)

  • httponly (bool) – True if the cookie HTTP only (optional)

  • version (int) – a decimal integer, identifies to which version of the state management specification the cookie conforms. (Optional, version=1 by default)

  • samesite (str) –

    Asserts that a cookie must not be sent with cross-origin requests, providing some protection against cross-site request forgery attacks. Generally the value should be one of: None, Lax or Strict. (optional)

    New in version 3.7.

Warning

In HTTP version 1.1, expires was deprecated and replaced with the easier-to-use max-age, but Internet Explorer (IE6, IE7, and IE8) does not support max-age.

Deletes cookie.

Parameters

  • name (str) – cookie name

  • domain (str) – optional cookie domain

  • path (str) – optional cookie path, '/' by default

content_length

Content-Length for outgoing response.

content_type

Content part of Content-Type for outgoing response.

charset

Charset aka encoding part of Content-Type for outgoing response.

The value converted to lower-case on attribute assigning.

last_modified

Last-Modified header for outgoing response.

This property accepts raw str values, datetime.datetime objects, Unix timestamps specified as an int or a float object, and the value None to unset the header.

coroutine prepare(request)
Parameters

request (aiohttp.web.Request) – HTTP request object, that the response answers.

Send HTTP header. You should not change any header data after calling this method.

The coroutine calls on_response_prepare signal handlers after default headers have been computed and directly before headers are sent.

coroutine write(data)

Send byte-ish data as the part of response BODY:

await resp.write(data)

prepare() must be invoked before the call.

Raises TypeError if data is not bytes, bytearray or memoryview instance.

Raises RuntimeError if prepare() has not been called.

Raises RuntimeError if write_eof() has been called.

coroutine write_eof()

A coroutine may be called as a mark of the HTTP response processing finish.

Internal machinery will call this method at the end of the request processing if needed.

After write_eof() call any manipulations with the response object are forbidden.

Response
class aiohttp.web.Response(*, body=None, status=200, reason=None, text=None, headers=None, content_type=None, charset=None, zlib_executor_size=sentinel, zlib_executor=None)

The most usable response class, inherited from StreamResponse.

Accepts body argument for setting the HTTP response BODY.

The actual body sending happens in overridden write_eof().

Parameters

  • body (bytes) – response’s BODY

  • status (int) – HTTP status code, 200 OK by default.

  • headers (collections.abc.Mapping) – HTTP headers that should be added to response’s ones.

  • text (str) – response’s BODY

  • content_type (str) – response’s content type. 'text/plain' if text is passed also, 'application/octet-stream' otherwise.

  • charset (str) – response’s charset. 'utf-8' if text is passed also, None otherwise.

  • zlib_executor_size (int) –

    length in bytes which will trigger zlib compression

    of body to happen in an executor

    New in version 3.5.

  • zlib_executor (int) –

    executor to use for zlib compression

    New in version 3.5.

body

Read-write attribute for storing response’s content aka BODY, bytes.

Setting body also recalculates content_length value.

Assigning str to body will make the body type of aiohttp.payload.StringPayload, which tries to encode the given data based on Content-Type HTTP header, while defaulting to UTF-8.

Resetting body (assigning None) sets content_length to None too, dropping Content-Length HTTP header.

text

Read-write attribute for storing response’s content, represented as string, str.

Setting text also recalculates content_length value and body value

Resetting text (assigning None) sets content_length to None too, dropping Content-Length HTTP header.

WebSocketResponse
class aiohttp.web.WebSocketResponse(*, timeout=10.0, receive_timeout=None, autoclose=True, autoping=True, heartbeat=None, protocols=(), compress=True, max_msg_size=4194304)

Class for handling server-side websockets, inherited from StreamResponse.

After starting (by prepare() call) the response you cannot use write() method but should to communicate with websocket client by send_str(), receive() and others.

To enable back-pressure from slow websocket clients treat methods ping(), pong(), send_str(), send_bytes(), send_json() as coroutines. By default write buffer size is set to 64k.

Parameters

  • autoping (bool) – Automatically send PONG on PING message from client, and handle PONG responses from client. Note that server does not send PING requests, you need to do this explicitly using ping() method.

  • heartbeat (float) – Send ping message every heartbeat seconds and wait pong response, close connection if pong response is not received. The timer is reset on any data reception.

  • receive_timeout (float) – Timeout value for receive operations. Default value is None (no timeout for receive operation)

  • compress (bool) – Enable per-message deflate extension support. False for disabled, default value is True.

  • max_msg_size (int) –

    maximum size of read websocket message, 4

    MB by default. To disable the size limit use 0.

    New in version 3.3.

The class supports async for statement for iterating over incoming messages:

ws = web.WebSocketResponse()
await ws.prepare(request)

    async for msg in ws:
        print(msg.data)
coroutine prepare(request)

Starts websocket. After the call you can use websocket methods.

Parameters

request (aiohttp.web.Request) – HTTP request object, that the response answers.

Raises

HTTPException – if websocket handshake has failed.

can_prepare(request)

Performs checks for request data to figure out if websocket can be started on the request.

If can_prepare() call is success then prepare() will success too.

Parameters

request (aiohttp.web.Request) – HTTP request object, that the response answers.

Returns

WebSocketReady instance.

WebSocketReady.ok is True on success, WebSocketReady.protocol is websocket subprotocol which is passed by client and accepted by server (one of protocols sequence from WebSocketResponse ctor). WebSocketReady.protocol may be None if client and server subprotocols are not overlapping.

Note

The method never raises exception.

closed

Read-only property, True if connection has been closed or in process of closing. CLOSE message has been received from peer.

close_code

Read-only property, close code from peer. It is set to None on opened connection.

ws_protocol

Websocket subprotocol chosen after start() call.

May be None if server and client protocols are not overlapping.

exception()

Returns last occurred exception or None.

coroutine ping(message=b'')

Send PING to peer.

Parameters

message – optional payload of ping message, str (converted to UTF-8 encoded bytes) or bytes.

Raises

RuntimeError – if connections is not started or closing.

Changed in version 3.0: The method is converted into coroutine

coroutine pong(message=b'')

Send unsolicited PONG to peer.

Parameters

message – optional payload of pong message, str (converted to UTF-8 encoded bytes) or bytes.

Raises

RuntimeError – if connections is not started or closing.

Changed in version 3.0: The method is converted into coroutine

coroutine send_str(data, compress=None)

Send data to peer as TEXT message.

Parameters

  • data (str) – data to send.

  • compress (int) – sets specific level of compression for single message, None for not overriding per-socket setting.

Raises

Changed in version 3.0: The method is converted into coroutine, compress parameter added.

coroutine send_bytes(data, compress=None)

Send data to peer as BINARY message.

Parameters

  • data – data to send.

  • compress (int) – sets specific level of compression for single message, None for not overriding per-socket setting.

Raises

Changed in version 3.0: The method is converted into coroutine, compress parameter added.

coroutine send_json(data, compress=None, *, dumps=json.dumps)

Send data to peer as JSON string.

Parameters

  • data – data to send.

  • compress (int) – sets specific level of compression for single message, None for not overriding per-socket setting.

  • dumps (callable) – any callable that accepts an object and returns a JSON string (json.dumps() by default).

Raises

Changed in version 3.0: The method is converted into coroutine, compress parameter added.

coroutine close(*, code=1000, message=b'')

A coroutine that initiates closing handshake by sending CLOSE message.

It is safe to call close() from different task.

Parameters

  • code (int) – closing code

  • message – optional payload of close message, str (converted to UTF-8 encoded bytes) or bytes.

Raises

RuntimeError – if connection is not started

coroutine receive(timeout=None)

A coroutine that waits upcoming data message from peer and returns it.

The coroutine implicitly handles PING, PONG and CLOSE without returning the message.

It process ping-pong game and performs closing handshake internally.

Note

Can only be called by the request handling task.

Parameters

timeout –

timeout for receive operation.

timeout value overrides response`s receive_timeout attribute.

Returns

WSMessage

Raises

RuntimeError – if connection is not started

coroutine receive_str(*, timeout=None)

A coroutine that calls receive() but also asserts the message type is TEXT.

Note

Can only be called by the request handling task.

Parameters

timeout –

timeout for receive operation.

timeout value overrides response`s receive_timeout attribute.

Return str

peer’s message content.

Raises

TypeError – if message is BINARY.

coroutine receive_bytes(*, timeout=None)

A coroutine that calls receive() but also asserts the message type is BINARY.

Note

Can only be called by the request handling task.

Parameters

timeout –

timeout for receive operation.

timeout value overrides response`s receive_timeout attribute.

Return bytes

peer’s message content.

Raises

TypeError – if message is TEXT.

coroutine receive_json(*, loads=json.loads, timeout=None)

A coroutine that calls receive_str() and loads the JSON string to a Python dict.

Note

Can only be called by the request handling task.

Parameters

  • loads (callable) – any callable that accepts str and returns dict with parsed JSON (json.loads() by default).

  • timeout –

    timeout for receive operation.

    timeout value overrides response`s receive_timeout attribute.

Return dict

loaded JSON content

Raises

See also

WebSockets handling

WebSocketReady
class aiohttp.web.WebSocketReady

A named tuple for returning result from WebSocketResponse.can_prepare().

Has bool check implemented, e.g.:

if not await ws.can_prepare(...):
    cannot_start_websocket()
ok

True if websocket connection can be established, False otherwise.

protocol

str represented selected websocket sub-protocol.

See also

WebSocketResponse.can_prepare()

json_response
aiohttp.web.json_response([data, ]*, text=None, body=None, status=200, reason=None, headers=None, content_type='application/json', dumps=json.dumps)

Return Response with predefined 'application/json' content type and data encoded by dumps parameter (json.dumps() by default).

HTTP Exceptions

Errors can also be returned by raising a HTTP exception instance from within the handler.

class aiohttp.web.HTTPException(*, headers=None, reason=None, text=None, content_type=None)

Low-level HTTP failure.

Parameters

  • headers (dict or multidict.CIMultiDict) – headers for the response

  • reason (str) – reason included in the response

  • text (str) – response’s body

  • content_type (str) – response’s content type. This is passed through to the Response initializer.

Sub-classes of HTTPException exist for the standard HTTP response codes as described in Exceptions and the expected usage is to simply raise the appropriate exception type to respond with a specific HTTP response code.

Since HTTPException is a sub-class of Response, it contains the methods and properties that allow you to directly manipulate details of the response.

status_code

HTTP status code for this exception class. This attribute is usually defined at the class level. self.status_code is passed to the Response initializer.

Application and Router
Application

Application is a synonym for web-server.

To get fully working example, you have to make application, register supported urls in router and pass it to aiohttp.web.run_app() or aiohttp.web.AppRunner.

Application contains a router instance and a list of callbacks that will be called during application finishing.

Application is a dict-like object, so you can use it for sharing data globally by storing arbitrary properties for later access from a handler via the Request.app property:

app = Application()
app['database'] = await aiopg.create_engine(**db_config)

async def handler(request):
    with (await request.app['database']) as conn:
        conn.execute("DELETE * FROM table")

Although Application is a dict-like object, it can’t be duplicated like one using Application.copy().

class aiohttp.web.Application(*, logger=<default>, router=None, middlewares=(), handler_args=None, client_max_size=1024**2, loop=None, debug=...)

The class inherits dict.

Parameters

  • logger –

    logging.Logger instance for storing application logs.

    By default the value is logging.getLogger("aiohttp.web")

  • router –

    aiohttp.abc.AbstractRouter instance, the system

    creates UrlDispatcher by default if router is None.

    Deprecated since version 3.3: The custom routers support is deprecated, the parameter will be removed in 4.0.

  • middlewares – list of middleware factories, see Middlewares for details.

  • handler_args – dict-like object that overrides keyword arguments of Application.make_handler()

  • client_max_size – client’s maximum size in a request, in bytes. If a POST request exceeds this value, it raises an HTTPRequestEntityTooLarge exception.

  • loop –

    event loop

    Deprecated since version 2.0: The parameter is deprecated. Loop is get set during freeze stage.

  • debug –

    Switches debug mode.

    Deprecated since version 3.5: Use asyncio Debug Mode instead.

router

Read-only property that returns router instance.

logger

logging.Logger instance for storing application logs.

loop

event loop used for processing HTTP requests.

Deprecated since version 3.5.

debug

Boolean value indicating whether the debug mode is turned on or off.

Deprecated since version 3.5: Use asyncio Debug Mode instead.

on_response_prepare

A Signal that is fired near the end of StreamResponse.prepare() with parameters request and response. It can be used, for example, to add custom headers to each response, or to modify the default headers computed by the application, directly before sending the headers to the client.

Signal handlers should have the following signature:

async def on_prepare(request, response):
    pass
on_startup

A Signal that is fired on application start-up.

Subscribers may use the signal to run background tasks in the event loop along with the application’s request handler just after the application start-up.

Signal handlers should have the following signature:

async def on_startup(app):
    pass

See also

Signals.

on_shutdown

A Signal that is fired on application shutdown.

Subscribers may use the signal for gracefully closing long running connections, e.g. websockets and data streaming.

Signal handlers should have the following signature:

async def on_shutdown(app):
    pass

It’s up to end user to figure out which web-handlers are still alive and how to finish them properly.

We suggest keeping a list of long running handlers in Application dictionary.

See also

Graceful shutdown and on_cleanup.

on_cleanup

A Signal that is fired on application cleanup.

Subscribers may use the signal for gracefully closing connections to database server etc.

Signal handlers should have the following signature:

async def on_cleanup(app):
    pass

See also

Signals and on_shutdown.

cleanup_ctx

A list of context generators for startup/cleanup handling.

Signal handlers should have the following signature:

async def context(app):
    # do startup stuff
    yield
    # do cleanup

New in version 3.1.

See also

Cleanup Context.

add_subapp(prefix, subapp)

Register nested sub-application under given path prefix.

In resolving process if request’s path starts with prefix then further resolving is passed to subapp.

Parameters

  • prefix (str) – path’s prefix for the resource.

  • subapp (Application) – nested application attached under prefix.

Returns

a PrefixedSubAppResource instance.

add_domain(domain, subapp)

Register nested sub-application that serves the domain name or domain name mask.

In resolving process if request.headers[‘host’] matches the pattern domain then further resolving is passed to subapp.

Parameters

  • domain (str) – domain or mask of domain for the resource.

  • subapp (Application) – nested application.

Returns

a MatchedSubAppResource instance.

add_routes(routes_table)

Register route definitions from routes_table.

The table is a list of RouteDef items or RouteTableDef.

Returns

list of registered AbstractRoute instances.

The method is a shortcut for app.router.add_routes(routes_table), see also UrlDispatcher.add_routes().

New in version 3.1.

Changed in version 3.7: Return value updated from None to list of AbstractRoute instances.

make_handler(loop=None, **kwargs)

Creates HTTP protocol factory for handling requests.

Parameters

  • loop –

    event loop used for processing HTTP requests.

    If param is None asyncio.get_event_loop() used for getting default event loop.

    Deprecated since version 2.0.

  • tcp_keepalive (bool) – Enable TCP Keep-Alive. Default: True.

  • keepalive_timeout (int) – Number of seconds before closing Keep-Alive connection. Default: 75 seconds (NGINX’s default value).

  • logger – Custom logger object. Default: aiohttp.log.server_logger.

  • access_log – Custom logging object. Default: aiohttp.log.access_logger.

  • access_log_class – Class for access_logger. Default: aiohttp.helpers.AccessLogger. Must to be a subclass of aiohttp.abc.AbstractAccessLogger.

  • access_log_format (str) – Access log format string. Default: helpers.AccessLogger.LOG_FORMAT.

  • max_line_size (int) – Optional maximum header line size. Default: 8190.

  • max_headers (int) – Optional maximum header size. Default: 32768.

  • max_field_size (int) – Optional maximum header field size. Default: 8190.

  • lingering_time (float) – Maximum time during which the server reads and ignores additional data coming from the client when lingering close is on. Use 0 to disable lingering on server channel closing.

You should pass result of the method as protocol_factory to create_server(), e.g.:

loop = asyncio.get_event_loop()

app = Application()

# setup route table
# app.router.add_route(...)

await loop.create_server(app.make_handler(),
                         '0.0.0.0', 8080)

Deprecated since version 3.2: The method is deprecated and will be removed in future aiohttp versions. Please use Application runners instead.

coroutine startup()

A coroutine that will be called along with the application’s request handler.

The purpose of the method is calling on_startup signal handlers.

coroutine shutdown()

A coroutine that should be called on server stopping but before cleanup().

The purpose of the method is calling on_shutdown signal handlers.

coroutine cleanup()

A coroutine that should be called on server stopping but after shutdown().

The purpose of the method is calling on_cleanup signal handlers.

Note

Application object has router attribute but has no add_route() method. The reason is: we want to support different router implementations (even maybe not url-matching based but traversal ones).

For sake of that fact we have very trivial ABC for AbstractRouter: it should have only AbstractRouter.resolve() coroutine.

No methods for adding routes or route reversing (getting URL by route name). All those are router implementation details (but, sure, you need to deal with that methods after choosing the router for your application).

Server

A protocol factory compatible with create_server().

class aiohttp.web.Server

The class is responsible for creating HTTP protocol objects that can handle HTTP connections.

connections

List of all currently opened connections.

requests_count

Amount of processed requests.

coroutine shutdown(timeout)

A coroutine that should be called to close all opened connections.

Router

For dispatching URLs to handlers aiohttp.web uses routers.

Router is any object that implements AbstractRouter interface.

aiohttp.web provides an implementation called UrlDispatcher.

Application uses UrlDispatcher as router() by default.

class aiohttp.web.UrlDispatcher

Straightforward url-matching router, implements collections.abc.Mapping for access to named routes.

Before running Application you should fill route table first by calling add_route() and add_static().

Handler lookup is performed by iterating on added routes in FIFO order. The first matching route will be used to call corresponding handler.

If on route creation you specify name parameter the result is named route.

Named route can be retrieved by app.router[name] call, checked for existence by name in app.router etc.

See also

Route classes

add_resource(path, *, name=None)

Append a resource to the end of route table.

path may be either constant string like '/a/b/c' or variable rule like '/a/{var}' (see handling variable paths)

Parameters

  • path (str) – resource path spec.

  • name (str) – optional resource name.

Returns

created resource instance (PlainResource or DynamicResource).

add_route(method, path, handler, *, name=None, expect_handler=None)

Append handler to the end of route table.

path may be either constant string like '/a/b/c' or

variable rule like '/a/{var}' (see handling variable paths)

Pay attention please: handler is converted to coroutine internally when it is a regular function.

Parameters

  • method (str) –

    HTTP method for route. Should be one of 'GET', 'POST', 'PUT', 'DELETE', 'PATCH', 'HEAD', 'OPTIONS' or '*' for any method.

    The parameter is case-insensitive, e.g. you can push 'get' as well as 'GET'.

  • path (str) – route path. Should be started with slash ('/').

  • handler (callable) – route handler.

  • name (str) – optional route name.

  • expect_handler (coroutine) – optional expect header handler.

Returns

new PlainRoute or DynamicRoute instance.

add_routes(routes_table)

Register route definitions from routes_table.

The table is a list of RouteDef items or RouteTableDef.

Returns

list of registered AbstractRoute instances.

New in version 2.3.

Changed in version 3.7: Return value updated from None to list of AbstractRoute instances.

add_get(path, handler, *, name=None, allow_head=True, **kwargs)

Shortcut for adding a GET handler. Calls the add_route() with method equals to 'GET'.

If allow_head is True (default) the route for method HEAD is added with the same handler as for GET.

If name is provided the name for HEAD route is suffixed with '-head'. For example router.add_get(path, handler, name='route') call adds two routes: first for GET with name 'route' and second for HEAD with name 'route-head'.

add_post(path, handler, **kwargs)

Shortcut for adding a POST handler. Calls the add_route() with

method equals to 'POST'.

add_head(path, handler, **kwargs)

Shortcut for adding a HEAD handler. Calls the add_route() with method equals to 'HEAD'.

add_put(path, handler, **kwargs)

Shortcut for adding a PUT handler. Calls the add_route() with method equals to 'PUT'.

add_patch(path, handler, **kwargs)

Shortcut for adding a PATCH handler. Calls the add_route() with method equals to 'PATCH'.

add_delete(path, handler, **kwargs)

Shortcut for adding a DELETE handler. Calls the add_route() with method equals to 'DELETE'.

add_view(path, handler, **kwargs)

Shortcut for adding a class-based view handler. Calls the add_route() with method equals to '*'.

New in version 3.0.

add_static(prefix, path, *, name=None, expect_handler=None, chunk_size=256 * 1024, response_factory=StreamResponse, show_index=False, follow_symlinks=False, append_version=False)

Adds a router and a handler for returning static files.

Useful for serving static content like images, javascript and css files.

On platforms that support it, the handler will transfer files more efficiently using the sendfile system call.

In some situations it might be necessary to avoid using the sendfile system call even if the platform supports it. This can be accomplished by by setting environment variable AIOHTTP_NOSENDFILE=1.

If a gzip version of the static content exists at file path + .gz, it will be used for the response.

Warning

Use add_static() for development only. In production, static content should be processed by web servers like nginx or apache.

Parameters

  • prefix (str) – URL path prefix for handled static files

  • path – path to the folder in file system that contains handled static files, str or pathlib.Path.

  • name (str) – optional route name.

  • expect_handler (coroutine) – optional expect header handler.

  • chunk_size (int) –

    size of single chunk for file downloading, 256Kb by default.

    Increasing chunk_size parameter to, say, 1Mb may increase file downloading speed but consumes more memory.

  • show_index (bool) – flag for allowing to show indexes of a directory, by default it’s not allowed and HTTP/403 will be returned on directory access.

  • follow_symlinks (bool) – flag for allowing to follow symlinks from a directory, by default it’s not allowed and HTTP/404 will be returned on access.

  • append_version (bool) – flag for adding file version (hash) to the url query string, this value will be used as default when you call to StaticRoute.url() and StaticRoute.url_for() methods.

Returns

new StaticRoute instance.

coroutine resolve(request)

A coroutine that returns AbstractMatchInfo for request.

The method never raises exception, but returns AbstractMatchInfo instance with:

  1. http_exception assigned to HTTPException instance.

  2. handler which raises HTTPNotFound or HTTPMethodNotAllowed on handler’s execution if there is no registered route for request.

    Middlewares can process that exceptions to render pretty-looking error page for example.

Used by internal machinery, end user unlikely need to call the method.

Note

The method uses Request.raw_path for pattern matching against registered routes.

resources()

The method returns a view for all registered resources.

The view is an object that allows to:

  1. Get size of the router table:

    len(app.router.resources())

  2. Iterate over registered resources:

    for resource in app.router.resources():
    print(resource)

  3. Make a check if the resources is registered in the router table:

    route in app.router.resources()
routes()

The method returns a view for all registered routes.

named_resources()

Returns a dict-like types.MappingProxyType view over all named resources.

The view maps every named resource’s name to the BaseResource instance. It supports the usual dict-like operations, except for any mutable operations (i.e. it’s read-only):

len(app.router.named_resources())

for name, resource in app.router.named_resources().items():
    print(name, resource)

"name" in app.router.named_resources()

app.router.named_resources()["name"]
Resource

Default router UrlDispatcher operates with resources.

Resource is an item in routing table which has a path, an optional unique name and at least one route.

web-handler lookup is performed in the following way:

  1. Router iterates over resources one-by-one.

  2. If resource matches to requested URL the resource iterates over own routes.

  3. If route matches to requested HTTP method (or '*' wildcard) the route’s handler is used as found web-handler. The lookup is finished.

  4. Otherwise router tries next resource from the routing table.

  5. If the end of routing table is reached and no resource / route pair found the router returns special AbstractMatchInfo instance with AbstractMatchInfo.http_exception is not None but HTTPException with either HTTP 404 Not Found or HTTP 405 Method Not Allowed status code. Registered AbstractMatchInfo.handler raises this exception on call.

User should never instantiate resource classes but give it by UrlDispatcher.add_resource() call.

After that he may add a route by calling Resource.add_route().

UrlDispatcher.add_route() is just shortcut for:

router.add_resource(path).add_route(method, handler)

Resource with a name is called named resource. The main purpose of named resource is constructing URL by route name for passing it into template engine for example:

url = app.router['resource_name'].url_for().with_query({'a': 1, 'b': 2})

Resource classes hierarchy:

AbstractResource
  Resource
    PlainResource
    DynamicResource
    StaticResource
class aiohttp.web.AbstractResource

A base class for all resources.

Inherited from collections.abc.Sized and collections.abc.Iterable.

len(resource) returns amount of routes belongs to the resource, for route in resource allows to iterate over these routes.

name

Read-only name of resource or None.

canonical

Read-only canonical path associate with the resource. For example /path/to or /path/{to}

New in version 3.3.

coroutine resolve(request)

Resolve resource by finding appropriate web-handler for (method, path) combination.

Returns

(match_info, allowed_methods) pair.

allowed_methods is a set or HTTP methods accepted by resource.

match_info is either UrlMappingMatchInfo if request is resolved or None if no route is found.

get_info()

A resource description, e.g. {'path': '/path/to'} or {'formatter': '/path/{to}', 'pattern': re.compile(r'^/path/(?P<to>[a-zA-Z][_a-zA-Z0-9]+)$

url_for(*args, **kwargs)

Construct an URL for route with additional params.

args and kwargs depend on a parameters list accepted by inherited resource class.

Returns

URL – resulting URL instance.

class aiohttp.web.Resource

A base class for new-style resources, inherits AbstractResource.

add_route(method, handler, *, expect_handler=None)

Add a web-handler to resource.

Parameters

  • method (str) –

    HTTP method for route. Should be one of 'GET', 'POST', 'PUT', 'DELETE', 'PATCH', 'HEAD', 'OPTIONS' or '*' for any method.

    The parameter is case-insensitive, e.g. you can push 'get' as well as 'GET'.

    The method should be unique for resource.

  • handler (callable) – route handler.

  • expect_handler (coroutine) – optional expect header handler.

Returns

new ResourceRoute instance.

class aiohttp.web.PlainResource

A resource, inherited from Resource.

The class corresponds to resources with plain-text matching, '/path/to' for example.

canonical

Read-only canonical path associate with the resource. Returns the path used to create the PlainResource. For example /path/to

New in version 3.3.

url_for()

Returns a URL for the resource.

class aiohttp.web.DynamicResource

A resource, inherited from Resource.

The class corresponds to resources with variable matching, e.g. '/path/{to}/{param}' etc.

canonical

Read-only canonical path associate with the resource. Returns the formatter obtained from the path used to create the DynamicResource. For example, from a path /get/{num:^\d+}, it returns /get/{num}

New in version 3.3.

url_for(**params)

Returns a URL for the resource.

Parameters

params –

– a variable substitutions for dynamic resource.

E.g. for '/path/{to}/{param}' pattern the method should be called as resource.url_for(to='val1', param='val2')

class aiohttp.web.StaticResource

A resource, inherited from Resource.

The class corresponds to resources for static file serving.

canonical

Read-only canonical path associate with the resource. Returns the prefix used to create the StaticResource. For example /prefix

New in version 3.3.

url_for(filename, append_version=None)

Returns a URL for file path under resource prefix.

Parameters

  • filename –

    – a file name substitution for static file handler.

    Accepts both str and pathlib.Path.

    E.g. an URL for '/prefix/dir/file.txt' should be generated as resource.url_for(filename='dir/file.txt')

  • append_version (bool) –

    – a flag for adding file version

    (hash) to the url query string for cache boosting

    By default has value from a constructor (False by default) When set to True - v=FILE_HASH query string param will be added When set to False has no impact

    if file not found has no impact

class aiohttp.web.PrefixedSubAppResource

A resource for serving nested applications. The class instance is returned by add_subapp call.

canonical

Read-only canonical path associate with the resource. Returns the prefix used to create the PrefixedSubAppResource. For example /prefix

New in version 3.3.

url_for(**kwargs)

The call is not allowed, it raises RuntimeError.

Route

Route has HTTP method (wildcard '*' is an option), web-handler and optional expect handler.

Every route belong to some resource.

Route classes hierarchy:

AbstractRoute
  ResourceRoute
  SystemRoute

ResourceRoute is the route used for resources, SystemRoute serves URL resolving errors like 404 Not Found and 405 Method Not Allowed.

class aiohttp.web.AbstractRoute

Base class for routes served by UrlDispatcher.

method

HTTP method handled by the route, e.g. GET, POST etc.

handler

handler that processes the route.

name

Name of the route, always equals to name of resource which owns the route.

resource

Resource instance which holds the route, None for SystemRoute.

url_for(*args, **kwargs)

Abstract method for constructing url handled by the route.

Actually it’s a shortcut for route.resource.url_for(...).

coroutine handle_expect_header(request)

100-continue handler.

class aiohttp.web.ResourceRoute

The route class for handling different HTTP methods for Resource.

class aiohttp.web.SystemRoute

The route class for handling URL resolution errors like like 404 Not Found and 405 Method Not Allowed.

status

HTTP status code

reason

HTTP status reason

RouteDef and StaticDef

Route definition, a description for not registered yet route.

Could be used for filing route table by providing a list of route definitions (Django style).

The definition is created by functions like get() or post(), list of definitions could be added to router by UrlDispatcher.add_routes() call:

from aiohttp import web

async def handle_get(request):
    ...


async def handle_post(request):
    ...

app.router.add_routes([web.get('/get', handle_get),
                       web.post('/post', handle_post),
class aiohttp.web.AbstractRouteDef

A base class for route definitions.

Inherited from abc.ABC.

New in version 3.1.

register(router)

Register itself into UrlDispatcher.

Abstract method, should be overridden by subclasses.

Returns

list of registered AbstractRoute objects.

Changed in version 3.7: Return value updated from None to list of AbstractRoute instances.

class aiohttp.web.RouteDef

A definition of not registered yet route.

Implements AbstractRouteDef.

New in version 2.3.

Changed in version 3.1: The class implements AbstractRouteDef interface.

method

HTTP method (GET, POST etc.) (str).

path

Path to resource, e.g. /path/to. Could contain {} brackets for variable resources (str).

handler

An async function to handle HTTP request.

kwargs

A dict of additional arguments.

class aiohttp.web.StaticDef

A definition of static file resource.

Implements AbstractRouteDef.

New in version 3.1.

prefix

A prefix used for static file handling, e.g. /static.

path

File system directory to serve, str or pathlib.Path (e.g. '/home/web-service/path/to/static'.

kwargs

A dict of additional arguments, see UrlDispatcher.add_static() for a list of supported options.

aiohttp.web.get(path, handler, *, name=None, allow_head=True, expect_handler=None)

Return RouteDef for processing GET requests. See UrlDispatcher.add_get() for information about parameters.

New in version 2.3.

aiohttp.web.post(path, handler, *, name=None, expect_handler=None)

Return RouteDef for processing POST requests. See UrlDispatcher.add_post() for information about parameters.

New in version 2.3.

aiohttp.web.head(path, handler, *, name=None, expect_handler=None)

Return RouteDef for processing HEAD requests. See UrlDispatcher.add_head() for information about parameters.

New in version 2.3.

aiohttp.web.put(path, handler, *, name=None, expect_handler=None)

Return RouteDef for processing PUT requests. See UrlDispatcher.add_put() for information about parameters.

New in version 2.3.

aiohttp.web.patch(path, handler, *, name=None, expect_handler=None)

Return RouteDef for processing PATCH requests. See UrlDispatcher.add_patch() for information about parameters.

New in version 2.3.

aiohttp.web.delete(path, handler, *, name=None, expect_handler=None)

Return RouteDef for processing DELETE requests. See UrlDispatcher.add_delete() for information about parameters.

New in version 2.3.

aiohttp.web.view(path, handler, *, name=None, expect_handler=None)

Return RouteDef for processing ANY requests. See UrlDispatcher.add_view() for information about parameters.

New in version 3.0.

aiohttp.web.static(prefix, path, *, name=None, expect_handler=None, chunk_size=256 * 1024, show_index=False, follow_symlinks=False, append_version=False)

Return StaticDef for processing static files.

See UrlDispatcher.add_static() for information about supported parameters.

New in version 3.1.

aiohttp.web.route(method, path, handler, *, name=None, expect_handler=None)

Return RouteDef for processing requests that decided by method. See UrlDispatcher.add_route() for information about parameters.

New in version 2.3.

RouteTableDef

A routes table definition used for describing routes by decorators (Flask style):

from aiohttp import web

routes = web.RouteTableDef()

@routes.get('/get')
async def handle_get(request):
    ...


@routes.post('/post')
async def handle_post(request):
    ...

app.router.add_routes(routes)


@routes.view("/view")
class MyView(web.View):
    async def get(self):
        ...

    async def post(self):
        ...
class aiohttp.web.RouteTableDef

A sequence of RouteDef instances (implements abc.collections.Sequence protocol).

In addition to all standard list methods the class provides also methods like get() and post() for adding new route definition.

New in version 2.3.

@get(path, *, allow_head=True, name=None, expect_handler=None)

Add a new RouteDef item for registering GET web-handler.

See UrlDispatcher.add_get() for information about parameters.

@post(path, *, name=None, expect_handler=None)

Add a new RouteDef item for registering POST web-handler.

See UrlDispatcher.add_post() for information about parameters.

@head(path, *, name=None, expect_handler=None)

Add a new RouteDef item for registering HEAD web-handler.

See UrlDispatcher.add_head() for information about parameters.

@put(path, *, name=None, expect_handler=None)

Add a new RouteDef item for registering PUT web-handler.

See UrlDispatcher.add_put() for information about parameters.

@patch(path, *, name=None, expect_handler=None)

Add a new RouteDef item for registering PATCH web-handler.

See UrlDispatcher.add_patch() for information about parameters.

@delete(path, *, name=None, expect_handler=None)

Add a new RouteDef item for registering DELETE web-handler.

See UrlDispatcher.add_delete() for information about parameters.

@view(path, *, name=None, expect_handler=None)

Add a new RouteDef item for registering ANY methods against a class-based view.

See UrlDispatcher.add_view() for information about parameters.

New in version 3.0.

static(prefix, path, *, name=None, expect_handler=None, chunk_size=256 * 1024, show_index=False, follow_symlinks=False, append_version=False)

Add a new StaticDef item for registering static files processor.

See UrlDispatcher.add_static() for information about supported parameters.

New in version 3.1.

@route(method, path, *, name=None, expect_handler=None)

Add a new RouteDef item for registering a web-handler for arbitrary HTTP method.

See UrlDispatcher.add_route() for information about parameters.

MatchInfo

After route matching web application calls found handler if any.

Matching result can be accessible from handler as Request.match_info attribute.

In general the result may be any object derived from AbstractMatchInfo (UrlMappingMatchInfo for default UrlDispatcher router).

class aiohttp.web.UrlMappingMatchInfo

Inherited from dict and AbstractMatchInfo. Dict items are filled by matching info and is resource-specific.

expect_handler

A coroutine for handling 100-continue.

handler

A coroutine for handling request.

route

Route instance for url matching.

View
class aiohttp.web.View(request)

Inherited from AbstractView.

Base class for class based views. Implementations should derive from View and override methods for handling HTTP verbs like get() or post():

class MyView(View):

    async def get(self):
        resp = await get_response(self.request)
        return resp

    async def post(self):
        resp = await post_response(self.request)
        return resp

app.router.add_view('/view', MyView)

The view raises 405 Method Not allowed (HTTPMethodNotAllowed) if requested web verb is not supported.

Parameters

request – instance of Request that has initiated a view processing.

request

Request sent to view’s constructor, read-only property.

Overridable coroutine methods: connect(), delete(), get(), head(), options(), patch(), post(), put(), trace().

See also

Class Based Views

Running Applications

To start web application there is AppRunner and site classes.

Runner is a storage for running application, sites are for running application on specific TCP or Unix socket, e.g.:

runner = web.AppRunner(app)
await runner.setup()
site = web.TCPSite(runner, 'localhost', 8080)
await site.start()
# wait for finish signal
await runner.cleanup()

New in version 3.0: AppRunner / ServerRunner and TCPSite / UnixSite / SockSite are added in aiohttp 3.0

class aiohttp.web.BaseRunner

A base class for runners. Use AppRunner for serving Application, ServerRunner for low-level Server.

server

Low-level web Server for handling HTTP requests, read-only attribute.

addresses

A list of served sockets addresses.

See socket.getsockname() for items type.

New in version 3.3.

sites

A read-only set of served sites (TCPSite / UnixSite / NamedPipeSite / SockSite instances).

coroutine setup()

Initialize the server. Should be called before adding sites.

coroutine cleanup()

Stop handling all registered sites and cleanup used resources.

class aiohttp.web.AppRunner(app, *, handle_signals=False, **kwargs)

A runner for Application. Used with conjunction with sites to serve on specific port.

Inherited from BaseRunner.

Parameters

  • app (Application) – web application instance to serve.

  • handle_signals (bool) – add signal handlers for signal.SIGINT and signal.SIGTERM (False by default).

  • kwargs – named parameters to pass into web protocol.

Supported kwargs:

Parameters

  • tcp_keepalive (bool) – Enable TCP Keep-Alive. Default: True.

  • keepalive_timeout (int) – Number of seconds before closing Keep-Alive connection. Default: 75 seconds (NGINX’s default value).

  • logger – Custom logger object. Default: aiohttp.log.server_logger.

  • access_log – Custom logging object. Default: aiohttp.log.access_logger.

  • access_log_class – Class for access_logger. Default: aiohttp.helpers.AccessLogger. Must to be a subclass of aiohttp.abc.AbstractAccessLogger.

  • access_log_format (str) – Access log format string. Default: helpers.AccessLogger.LOG_FORMAT.

  • max_line_size (int) – Optional maximum header line size. Default: 8190.

  • max_headers (int) – Optional maximum header size. Default: 32768.

  • max_field_size (int) – Optional maximum header field size. Default: 8190.

  • lingering_time (float) – Maximum time during which the server reads and ignores additional data coming from the client when lingering close is on. Use 0 to disable lingering on server channel closing.

  • read_bufsize (int) –

    Size of the read buffer (BaseRequest.content).

    None by default, it means that the session global value is used.

    New in version 3.7.

app

Read-only attribute for accessing to Application served instance.

coroutine setup()

Initialize application. Should be called before adding sites.

The method calls Application.on_startup registered signals.

coroutine cleanup()

Stop handling all registered sites and cleanup used resources.

Application.on_shutdown and Application.on_cleanup signals are called internally.

class aiohttp.web.ServerRunner(web_server, *, handle_signals=False, **kwargs)

A runner for low-level Server. Used with conjunction with sites to serve on specific port.

Inherited from BaseRunner.

Parameters

  • web_server (Server) – low-level web server instance to serve.

  • handle_signals (bool) – add signal handlers for signal.SIGINT and signal.SIGTERM (False by default).

  • kwargs – named parameters to pass into web protocol.

See also

Low Level Server demonstrates low-level server usage

class aiohttp.web.BaseSite

An abstract class for handled sites.

name

An identifier for site, read-only str property. Could be a handled URL or UNIX socket path.

coroutine start()

Start handling a site.

coroutine stop()

Stop handling a site.

class aiohttp.web.TCPSite(runner, host=None, port=None, *, shutdown_timeout=60.0, ssl_context=None, backlog=128, reuse_address=None, reuse_port=None)

Serve a runner on TCP socket.

Parameters

  • runner – a runner to serve.

  • host (str) – HOST to listen on, all interfaces if None (default).

  • port (int) – PORT to listed on, 8080 if None (default).

  • shutdown_timeout (float) – a timeout for closing opened connections on BaseSite.stop() call.

  • ssl_context – a ssl.SSLContext instance for serving SSL/TLS secure server, None for plain HTTP server (default).

  • backlog (int) –

    a number of unaccepted connections that the system will allow before refusing new connections, see socket.listen() for details.

    128 by default.

  • reuse_address (bool) – tells the kernel to reuse a local socket in TIME_WAIT state, without waiting for its natural timeout to expire. If not specified will automatically be set to True on UNIX.

  • reuse_port (bool) – tells the kernel to allow this endpoint to be bound to the same port as other existing endpoints are bound to, so long as they all set this flag when being created. This option is not supported on Windows.

class aiohttp.web.UnixSite(runner, path, *, shutdown_timeout=60.0, ssl_context=None, backlog=128)

Serve a runner on UNIX socket.

Parameters

  • runner – a runner to serve.

  • path (str) – PATH to UNIX socket to listen.

  • shutdown_timeout (float) – a timeout for closing opened connections on BaseSite.stop() call.

  • ssl_context – a ssl.SSLContext instance for serving SSL/TLS secure server, None for plain HTTP server (default).

  • backlog (int) –

    a number of unaccepted connections that the system will allow before refusing new connections, see socket.listen() for details.

    128 by default.

class aiohttp.web.NamedPipeSite(runner, path, *, shutdown_timeout=60.0)

Serve a runner on Named Pipe in Windows.

Parameters

  • runner – a runner to serve.

  • path (str) – PATH of named pipe to listen.

  • shutdown_timeout (float) – a timeout for closing opened connections on BaseSite.stop() call.

class aiohttp.web.SockSite(runner, sock, *, shutdown_timeout=60.0, ssl_context=None, backlog=128)

Serve a runner on UNIX socket.

Parameters

  • runner – a runner to serve.

  • sock – socket.socket to listen.

  • shutdown_timeout (float) – a timeout for closing opened connections on BaseSite.stop() call.

  • ssl_context – a ssl.SSLContext instance for serving SSL/TLS secure server, None for plain HTTP server (default).

  • backlog (int) –

    a number of unaccepted connections that the system will allow before refusing new connections, see socket.listen() for details.

    128 by default.

Utilities
class aiohttp.web.FileField

A namedtuple instance that is returned as multidict value by Request.POST() if field is uploaded file.

name

Field name

filename

File name as specified by uploading (client) side.

file

An io.IOBase instance with content of uploaded file.

content_type

MIME type of uploaded file, 'text/plain' by default.

See also

File Uploads

aiohttp.web.run_app(app, *, host=None, port=None, path=None, sock=None, shutdown_timeout=60.0, ssl_context=None, print=print, backlog=128, access_log_class=aiohttp.helpers.AccessLogger, access_log_format=aiohttp.helpers.AccessLogger.LOG_FORMAT, access_log=aiohttp.log.access_logger, handle_signals=True, reuse_address=None, reuse_port=None)

A utility function for running an application, serving it until keyboard interrupt and performing a Graceful shutdown.

Suitable as handy tool for scaffolding aiohttp based projects. Perhaps production config will use more sophisticated runner but it good enough at least at very beginning stage.

The server will listen on any host or Unix domain socket path you supply. If no hosts or paths are supplied, or only a port is supplied, a TCP server listening on 0.0.0.0 (all hosts) will be launched.

Distributing HTTP traffic to multiple hosts or paths on the same application process provides no performance benefit as the requests are handled on the same event loop. See Server Deployment for ways of distributing work for increased performance.

Parameters

  • app – Application instance to run or a coroutine that returns an application.

  • host (str) – TCP/IP host or a sequence of hosts for HTTP server. Default is '0.0.0.0' if port has been specified or if path is not supplied.

  • port (int) – TCP/IP port for HTTP server. Default is 8080 for plain text HTTP and 8443 for HTTP via SSL (when ssl_context parameter is specified).

  • path (str) – file system path for HTTP server Unix domain socket. A sequence of file system paths can be used to bind multiple domain sockets. Listening on Unix domain sockets is not supported by all operating systems.

  • sock (socket) – a preexisting socket object to accept connections on. A sequence of socket objects can be passed.

  • shutdown_timeout (int) –

    a delay to wait for graceful server shutdown before disconnecting all open client sockets hard way.

    A system with properly Graceful shutdown implemented never waits for this timeout but closes a server in a few milliseconds.

  • ssl_context – ssl.SSLContext for HTTPS server, None for HTTP connection.

  • print – a callable compatible with print(). May be used to override STDOUT output or suppress it. Passing None disables output.

  • backlog (int) – the number of unaccepted connections that the system will allow before refusing new connections (128 by default).

  • access_log_class – class for access_logger. Default: aiohttp.helpers.AccessLogger. Must to be a subclass of aiohttp.abc.AbstractAccessLogger.

  • access_log – logging.Logger instance used for saving access logs. Use None for disabling logs for sake of speedup.

  • access_log_format – access log format, see Format specification for details.

  • handle_signals (bool) – override signal TERM handling to gracefully exit the application.

  • reuse_address (bool) – tells the kernel to reuse a local socket in TIME_WAIT state, without waiting for its natural timeout to expire. If not specified will automatically be set to True on UNIX.

  • reuse_port (bool) – tells the kernel to allow this endpoint to be bound to the same port as other existing endpoints are bound to, so long as they all set this flag when being created. This option is not supported on Windows.

New in version 3.0: Support access_log_class parameter.

Support reuse_address, reuse_port parameter.

New in version 3.1: Accept a coroutine as app parameter.

Constants
class aiohttp.web.ContentCoding

An enum.Enum class of available Content Codings.

deflate

DEFLATE compression

gzip

GZIP compression

identity

no compression

Middlewares
Normalize path middleware
aiohttp.web.normalize_path_middleware(*, append_slash=True, remove_slash=False, merge_slashes=True, redirect_class=HTTPPermanentRedirect)

Middleware factory which produces a middleware that normalizes the path of a request. By normalizing it means:

  • Add or remove a trailing slash to the path.

  • Double slashes are replaced by one.

The middleware returns as soon as it finds a path that resolves correctly. The order if both merge and append/remove are enabled is:

  1. merge_slashes

  2. append_slash or remove_slash

  3. both merge_slashes and append_slash or remove_slash

If the path resolves with at least one of those conditions, it will redirect to the new path.

Only one of append_slash and remove_slash can be enabled. If both are True the factory will raise an AssertionError

If append_slash is True the middleware will append a slash when needed. If a resource is defined with trailing slash and the request comes without it, it will append it automatically.

If remove_slash is True, append_slash must be False. When enabled the middleware will remove trailing slashes and redirect if the resource is defined.

If merge_slashes is True, merge multiple consecutive slashes in the path into one.

New in version 3.4: Support for remove_slash

Logging

aiohttp uses standard logging for tracking the library activity.

We have the following loggers enumerated by names:

  • 'aiohttp.access'

  • 'aiohttp.client'

  • 'aiohttp.internal'

  • 'aiohttp.server'

  • 'aiohttp.web'

  • 'aiohttp.websocket'

You may subscribe to these loggers for getting logging messages. The page does not provide instructions for logging subscribing while the most friendly method is logging.config.dictConfig() for configuring whole loggers in your application.

Logging does not work out of the box. It requires at least minimal 'logging' configuration. Example of minimal working logger setup:

import logging
from aiohttp import web

app = web.Application()
logging.basicConfig(level=logging.DEBUG)
web.run_app(app, port=5000)

New in version 4.0.0.

Access logs

Access logs are enabled by default. If the debug flag is set, and the default logger 'aiohttp.access' is used, access logs will be output to stderr if no handlers are attached. Furthermore, if the default logger has no log level set, the log level will be set to logging.DEBUG.

This logging may be controlled by aiohttp.web.AppRunner() and aiohttp.web.run_app().

To override the default logger, pass an instance of logging.Logger to override the default logger.

Note

Use web.run_app(app, access_log=None) to disable access logs.

In addition, access_log_format may be used to specify the log format.

Format specification

The library provides custom micro-language to specifying info about request and response:

Option

Meaning

%%

The percent sign

%a

Remote IP-address (IP-address of proxy if using reverse proxy)

%t

Time when the request was started to process

%P

The process ID of the child that serviced the request

%r

First line of request

%s

Response status code

%b

Size of response in bytes, including HTTP headers

%T

The time taken to serve the request, in seconds

%Tf

The time taken to serve the request, in seconds with fraction in %.06f format

%D

The time taken to serve the request, in microseconds

%{FOO}i

request.headers['FOO']

%{FOO}o

response.headers['FOO']

The default access log format is:

'%a %t "%r" %s %b "%{Referer}i" "%{User-Agent}i"'

New in version 2.3.0.

access_log_class introduced.

Example of a drop-in replacement for the default access logger:

from aiohttp.abc import AbstractAccessLogger

class AccessLogger(AbstractAccessLogger):

    def log(self, request, response, time):
        self.logger.info(f'{request.remote} '
                         f'"{request.method} {request.path} '
                         f'done in {time}s: {response.status}')
Gunicorn access logs

When Gunicorn is used for deployment, its default access log format will be automatically replaced with the default aiohttp’s access log format.

If Gunicorn’s option access_logformat is specified explicitly, it should use aiohttp’s format specification.

Gunicorn’s access log works only if accesslog is specified explicitly in your config or as a command line option. This configuration can be either a path or '-'. If the application uses a custom logging setup intercepting the 'gunicorn.access' logger, accesslog should be set to '-' to prevent Gunicorn to create an empty access log file upon every startup.

Error logs

aiohttp.web uses a logger named 'aiohttp.server' to store errors given on web requests handling.

This log is enabled by default.

To use a different logger name, pass logger (logging.Logger instance) to the aiohttp.web.AppRunner() constructor.

Testing

Testing aiohttp web servers

aiohttp provides plugin for pytest making writing web server tests extremely easy, it also provides test framework agnostic utilities for testing with other frameworks such as unittest.

Before starting to write your tests, you may also be interested on reading how to write testable services that interact with the loop.

For using pytest plugin please install pytest-aiohttp library:

$ pip install pytest-aiohttp

If you don’t want to install pytest-aiohttp for some reason you may insert pytest_plugins = 'aiohttp.pytest_plugin' line into conftest.py instead for the same functionality.

Provisional Status

The module is a provisional.

aiohttp has a year and half period for removing deprecated API (Policy for Backward Incompatible Changes).

But for aiohttp.test_tools the deprecation period could be reduced.

Moreover we may break backward compatibility without deprecation period for some very strong reason.

The Test Client and Servers

aiohttp test utils provides a scaffolding for testing aiohttp-based web servers.

They are consist of two parts: running test server and making HTTP requests to this server.

TestServer runs aiohttp.web.Application based server, RawTestServer starts aiohttp.web.WebServer low level server.

For performing HTTP requests to these servers you have to create a test client: TestClient instance.

The client incapsulates aiohttp.ClientSession by providing proxy methods to the client for common operations such as ws_connect, get, post, etc.

Pytest

The aiohttp_client fixture available from pytest-aiohttp plugin allows you to create a client to make requests to test your app.

A simple would be:

from aiohttp import web

async def hello(request):
    return web.Response(text='Hello, world')

async def test_hello(aiohttp_client, loop):
    app = web.Application()
    app.router.add_get('/', hello)
    client = await aiohttp_client(app)
    resp = await client.get('/')
    assert resp.status == 200
    text = await resp.text()
    assert 'Hello, world' in text

It also provides access to the app instance allowing tests to check the state of the app. Tests can be made even more succinct with a fixture to create an app test client:

import pytest
from aiohttp import web


async def previous(request):
    if request.method == 'POST':
        request.app['value'] = (await request.post())['value']
        return web.Response(body=b'thanks for the data')
    return web.Response(
        body='value: {}'.format(request.app['value']).encode('utf-8'))

@pytest.fixture
def cli(loop, aiohttp_client):
    app = web.Application()
    app.router.add_get('/', previous)
    app.router.add_post('/', previous)
    return loop.run_until_complete(aiohttp_client(app))

async def test_set_value(cli):
    resp = await cli.post('/', data={'value': 'foo'})
    assert resp.status == 200
    assert await resp.text() == 'thanks for the data'
    assert cli.server.app['value'] == 'foo'

async def test_get_value(cli):
    cli.server.app['value'] = 'bar'
    resp = await cli.get('/')
    assert resp.status == 200
    assert await resp.text() == 'value: bar'

Pytest tooling has the following fixtures:

pytest_aiohttp.aiohttp_server(app, *, port=None, **kwargs)

A fixture factory that creates TestServer:

async def test_f(aiohttp_server):
    app = web.Application()
    # fill route table

    server = await aiohttp_server(app)

The server will be destroyed on exit from test function.

app is the aiohttp.web.Application used

to start server.

port optional, port the server is run at, if not provided a random unused port is used.

New in version 3.0.

kwargs are parameters passed to

aiohttp.web.Application.make_handler()

Changed in version 3.0.

Deprecated since version 3.2: The fixture was renamed from test_server to aiohttp_server.

pytest_aiohttp.aiohttp_client(app, server_kwargs=None, **kwargs)
pytest_aiohttp.aiohttp_client(server, **kwargs)
pytest_aiohttp.aiohttp_client(raw_server, **kwargs)

A fixture factory that creates TestClient for access to tested server:

async def test_f(aiohttp_client):
    app = web.Application()
    # fill route table

    client = await aiohttp_client(app)
    resp = await client.get('/')

client and responses are cleaned up after test function finishing.

The fixture accepts aiohttp.web.Application, aiohttp.test_utils.TestServer or aiohttp.test_utils.RawTestServer instance.

server_kwargs are parameters passed to the test server if an app is passed, else ignored.

kwargs are parameters passed to aiohttp.test_utils.TestClient constructor.

Changed in version 3.0: The fixture was renamed from test_client to aiohttp_client.

pytest_aiohttp.aiohttp_raw_server(handler, *, port=None, **kwargs)

A fixture factory that creates RawTestServer instance from given web handler.:

async def test_f(aiohttp_raw_server, aiohttp_client):

    async def handler(request):
        return web.Response(text="OK")

    raw_server = await aiohttp_raw_server(handler)
    client = await aiohttp_client(raw_server)
    resp = await client.get('/')

handler should be a coroutine which accepts a request and returns response, e.g.

port optional, port the server is run at, if not provided a random unused port is used.

New in version 3.0.

pytest_aiohttp.aiohttp_unused_port

Function to return an unused port number for IPv4 TCP protocol:

async def test_f(aiohttp_client, aiohttp_unused_port):
    port = aiohttp_unused_port()
    app = web.Application()
    # fill route table

    client = await aiohttp_client(app, server_kwargs={'port': port})
    ...

Changed in version 3.0: The fixture was renamed from unused_port to aiohttp_unused_port.

Unittest

To test applications with the standard library’s unittest or unittest-based functionality, the AioHTTPTestCase is provided:

from aiohttp.test_utils import AioHTTPTestCase, unittest_run_loop
from aiohttp import web

class MyAppTestCase(AioHTTPTestCase):

    async def get_application(self):
        """
        Override the get_app method to return your application.
        """
        async def hello(request):
            return web.Response(text='Hello, world')

        app = web.Application()
        app.router.add_get('/', hello)
        return app

    # the unittest_run_loop decorator can be used in tandem with
    # the AioHTTPTestCase to simplify running
    # tests that are asynchronous
    @unittest_run_loop
    async def test_example(self):
        resp = await self.client.request("GET", "/")
        assert resp.status == 200
        text = await resp.text()
        assert "Hello, world" in text

    # a vanilla example
    def test_example_vanilla(self):
        async def test_get_route():
            url = "/"
            resp = await self.client.request("GET", url)
            assert resp.status == 200
            text = await resp.text()
            assert "Hello, world" in text

        self.loop.run_until_complete(test_get_route())
class aiohttp.test_utils.AioHTTPTestCase

A base class to allow for unittest web applications using aiohttp.

Derived from unittest.TestCase

Provides the following:

client

an aiohttp test client, TestClient instance.

server

an aiohttp test server, TestServer instance.

New in version 2.3.

loop

The event loop in which the application and server are running.

Deprecated since version 3.5.

app

The application returned by get_app() (aiohttp.web.Application instance).

coroutine get_client()

This async method can be overridden to return the TestClient object used in the test.

Returns

TestClient instance.

New in version 2.3.

coroutine get_server()

This async method can be overridden to return the TestServer object used in the test.

Returns

TestServer instance.

New in version 2.3.

coroutine get_application()

This async method should be overridden to return the aiohttp.web.Application object to test.

Returns

aiohttp.web.Application instance.

coroutine setUpAsync()

This async method do nothing by default and can be overridden to execute asynchronous code during the setUp stage of the TestCase.

New in version 2.3.

coroutine tearDownAsync()

This async method do nothing by default and can be overridden to execute asynchronous code during the tearDown stage of the TestCase.

New in version 2.3.

setUp()

Standard test initialization method.

tearDown()

Standard test finalization method.

Note

The TestClient’s methods are asynchronous: you have to execute function on the test client using asynchronous methods.

A basic test class wraps every test method by unittest_run_loop() decorator:

class TestA(AioHTTPTestCase):

    @unittest_run_loop
    async def test_f(self):
        resp = await self.client.get('/')
unittest_run_loop:

A decorator dedicated to use with asynchronous methods of an AioHTTPTestCase.

Handles executing an asynchronous function, using the AioHTTPTestCase.loop of the AioHTTPTestCase.

Faking request object

aiohttp provides test utility for creating fake aiohttp.web.Request objects: aiohttp.test_utils.make_mocked_request(), it could be useful in case of simple unit tests, like handler tests, or simulate error conditions that hard to reproduce on real server:

from aiohttp import web
from aiohttp.test_utils import make_mocked_request

def handler(request):
    assert request.headers.get('token') == 'x'
    return web.Response(body=b'data')

def test_handler():
    req = make_mocked_request('GET', '/', headers={'token': 'x'})
    resp = handler(req)
    assert resp.body == b'data'

Warning

We don’t recommend to apply make_mocked_request() everywhere for testing web-handler’s business object – please use test client and real networking via ‘localhost’ as shown in examples before.

make_mocked_request() exists only for testing complex cases (e.g. emulating network errors) which are extremely hard or even impossible to test by conventional way.

aiohttp.test_utils.make_mocked_request(method, path, headers=None, *, version=HttpVersion(1, 1), closing=False, app=None, match_info=sentinel, reader=sentinel, writer=sentinel, transport=sentinel, payload=sentinel, sslcontext=None, loop=...)

Creates mocked web.Request testing purposes.

Useful in unit tests, when spinning full web server is overkill or specific conditions and errors are hard to trigger.

Parameters

  • method (str) – str, that represents HTTP method, like; GET, POST.

  • path (str) – str, The URL including PATH INFO without the host or scheme

  • headers (dict, multidict.CIMultiDict, list of pairs) – mapping containing the headers. Can be anything accepted by the multidict.CIMultiDict constructor.

  • match_info (dict) – mapping containing the info to match with url parameters.

  • version (aiohttp.protocol.HttpVersion) – namedtuple with encoded HTTP version

  • closing (bool) – flag indicates that connection should be closed after response.

  • app (aiohttp.web.Application) – the aiohttp.web application attached for fake request

  • writer (aiohttp.StreamWriter) – object for managing outcoming data

  • transport (asyncio.transports.Transport) – asyncio transport instance

  • payload (aiohttp.StreamReader) – raw payload reader object

  • sslcontext (ssl.SSLContext) – ssl.SSLContext object, for HTTPS connection

  • loop (asyncio.AbstractEventLoop) – An event loop instance, mocked loop by default.

Returns

aiohttp.web.Request object.

New in version 2.3: match_info parameter.

Framework Agnostic Utilities

High level test creation:

from aiohttp.test_utils import TestClient, TestServer, loop_context
from aiohttp import request

# loop_context is provided as a utility. You can use any
# asyncio.BaseEventLoop class in its place.
with loop_context() as loop:
    app = _create_example_app()
    with TestClient(TestServer(app), loop=loop) as client:

        async def test_get_route():
            nonlocal client
            resp = await client.get("/")
            assert resp.status == 200
            text = await resp.text()
            assert "Hello, world" in text

        loop.run_until_complete(test_get_route())

If it’s preferred to handle the creation / teardown on a more granular basis, the TestClient object can be used directly:

from aiohttp.test_utils import TestClient, TestServer

with loop_context() as loop:
    app = _create_example_app()
    client = TestClient(TestServer(app), loop=loop)
    loop.run_until_complete(client.start_server())
    root = "http://127.0.0.1:{}".format(port)

    async def test_get_route():
        resp = await client.get("/")
        assert resp.status == 200
        text = await resp.text()
        assert "Hello, world" in text

    loop.run_until_complete(test_get_route())
    loop.run_until_complete(client.close())

A full list of the utilities provided can be found at the api reference

Testing API Reference
Test server

Runs given aiohttp.web.Application instance on random TCP port.

After creation the server is not started yet, use start_server() for actual server starting and close() for stopping/cleanup.

Test server usually works in conjunction with aiohttp.test_utils.TestClient which provides handy client methods for accessing to the server.

class aiohttp.test_utils.BaseTestServer(*, scheme='http', host='127.0.0.1', port=None)

Base class for test servers.

Parameters

  • scheme (str) – HTTP scheme, non-protected "http" by default.

  • host (str) – a host for TCP socket, IPv4 local host ('127.0.0.1') by default.

  • port (int) –

    optional port for TCP socket, if not provided a random unused port is used.

    New in version 3.0.

scheme

A scheme for tested application, 'http' for non-protected run and 'https' for TLS encrypted server.

host

host used to start a test server.

port

port used to start the test server.

handler

aiohttp.web.WebServer used for HTTP requests serving.

server

asyncio.AbstractServer used for managing accepted connections.

coroutine start_server(loop=None, **kwargs)
Parameters

loop (asyncio.AbstractEventLoop) – the event_loop to use

Start a test server.

coroutine close()

Stop and finish executed test server.

make_url(path)

Return an absolute URL for given path.

class aiohttp.test_utils.RawTestServer(handler, *, scheme='http', host='127.0.0.1')

Low-level test server (derived from BaseTestServer).

Parameters

  • handler –

    a coroutine for handling web requests. The handler should accept aiohttp.web.BaseRequest and return a response instance, e.g. StreamResponse or Response.

    The handler could raise HTTPException as a signal for non-200 HTTP response.

  • scheme (str) – HTTP scheme, non-protected "http" by default.

  • host (str) – a host for TCP socket, IPv4 local host ('127.0.0.1') by default.

  • port (int) –

    optional port for TCP socket, if not provided a random unused port is used.

    New in version 3.0.

class aiohttp.test_utils.TestServer(app, *, scheme='http', host='127.0.0.1')

Test server (derived from BaseTestServer) for starting Application.

Parameters

  • app – aiohttp.web.Application instance to run.

  • scheme (str) – HTTP scheme, non-protected "http" by default.

  • host (str) – a host for TCP socket, IPv4 local host ('127.0.0.1') by default.

  • port (int) –

    optional port for TCP socket, if not provided a random unused port is used.

    New in version 3.0.

app

aiohttp.web.Application instance to run.

Test Client
class aiohttp.test_utils.TestClient(app_or_server, *, loop=None, scheme='http', host='127.0.0.1', cookie_jar=None, **kwargs)

A test client used for making calls to tested server.

Parameters

  • app_or_server –

    BaseTestServer instance for making client requests to it.

    In order to pass a aiohttp.web.Application you need to convert it first to TestServer first with TestServer(app).

  • cookie_jar – an optional aiohttp.CookieJar instance, may be useful with CookieJar(unsafe=True) option.

  • scheme (str) – HTTP scheme, non-protected "http" by default.

  • loop (asyncio.AbstractEventLoop) – the event_loop to use

  • host (str) – a host for TCP socket, IPv4 local host ('127.0.0.1') by default.

scheme

A scheme for tested application, 'http' for non-protected run and 'https' for TLS encrypted server.

host

host used to start a test server.

port

port used to start the server

server

BaseTestServer test server instance used in conjunction with client.

app

An alias for self.server.app. return None if self.server is not TestServer instance(e.g. RawTestServer instance for test low-level server).

session

An internal aiohttp.ClientSession.

Unlike the methods on the TestClient, client session requests do not automatically include the host in the url queried, and will require an absolute path to the resource.

coroutine start_server(**kwargs)

Start a test server.

coroutine close()

Stop and finish executed test server.

make_url(path)

Return an absolute URL for given path.

coroutine request(method, path, *args, **kwargs)

Routes a request to tested http server.

The interface is identical to aiohttp.ClientSession.request(), except the loop kwarg is overridden by the instance used by the test server.

coroutine get(path, *args, **kwargs)

Perform an HTTP GET request.

coroutine post(path, *args, **kwargs)

Perform an HTTP POST request.

coroutine options(path, *args, **kwargs)

Perform an HTTP OPTIONS request.

coroutine head(path, *args, **kwargs)

Perform an HTTP HEAD request.

coroutine put(path, *args, **kwargs)

Perform an HTTP PUT request.

coroutine patch(path, *args, **kwargs)

Perform an HTTP PATCH request.

coroutine delete(path, *args, **kwargs)

Perform an HTTP DELETE request.

coroutine ws_connect(path, *args, **kwargs)

Initiate websocket connection.

The api corresponds to aiohttp.ClientSession.ws_connect().

Utilities
aiohttp.test_utils.make_mocked_coro(return_value)

Creates a coroutine mock.

Behaves like a coroutine which returns return_value. But it is also a mock object, you might test it as usual Mock:

mocked = make_mocked_coro(1)
assert 1 == await mocked(1, 2)
mocked.assert_called_with(1, 2)
Parameters

return_value – A value that the the mock object will return when called.

Returns

A mock object that behaves as a coroutine which returns return_value when called.

aiohttp.test_utils.unused_port()

Return an unused port number for IPv4 TCP protocol.

Return int

ephemeral port number which could be reused by test server.

aiohttp.test_utils.loop_context(loop_factory=<function asyncio.new_event_loop>)

A contextmanager that creates an event_loop, for test purposes.

Handles the creation and cleanup of a test loop.

aiohttp.test_utils.setup_test_loop(loop_factory=<function asyncio.new_event_loop>)

Create and return an asyncio.AbstractEventLoop instance.

The caller should also call teardown_test_loop, once they are done with the loop.

Note

As side effect the function changes asyncio default loop by asyncio.set_event_loop() call.

Previous default loop is not restored.

It should not be a problem for test suite: every test expects a new test loop instance anyway.

Changed in version 3.1: The function installs a created event loop as default.

aiohttp.test_utils.teardown_test_loop(loop)

Teardown and cleanup an event_loop created by setup_test_loop.

Parameters

loop (asyncio.AbstractEventLoop) – the loop to teardown

Server Deployment

There are several options for aiohttp server deployment:

  • Standalone server

  • Running a pool of backend servers behind of nginx, HAProxy or other reverse proxy server

  • Using gunicorn behind of reverse proxy

Every method has own benefits and disadvantages.

Standalone

Just call aiohttp.web.run_app() function passing aiohttp.web.Application instance.

The method is very simple and could be the best solution in some trivial cases. But it does not utilize all CPU cores.

For running multiple aiohttp server instances use reverse proxies.

Nginx+supervisord

Running aiohttp servers behind nginx makes several advantages.

At first, nginx is the perfect frontend server. It may prevent many attacks based on malformed http protocol etc.

Second, running several aiohttp instances behind nginx allows to utilize all CPU cores.

Third, nginx serves static files much faster than built-in aiohttp static file support.

But this way requires more complex configuration.

Nginx configuration

Here is short extraction about writing Nginx configuration file. It does not cover all available Nginx options.

For full reference read Nginx tutorial and official Nginx documentation.

First configure HTTP server itself:

http {
  server {
    listen 80;
    client_max_body_size 4G;

    server_name example.com;

    location / {
      proxy_set_header Host $http_host;
      proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
      proxy_redirect off;
      proxy_buffering off;
      proxy_pass http://aiohttp;
    }

    location /static {
      # path for static files
      root /path/to/app/static;
    }

  }
}

This config listens on port 80 for server named example.com and redirects everything to aiohttp backend group.

Also it serves static files from /path/to/app/static path as example.com/static.

Next we need to configure aiohttp upstream group:

http {
  upstream aiohttp {
    # fail_timeout=0 means we always retry an upstream even if it failed
    # to return a good HTTP response

    # Unix domain servers
    server unix:/tmp/example_1.sock fail_timeout=0;
    server unix:/tmp/example_2.sock fail_timeout=0;
    server unix:/tmp/example_3.sock fail_timeout=0;
    server unix:/tmp/example_4.sock fail_timeout=0;

    # Unix domain sockets are used in this example due to their high performance,
    # but TCP/IP sockets could be used instead:
    # server 127.0.0.1:8081 fail_timeout=0;
    # server 127.0.0.1:8082 fail_timeout=0;
    # server 127.0.0.1:8083 fail_timeout=0;
    # server 127.0.0.1:8084 fail_timeout=0;
  }
}

All HTTP requests for http://example.com except ones for http://example.com/static will be redirected to example1.sock, example2.sock, example3.sock or example4.sock backend servers. By default, Nginx uses round-robin algorithm for backend selection.

Note

Nginx is not the only existing reverse proxy server but the most popular one. Alternatives like HAProxy may be used as well.

Supervisord

After configuring Nginx we need to start our aiohttp backends. Better to use some tool for starting them automatically after system reboot or backend crash.

There are very many ways to do it: Supervisord, Upstart, Systemd, Gaffer, Circus, Runit etc.

Here we’ll use Supervisord for example:

[program:aiohttp]
numprocs = 4
numprocs_start = 1
process_name = example_%(process_num)s

; Unix socket paths are specified by command line.
command=/path/to/aiohttp_example.py --path=/tmp/example_%(process_num)s.sock

; We can just as easily pass TCP port numbers:
; command=/path/to/aiohttp_example.py --port=808%(process_num)s

user=nobody
autostart=true
autorestart=true
aiohttp server

The last step is preparing aiohttp server for working with supervisord.

Assuming we have properly configured aiohttp.web.Application and port is specified by command line, the task is trivial:

# aiohttp_example.py
import argparse
from aiohttp import web

parser = argparse.ArgumentParser(description="aiohttp server example")
parser.add_argument('--path')
parser.add_argument('--port')


if __name__ == '__main__':
    app = web.Application()
    # configure app

    args = parser.parse_args()
    web.run_app(app, path=args.path, port=args.port)

For real use cases we perhaps need to configure other things like logging etc., but it’s out of scope of the topic.

Nginx+Gunicorn

aiohttp can be deployed using Gunicorn, which is based on a pre-fork worker model. Gunicorn launches your app as worker processes for handling incoming requests.

In opposite to deployment with bare Nginx the solution does not need to manually run several aiohttp processes and use tool like supervisord for monitoring it. But nothing is for free: running aiohttp application under gunicorn is slightly slower.

Prepare environment

You firstly need to setup your deployment environment. This example is based on Ubuntu 16.04.

Create a directory for your application:

>> mkdir myapp
>> cd myapp

Create Python virtual environment:

>> python3 -m venv venv
>> source venv/bin/activate

Now that the virtual environment is ready, we’ll proceed to install aiohttp and gunicorn:

>> pip install gunicorn
>> pip install aiohttp
Application

Lets write a simple application, which we will save to file. We’ll name this file my_app_module.py:

from aiohttp import web

async def index(request):
    return web.Response(text="Welcome home!")


my_web_app = web.Application()
my_web_app.router.add_get('/', index)
Application factory

As an option an entry point could be a coroutine that accepts no parameters and returns an application instance:

from aiohttp import web

async def index(request):
    return web.Response(text="Welcome home!")


async def my_web_app():
    app = web.Application()
    app.router.add_get('/', index)
    return app
Start Gunicorn

When Running Gunicorn, you provide the name of the module, i.e. my_app_module, and the name of the app or application factory, i.e. my_web_app, along with other Gunicorn Settings provided as command line flags or in your config file.

In this case, we will use:

  • the --bind flag to set the server’s socket address;

  • the --worker-class flag to tell Gunicorn that we want to use a custom worker subclass instead of one of the Gunicorn default worker types;

  • you may also want to use the --workers flag to tell Gunicorn how many worker processes to use for handling requests. (See the documentation for recommendations on How Many Workers?)

  • you may also want to use the --accesslog flag to enable the access log to be populated. (See logging for more information.)

The custom worker subclass is defined in aiohttp.GunicornWebWorker:

>> gunicorn my_app_module:my_web_app --bind localhost:8080 --worker-class aiohttp.GunicornWebWorker
[2017-03-11 18:27:21 +0000] [1249] [INFO] Starting gunicorn 19.7.1
[2017-03-11 18:27:21 +0000] [1249] [INFO] Listening at: http://127.0.0.1:8080 (1249)
[2017-03-11 18:27:21 +0000] [1249] [INFO] Using worker: aiohttp.worker.GunicornWebWorker
[2015-03-11 18:27:21 +0000] [1253] [INFO] Booting worker with pid: 1253

Gunicorn is now running and ready to serve requests to your app’s worker processes.

Note

If you want to use an alternative asyncio event loop uvloop, you can use the aiohttp.GunicornUVLoopWebWorker worker class.

Proxy through NGINX

We can proxy our gunicorn workers through NGINX with a configuration like this:

worker_processes 1;
user nobody nogroup;
events {
    worker_connections 1024;
}
http {
    ## Main Server Block
    server {
        ## Open by default.
        listen                80 default_server;
        server_name           main;
        client_max_body_size  200M;

        ## Main site location.
        location / {
            proxy_pass                          http://127.0.0.1:8080;
            proxy_set_header                    Host $host;
            proxy_set_header X-Forwarded-Host   $server_name;
            proxy_set_header X-Real-IP          $remote_addr;
        }
    }
}

Since gunicorn listens for requests at our localhost address on port 8080, we can use the proxy_pass directive to send web traffic to our workers. If everything is configured correctly, we should reach our application at the ip address of our web server.

Proxy through NGINX + SSL

Here is an example NGINX configuration setup to accept SSL connections:

worker_processes 1;
user nobody nogroup;
events {
    worker_connections 1024;
}
http {
    ## SSL Redirect
    server {
        listen 80       default;
        return 301      https://$host$request_uri;
    }

    ## Main Server Block
    server {
        # Open by default.
        listen                443 ssl default_server;
        listen                [::]:443 ssl default_server;
        server_name           main;
        client_max_body_size  200M;

        ssl_certificate       /etc/secrets/cert.pem;
        ssl_certificate_key   /etc/secrets/key.pem;

        ## Main site location.
        location / {
            proxy_pass                          http://127.0.0.1:8080;
            proxy_set_header                    Host $host;
            proxy_set_header X-Forwarded-Host   $server_name;
            proxy_set_header X-Real-IP          $remote_addr;
        }
    }
}

The first server block accepts regular http connections on port 80 and redirects them to our secure SSL connection. The second block matches our previous example except we need to change our open port to https and specify where our SSL certificates are being stored with the ssl_certificate and ssl_certificate_key directives.

During development, you may want to create your own self-signed certificates for testing purposes and use another service like Let’s Encrypt when you are ready to move to production.

More information

See the official documentation for more information about suggested nginx configuration. You can also find out more about configuring for secure https connections as well.

Logging configuration

aiohttp and gunicorn use different format for specifying access log.

By default aiohttp uses own defaults:

'%a %t "%r" %s %b "%{Referer}i" "%{User-Agent}i"'

For more information please read Format Specification for Access Log.

Proxy through Apache at your own risk

Issues have been reported using Apache2 in front of aiohttp server: #2687 Intermittent 502 proxy errors when running behind Apache <https://github.com/aio-libs/aiohttp/issues/2687>.

Utilities

Miscellaneous API Shared between Client And Server.

Abstract Base Classes

Abstract routing

aiohttp has abstract classes for managing web interfaces.

The most part of aiohttp.web is not intended to be inherited but few of them are.

aiohttp.web is built on top of few concepts: application, router, request and response.

router is a plugable part: a library user may build a router from scratch, all other parts should work with new router seamlessly.

AbstractRouter has the only mandatory method: AbstractRouter.resolve() coroutine. It must return an AbstractMatchInfo instance.

If the requested URL handler is found AbstractMatchInfo.handler() is a web-handler for requested URL and AbstractMatchInfo.http_exception is None.

Otherwise AbstractMatchInfo.http_exception is an instance of HTTPException like 404: NotFound or 405: Method Not Allowed. AbstractMatchInfo.handler() raises http_exception on call.

class aiohttp.abc.AbstractRouter

Abstract router, aiohttp.web.Application accepts it as router parameter and returns as aiohttp.web.Application.router.

coroutine resolve(request)

Performs URL resolving. It’s an abstract method, should be overridden in router implementation.

Parameters

request – aiohttp.web.Request instance for resolving, the request has aiohttp.web.Request.match_info equals to None at resolving stage.

Returns

AbstractMatchInfo instance.

class aiohttp.abc.AbstractMatchInfo

Abstract match info, returned by AbstractRouter.resolve() call.

http_exception

aiohttp.web.HTTPException if no match was found, None otherwise.

coroutine handler(request)

Abstract method performing web-handler processing.

Parameters

request – aiohttp.web.Request instance for resolving, the request has aiohttp.web.Request.match_info equals to None at resolving stage.

Returns

aiohttp.web.StreamResponse or descendants.

Raise

aiohttp.web.HTTPException on error

coroutine expect_handler(request)

Abstract method for handling 100-continue processing.

Abstract Class Based Views

For class based view support aiohttp has abstract AbstractView class which is awaitable (may be uses like await Cls() or yield from Cls() and has a request as an attribute.

class aiohttp.abc.AbstractView

An abstract class, base for all class based views implementations.

Methods __iter__ and __await__ should be overridden.

request

aiohttp.web.Request instance for performing the request.

Abstract Abstract Access Logger
class aiohttp.abc.AbstractAccessLogger

An abstract class, base for all RequestHandler access_logger implementations

Method log should be overridden.

log(request, response, time)
Parameters

Working with Multipart

aiohttp supports a full featured multipart reader and writer. Both are designed with streaming processing in mind to avoid unwanted footprint which may be significant if you’re dealing with large payloads, but this also means that most I/O operation are only possible to be executed a single time.

Reading Multipart Responses

Assume you made a request, as usual, and want to process the response multipart data:

async with aiohttp.request(...) as resp:
    pass

First, you need to wrap the response with a MultipartReader.from_response(). This needs to keep the implementation of MultipartReader separated from the response and the connection routines which makes it more portable:

reader = aiohttp.MultipartReader.from_response(resp)

Let’s assume with this response you’d received some JSON document and multiple files for it, but you don’t need all of them, just a specific one.

So first you need to enter into a loop where the multipart body will be processed:

metadata = None
filedata = None
while True:
    part = await reader.next()

The returned type depends on what the next part is: if it’s a simple body part then you’ll get BodyPartReader instance here, otherwise, it will be another MultipartReader instance for the nested multipart. Remember, that multipart format is recursive and supports multiple levels of nested body parts. When there are no more parts left to fetch, None value will be returned - that’s the signal to break the loop:

if part is None:
    break

Both BodyPartReader and MultipartReader provides access to body part headers: this allows you to filter parts by their attributes:

if part.headers[aiohttp.hdrs.CONTENT_TYPE] == 'application/json':
    metadata = await part.json()
    continue

Nor BodyPartReader or MultipartReader instances does not read the whole body part data without explicitly asking for. BodyPartReader provides a set of helpers methods to fetch popular content types in friendly way:

Each of these methods automatically recognizes if content is compressed by using gzip and deflate encoding (while it respects identity one), or if transfer encoding is base64 or quoted-printable - in each case the result will get automatically decoded. But in case you need to access to raw binary data as it is, there are BodyPartReader.read() and BodyPartReader.read_chunk() coroutine methods as well to read raw binary data as it is all-in-single-shot or by chunks respectively.

When you have to deal with multipart files, the BodyPartReader.filename property comes to help. It’s a very smart helper which handles Content-Disposition handler right and extracts the right filename attribute from it:

if part.filename != 'secret.txt':
    continue

If current body part does not matches your expectation and you want to skip it - just continue a loop to start a next iteration of it. Here is where magic happens. Before fetching the next body part await reader.next() it ensures that the previous one was read completely. If it was not, all its content sends to the void in term to fetch the next part. So you don’t have to care about cleanup routines while you’re within a loop.

Once you’d found a part for the file you’d searched for, just read it. Let’s handle it as it is without applying any decoding magic:

filedata = await part.read(decode=False)

Later you may decide to decode the data. It’s still simple and possible to do:

filedata = part.decode(filedata)

Once you are done with multipart processing, just break a loop:

break
Sending Multipart Requests

MultipartWriter provides an interface to build multipart payload from the Python data and serialize it into chunked binary stream. Since multipart format is recursive and supports deeply nesting, you can use with statement to design your multipart data closer to how it will be:

with aiohttp.MultipartWriter('mixed') as mpwriter:
    ...
    with aiohttp.MultipartWriter('related') as subwriter:
        ...
    mpwriter.append(subwriter)

    with aiohttp.MultipartWriter('related') as subwriter:
        ...
        with aiohttp.MultipartWriter('related') as subsubwriter:
            ...
        subwriter.append(subsubwriter)
    mpwriter.append(subwriter)

    with aiohttp.MultipartWriter('related') as subwriter:
        ...
    mpwriter.append(subwriter)

The MultipartWriter.append() is used to join new body parts into a single stream. It accepts various inputs and determines what default headers should be used for.

For text data default Content-Type is text/plain; charset=utf-8:

mpwriter.append('hello')

For binary data application/octet-stream is used:

mpwriter.append(b'aiohttp')

You can always override these default by passing your own headers with the second argument:

mpwriter.append(io.BytesIO(b'GIF89a...'),
                {'CONTENT-TYPE': 'image/gif'})

For file objects Content-Type will be determined by using Python’s mod:mimetypes module and additionally Content-Disposition header will include the file’s basename:

part = root.append(open(__file__, 'rb'))

If you want to send a file with a different name, just handle the Payload instance which MultipartWriter.append() will always return and set Content-Disposition explicitly by using the Payload.set_content_disposition() helper:

part.set_content_disposition('attachment', filename='secret.txt')

Additionally, you may want to set other headers here:

part.headers[aiohttp.hdrs.CONTENT_ID] = 'X-12345'

If you’d set Content-Encoding, it will be automatically applied to the data on serialization (see below):

part.headers[aiohttp.hdrs.CONTENT_ENCODING] = 'gzip'

There are also MultipartWriter.append_json() and MultipartWriter.append_form() helpers which are useful to work with JSON and form urlencoded data, so you don’t have to encode it every time manually:

mpwriter.append_json({'test': 'passed'})
mpwriter.append_form([('key', 'value')])

When it’s done, to make a request just pass a root MultipartWriter instance as aiohttp.ClientSession.request() data argument:

await session.post('http://example.com', data=mpwriter)

Behind the scenes MultipartWriter.write() will yield chunks of every part and if body part has Content-Encoding or Content-Transfer-Encoding they will be applied on streaming content.

Please note, that on MultipartWriter.write() all the file objects will be read until the end and there is no way to repeat a request without rewinding their pointers to the start.

Example MJPEG Streaming multipart/x-mixed-replace. By default MultipartWriter.write() appends closing --boundary-- and breaks your content. Providing close_boundary = False prevents this.:

my_boundary = 'some-boundary'
response = web.StreamResponse(
    status=200,
    reason='OK',
    headers={
        'Content-Type': 'multipart/x-mixed-replace;boundary={}'.format(my_boundary)
    }
)
while True:
    frame = get_jpeg_frame()
    with MultipartWriter('image/jpeg', boundary=my_boundary) as mpwriter:
        mpwriter.append(frame, {
            'Content-Type': 'image/jpeg'
        })
        await mpwriter.write(response, close_boundary=False)
    await response.drain()
Hacking Multipart

The Internet is full of terror and sometimes you may find a server which implements multipart support in strange ways when an oblivious solution does not work.

For instance, is server used cgi.FieldStorage then you have to ensure that no body part contains a Content-Length header:

for part in mpwriter:
    part.headers.pop(aiohttp.hdrs.CONTENT_LENGTH, None)

On the other hand, some server may require to specify Content-Length for the whole multipart request. aiohttp does not do that since it sends multipart using chunked transfer encoding by default. To overcome this issue, you have to serialize a MultipartWriter by our own in the way to calculate its size:

class Writer:
    def __init__(self):
        self.buffer = bytearray()

    async def write(self, data):
        self.buffer.extend(data)

writer = Writer()
await mpwriter.write(writer)
await aiohttp.post('http://example.com',
                   data=writer.buffer, headers=mpwriter.headers)

Sometimes the server response may not be well formed: it may or may not contains nested parts. For instance, we request a resource which returns JSON documents with the files attached to it. If the document has any attachments, they are returned as a nested multipart. If it has not it responds as plain body parts:

CONTENT-TYPE: multipart/mixed; boundary=--:

--:
CONTENT-TYPE: application/json

{"_id": "foo"}
--:
CONTENT-TYPE: multipart/related; boundary=----:

----:
CONTENT-TYPE: application/json

{"_id": "bar"}
----:
CONTENT-TYPE: text/plain
CONTENT-DISPOSITION: attachment; filename=bar.txt

bar! bar! bar!
----:--
--:
CONTENT-TYPE: application/json

{"_id": "boo"}
--:
CONTENT-TYPE: multipart/related; boundary=----:

----:
CONTENT-TYPE: application/json

{"_id": "baz"}
----:
CONTENT-TYPE: text/plain
CONTENT-DISPOSITION: attachment; filename=baz.txt

baz! baz! baz!
----:--
--:--

Reading such kind of data in single stream is possible, but is not clean at all:

result = []
while True:
    part = await reader.next()

    if part is None:
        break

    if isinstance(part, aiohttp.MultipartReader):
        # Fetching files
        while True:
            filepart = await part.next()
            if filepart is None:
                break
            result[-1].append((await filepart.read()))

    else:
        # Fetching document
        result.append([(await part.json())])

Let’s hack a reader in the way to return pairs of document and reader of the related files on each iteration:

class PairsMultipartReader(aiohttp.MultipartReader):

    # keep reference on the original reader
    multipart_reader_cls = aiohttp.MultipartReader

    async def next(self):
        """Emits a tuple of document object (:class:`dict`) and multipart
        reader of the followed attachments (if any).

        :rtype: tuple
        """
        reader = await super().next()

        if self._at_eof:
            return None, None

        if isinstance(reader, self.multipart_reader_cls):
            part = await reader.next()
            doc = await part.json()
        else:
            doc = await reader.json()

        return doc, reader

And this gives us a more cleaner solution:

reader = PairsMultipartReader.from_response(resp)
result = []
while True:
    doc, files_reader = await reader.next()

    if doc is None:
        break

    files = []
    while True:
        filepart = await files_reader.next()
        if file.part is None:
            break
        files.append((await filepart.read()))

    result.append((doc, files))

See also

Multipart reference

Multipart reference

class aiohttp.MultipartResponseWrapper(resp, stream)

Wrapper around the MultipartBodyReader to take care about underlying connection and close it when it needs in.

at_eof()

Returns True when all response data had been read.

Return type

bool

coroutine next()

Emits next multipart reader object.

coroutine release()

Releases the connection gracefully, reading all the content to the void.

class aiohttp.BodyPartReader(boundary, headers, content)

Multipart reader for single body part.

coroutine read(*, decode=False)

Reads body part data.

Parameters

decode (bool) – Decodes data following by encoding method from Content-Encoding header. If it missed data remains untouched

Return type

bytearray

coroutine read_chunk(size=chunk_size)

Reads body part content chunk of the specified size.

Parameters

size (int) – chunk size

Return type

bytearray

coroutine readline()

Reads body part by line by line.

Return type

bytearray

coroutine release()

Like read(), but reads all the data to the void.

Return type

None

coroutine text(*, encoding=None)

Like read(), but assumes that body part contains text data.

Parameters

encoding (str) – Custom text encoding. Overrides specified in charset param of Content-Type header

Return type

str

coroutine json(*, encoding=None)

Like read(), but assumes that body parts contains JSON data.

Parameters

encoding (str) – Custom JSON encoding. Overrides specified in charset param of Content-Type header

coroutine form(*, encoding=None)

Like read(), but assumes that body parts contains form urlencoded data.

Parameters

encoding (str) – Custom form encoding. Overrides specified in charset param of Content-Type header

at_eof()

Returns True if the boundary was reached or False otherwise.

Return type

bool

decode(data)

Decodes data according the specified Content-Encoding or Content-Transfer-Encoding headers value.

Supports gzip, deflate and identity encodings for Content-Encoding header.

Supports base64, quoted-printable, binary encodings for Content-Transfer-Encoding header.

Parameters

data (bytearray) – Data to decode.

Raises

RuntimeError - if encoding is unknown.

Return type

bytes

get_charset(default=None)

Returns charset parameter from Content-Type header or default.

name

A field name specified in Content-Disposition header or None if missed or header is malformed.

Readonly str property.

filename

A field filename specified in Content-Disposition header or None if missed or header is malformed.

Readonly str property.

class aiohttp.MultipartReader(headers, content)

Multipart body reader.

classmethod from_response(cls, response)

Constructs reader instance from HTTP response.

Parameters

response – ClientResponse instance

at_eof()

Returns True if the final boundary was reached or False otherwise.

Return type

bool

coroutine next()

Emits the next multipart body part.

coroutine release()

Reads all the body parts to the void till the final boundary.

coroutine fetch_next_part()

Returns the next body part reader.

class aiohttp.MultipartWriter(subtype='mixed', boundary=None, close_boundary=True)

Multipart body writer.

boundary may be an ASCII-only string.

boundary

The string (str) representation of the boundary.

Changed in version 3.0: Property type was changed from bytes to str.

append(obj, headers=None)

Append an object to writer.

append_payload(payload)

Adds a new body part to multipart writer.

append_json(obj, headers=None)

Helper to append JSON part.

append_form(obj, headers=None)

Helper to append form urlencoded part.

size

Size of the payload.

coroutine write(writer, close_boundary=True)

Write body.

Parameters

close_boundary (bool) – The (bool) that will emit boundary closing. You may want to disable when streaming (multipart/x-mixed-replace)

New in version 3.4: Support close_boundary argument.

Streaming API

aiohttp uses streams for retrieving BODIES: aiohttp.web.Request.content and aiohttp.ClientResponse.content are properties with stream API.

class aiohttp.StreamReader

The reader from incoming stream.

User should never instantiate streams manually but use existing aiohttp.web.Request.content and aiohttp.ClientResponse.content properties for accessing raw BODY data.

Reading Methods
coroutine StreamReader.read(n=- 1)

Read up to n bytes. If n is not provided, or set to -1, read until EOF and return all read bytes.

If the EOF was received and the internal buffer is empty, return an empty bytes object.

Parameters

n (int) – how many bytes to read, -1 for the whole stream.

Return bytes

the given data

coroutine StreamReader.readany()

Read next data portion for the stream.

Returns immediately if internal buffer has a data.

Return bytes

the given data

coroutine StreamReader.readexactly(n)

Read exactly n bytes.

Raise an asyncio.IncompleteReadError if the end of the stream is reached before n can be read, the asyncio.IncompleteReadError.partial attribute of the exception contains the partial read bytes.

Parameters

n (int) – how many bytes to read.

Return bytes

the given data

coroutine StreamReader.readline()

Read one line, where “line” is a sequence of bytes ending with \n.

If EOF is received, and \n was not found, the method will return the partial read bytes.

If the EOF was received and the internal buffer is empty, return an empty bytes object.

Return bytes

the given line

coroutine StreamReader.readchunk()

Read a chunk of data as it was received by the server.

Returns a tuple of (data, end_of_HTTP_chunk).

When chunked transfer encoding is used, end_of_HTTP_chunk is a bool indicating if the end of the data corresponds to the end of a HTTP chunk, otherwise it is always False.

Return tuple[bytes, bool]

a chunk of data and a bool that is True when the end of the returned chunk corresponds to the end of a HTTP chunk.

Asynchronous Iteration Support

Stream reader supports asynchronous iteration over BODY.

By default it iterates over lines:

async for line in response.content:
    print(line)

Also there are methods for iterating over data chunks with maximum size limit and over any available data.

async-for StreamReader.iter_chunked(n)

Iterates over data chunks with maximum size limit:

async for data in response.content.iter_chunked(1024):
    print(data)
async-for StreamReader.iter_any()

Iterates over data chunks in order of intaking them into the stream:

async for data in response.content.iter_any():
    print(data)
async-for StreamReader.iter_chunks()

Iterates over data chunks as received from the server:

async for data, _ in response.content.iter_chunks():
    print(data)

If chunked transfer encoding is used, the original http chunks formatting can be retrieved by reading the second element of returned tuples:

buffer = b""

async for data, end_of_http_chunk in response.content.iter_chunks():
    buffer += data
    if end_of_http_chunk:
        print(buffer)
        buffer = b""
Helpers
StreamReader.exception()

Get the exception occurred on data reading.

aiohttp.is_eof()

Return True if EOF was reached.

Internal buffer may be not empty at the moment.

See also

StreamReader.at_eof()

StreamReader.at_eof()

Return True if the buffer is empty and EOF was reached.

StreamReader.read_nowait(n=None)

Returns data from internal buffer if any, empty bytes object otherwise.

Raises RuntimeError if other coroutine is waiting for stream.

Parameters

n (int) – how many bytes to read, -1 for the whole internal buffer.

Return bytes

the given data

StreamReader.unread_data(data)

Rollback reading some data from stream, inserting it to buffer head.

Parameters

data (bytes) – data to push back into the stream.

Warning

The method does not wake up waiters.

E.g. read() will not be resumed.

coroutine aiohttp.wait_eof()

Wait for EOF. The given data may be accessible by upcoming read calls.

Signals

Signal is a list of registered asynchronous callbacks.

The signal’s life-cycle has two stages: after creation its content could be filled by using standard list operations: sig.append() etc.

After sig.freeze() call the signal is frozen: adding, removing and dropping callbacks are forbidden.

The only available operation is calling previously registered callbacks by await sig.send(data).

For concrete usage examples see signals in aiohttp.web chapter.

Changed in version 3.0: sig.send() call is forbidden for non-frozen signal.

Support for regular (non-async) callbacks is dropped. All callbacks should be async functions.

class aiohttp.Signal

The signal, implements collections.abc.MutableSequence interface.

coroutine send(*args, **kwargs)

Call all registered callbacks one by one starting from the begin of list.

frozen

True if freeze() was called, read-only property.

freeze()

Freeze the list. After the call any content modification is forbidden.

Common data structures

Common data structures used by aiohttp internally.

FrozenList

A list-like structure which implements collections.abc.MutableSequence.

The list is mutable unless FrozenList.freeze() is called, after that the list modification raises RuntimeError.

class aiohttp.FrozenList(items)

Construct a new non-frozen list from items iterable.

The list implements all collections.abc.MutableSequence methods plus two additional APIs.

frozen

A read-only property, True is the list is frozen (modifications are forbidden).

freeze()

Freeze the list. There is no way to thaw it back.

ChainMapProxy

An immutable version of collections.ChainMap. Internally the proxy is a list of mappings (dictionaries), if the requested key is not present in the first mapping the second is looked up and so on.

The class supports collections.abc.Mapping interface.

class aiohttp.ChainMapProxy(maps)

Create a new chained mapping proxy from a list of mappings (maps).

New in version 3.2.

WebSocket utilities

class aiohttp.WSCloseCode

An IntEnum for keeping close message code.

OK

A normal closure, meaning that the purpose for which the connection was established has been fulfilled.

GOING_AWAY

An endpoint is “going away”, such as a server going down or a browser having navigated away from a page.

PROTOCOL_ERROR

An endpoint is terminating the connection due to a protocol error.

UNSUPPORTED_DATA

An endpoint is terminating the connection because it has received a type of data it cannot accept (e.g., an endpoint that understands only text data MAY send this if it receives a binary message).

INVALID_TEXT

An endpoint is terminating the connection because it has received data within a message that was not consistent with the type of the message (e.g., non-UTF-8 RFC 3629 data within a text message).

POLICY_VIOLATION

An endpoint is terminating the connection because it has received a message that violates its policy. This is a generic status code that can be returned when there is no other more suitable status code (e.g., unsupported_data or message_too_big) or if there is a need to hide specific details about the policy.

MESSAGE_TOO_BIG

An endpoint is terminating the connection because it has received a message that is too big for it to process.

MANDATORY_EXTENSION

An endpoint (client) is terminating the connection because it has expected the server to negotiate one or more extension, but the server did not return them in the response message of the WebSocket handshake. The list of extensions that are needed should appear in the /reason/ part of the Close frame. Note that this status code is not used by the server, because it can fail the WebSocket handshake instead.

INTERNAL_ERROR

A server is terminating the connection because it encountered an unexpected condition that prevented it from fulfilling the request.

SERVICE_RESTART

The service is restarted. a client may reconnect, and if it chooses to do, should reconnect using a randomized delay of 5-30s.

TRY_AGAIN_LATER

The service is experiencing overload. A client should only connect to a different IP (when there are multiple for the target) or reconnect to the same IP upon user action.

class aiohttp.WSMsgType

An IntEnum for describing WSMessage type.

CONTINUATION

A mark for continuation frame, user will never get the message with this type.

TEXT

Text message, the value has str type.

BINARY

Binary message, the value has bytes type.

PING

Ping frame (sent by client peer).

PONG

Pong frame, answer on ping. Sent by server peer.

CLOSE

Close frame.

CLOSED FRAME

Actually not frame but a flag indicating that websocket was closed.

ERROR

Actually not frame but a flag indicating that websocket was received an error.

class aiohttp.WSMessage

Websocket message, returned by .receive() calls.

type

Message type, WSMsgType instance.

data

Message payload.

  1. str for WSMsgType.TEXT messages.

  2. bytes for WSMsgType.BINARY messages.

  3. WSCloseCode for WSMsgType.CLOSE messages.

  4. bytes for WSMsgType.PING messages.

  5. bytes for WSMsgType.PONG messages.

extra

Additional info, str.

Makes sense only for WSMsgType.CLOSE messages, contains optional message description.

json(*, loads=json.loads)

Returns parsed JSON data.

Parameters

loads – optional JSON decoder function.

FAQ

Are there plans for an @app.route decorator like in Flask?

As of aiohttp 2.3, RouteTableDef provides an API similar to Flask’s @app.route. See Alternative ways for registering routes.

Unlike Flask’s @app.route, RouteTableDef does not require an app in the module namespace (which often leads to circular imports).

Instead, a RouteTableDef is decoupled from an application instance:

routes = web.RouteTableDef()

@routes.get('/get')
async def handle_get(request):
    ...


@routes.post('/post')
async def handle_post(request):
    ...

app.router.add_routes(routes)

Does aiohttp have a concept like Flask’s “blueprint” or Django’s “app”?

If you’re writing a large application, you may want to consider using nested applications, which are similar to Flask’s “blueprints” or Django’s “apps”.

See: Nested applications.

How do I create a route that matches urls with a given prefix?

You can do something like the following:

app.router.add_route('*', '/path/to/{tail:.+}', sink_handler)

The first argument, *, matches any HTTP method (GET, POST, OPTIONS, etc). The second argument matches URLS with the desired prefix. The third argument is the handler function.

Where do I put my database connection so handlers can access it?

aiohttp.web.Application object supports the dict interface and provides a place to store your database connections or any other resource you want to share between handlers.

async def go(request):
    db = request.app['db']
    cursor = await db.cursor()
    await cursor.execute('SELECT 42')
    # ...
    return web.Response(status=200, text='ok')


async def init_app(loop):
    app = Application(loop=loop)
    db = await create_connection(user='user', password='123')
    app['db'] = db
    app.router.add_get('/', go)
    return app

How can middleware store data for web handlers to use?

Both aiohttp.web.Request and aiohttp.web.Application support the dict interface.

Therefore, data may be stored inside a request object.

async def handler(request):
    request['unique_key'] = data

See https://github.com/aio-libs/aiohttp_session code for an example. The aiohttp_session.get_session(request) method uses SESSION_KEY for saving request-specific session information.

As of aiohttp 3.0, all response objects are dict-like structures as well.

Can a handler receive incoming events from different sources in parallel?

Yes.

As an example, we may have two event sources:

  1. WebSocket for events from an end user

  2. Redis PubSub for events from other parts of the application

The most native way to handle this is to create a separate task for PubSub handling.

Parallel aiohttp.web.WebSocketResponse.receive() calls are forbidden; a single task should perform WebSocket reading. However, other tasks may use the same WebSocket object for sending data to peers.

async def handler(request):

    ws = web.WebSocketResponse()
    await ws.prepare(request)
    task = request.app.loop.create_task(
        read_subscription(ws,
                          request.app['redis']))
    try:
        async for msg in ws:
            # handle incoming messages
            # use ws.send_str() to send data back
            ...

    finally:
        task.cancel()

async def read_subscription(ws, redis):
    channel, = await redis.subscribe('channel:1')

    try:
        async for msg in channel.iter():
            answer = process_the_message(msg)  # your function here
            await ws.send_str(answer)
    finally:
        await redis.unsubscribe('channel:1')

How do I programmatically close a WebSocket server-side?

Let’s say we have an application with two endpoints:

  1. /echo a WebSocket echo server that authenticates the user

  2. /logout_user that, when invoked, closes all open WebSockets for that user.

One simple solution is to keep a shared registry of WebSocket responses for a user in the aiohttp.web.Application instance and call aiohttp.web.WebSocketResponse.close() on all of them in /logout_user handler:

async def echo_handler(request):

    ws = web.WebSocketResponse()
    user_id = authenticate_user(request)
    await ws.prepare(request)
    request.app['websockets'][user_id].add(ws)
    try:
        async for msg in ws:
            ws.send_str(msg.data)
    finally:
        request.app['websockets'][user_id].remove(ws)

    return ws


async def logout_handler(request):

    user_id = authenticate_user(request)

    ws_closers = [ws.close()
                  for ws in request.app['websockets'][user_id]
                  if not ws.closed]

    # Watch out, this will keep us from returing the response
    # until all are closed
    ws_closers and await asyncio.gather(*ws_closers)

    return web.Response(text='OK')


def main():
    loop = asyncio.get_event_loop()
    app = web.Application(loop=loop)
    app.router.add_route('GET', '/echo', echo_handler)
    app.router.add_route('POST', '/logout', logout_handler)
    app['websockets'] = defaultdict(set)
    web.run_app(app, host='localhost', port=8080)

How do I make a request from a specific IP address?

If your system has several IP interfaces, you may choose one which will be used used to bind a socket locally:

conn = aiohttp.TCPConnector(local_addr=('127.0.0.1', 0), loop=loop)
async with aiohttp.ClientSession(connector=conn) as session:
    ...

See also

aiohttp.TCPConnector and local_addr parameter.

What is the API stability and deprecation policy?

aiohttp follows strong Semantic Versioning (SemVer).

Obsolete attributes and methods are marked as deprecated in the documentation and raise DeprecationWarning upon usage.

Assume aiohttp X.Y.Z where X is major version, Y is minor version and Z is bugfix number.

For example, if the latest released version is aiohttp==3.0.6:

3.0.7 fixes some bugs but have no new features.

3.1.0 introduces new features and can deprecate some API but never remove it, also all bug fixes from previous release are merged.

4.0.0 removes all deprecations collected from 3.Y versions except deprecations from the last 3.Y release. These deprecations will be removed by 5.0.0.

Unfortunately we may have to break these rules when a security vulnerability is found. If a security problem cannot be fixed without breaking backward compatibility, a bugfix release may break compatibility. This is unlikely, but possible.

All backward incompatible changes are explicitly marked in the changelog.

How do I enable gzip compression globally for my entire application?

It’s impossible. Choosing what to compress and what not to compress is is a tricky matter.

If you need global compression, write a custom middleware. Or enable compression in NGINX (you are deploying aiohttp behind reverse proxy, right?).

How do I manage a ClientSession within a web server?

aiohttp.ClientSession should be created once for the lifetime of the server in order to benefit from connection pooling.

Sessions save cookies internally. If you don’t need cookie processing, use aiohttp.DummyCookieJar. If you need separate cookies for different http calls but process them in logical chains, use a single aiohttp.TCPConnector with separate client sessions and connector_owner=False.

How do I access database connections from a subapplication?

Restricting access from subapplication to main (or outer) app is a deliberate choice.

A subapplication is an isolated unit by design. If you need to share a database object, do it explicitly:

subapp['db'] = mainapp['db']
mainapp.add_subapp('/prefix', subapp)

How do I perform operations in a request handler after sending the response?

Middlewares can be written to handle post-response operations, but they run after every request. You can explicitly send the response by calling aiohttp.web.Response.write_eof(), which starts sending before the handler returns, giving you a chance to execute follow-up operations:

def ping_handler(request):
    """Send PONG and increase DB counter."""

    # explicitly send the response
    resp = web.json_response({'message': 'PONG'})
    await resp.prepare(request)
    await resp.write_eof()

    # increase the pong count
    APP['db'].inc_pong()

    return resp

A aiohttp.web.Response object must be returned. This is required by aiohttp web contracts, even though the response already been sent.

How do I make sure my custom middleware response will behave correctly?

Sometimes your middleware handlers might need to send a custom response. This is just fine as long as you always create a new aiohttp.web.Response object when required.

The response object is a Finite State Machine. Once it has been dispatched by the server, it will reach its final state and cannot be used again.

The following middleware will make the server hang, once it serves the second response:

from aiohttp import web

def misbehaved_middleware():
    # don't do this!
    cached = web.Response(status=200, text='Hi, I am cached!')

    @web.middleware
    async def middleware(request, handler):
        # ignoring response for the sake of this example
        _res = handler(request)
        return cached

    return middleware

The rule of thumb is one request, one response.

Why is creating a ClientSession outside of an event loop dangerous?

Short answer is: life-cycle of all asyncio objects should be shorter than life-cycle of event loop.

Full explanation is longer. All asyncio object should be correctly finished/disconnected/closed before event loop shutdown. Otherwise user can get unexpected behavior. In the best case it is a warning about unclosed resource, in the worst case the program just hangs, awaiting for coroutine is never resumed etc.

Consider the following code from mod.py:

import aiohttp

session = aiohttp.ClientSession()

async def fetch(url):
    async with session.get(url) as resp:
        return await resp.text()

The session grabs current event loop instance and stores it in a private variable.

The main module imports the module and installs uvloop (an alternative fast event loop implementation).

main.py:

import asyncio
import uvloop
import mod

asyncio.set_event_loop_policy(uvloop.EventLoopPolicy())
asyncio.run(main())

The code is broken: session is bound to default asyncio loop on import time but the loop is changed after the import by set_event_loop(). As result fetch() call hangs.

To avoid import dependency hell aiohttp encourages creation of ClientSession from async function. The same policy works for web.Application too.

Another use case is unit test writing. Very many test libraries (aiohttp test tools first) creates a new loop instance for every test function execution. It’s done for sake of tests isolation. Otherwise pending activity (timers, network packets etc.) from previous test may interfere with current one producing very cryptic and unstable test failure.

Note: class variables are hidden globals actually. The following code has the same problem as mod.py example, session variable is the hidden global object:

class A:
    session = aiohttp.ClientSession()

    async def fetch(self, url):
        async with session.get(url) as resp:
            return await resp.text()

Miscellaneous

Helpful pages.

Essays

Router refactoring in 0.21
Rationale

First generation (v1) of router has mapped (method, path) pair to web-handler. Mapping is named route. Routes used to have unique names if any.

The main mistake with the design is coupling the route to (method, path) pair while really URL construction operates with resources (location is a synonym). HTTP method is not part of URI but applied on sending HTTP request only.

Having different route names for the same path is confusing. Moreover named routes constructed for the same path should have unique non overlapping names which is cumbersome is certain situations.

From other side sometimes it’s desirable to bind several HTTP methods to the same web handler. For v1 router it can be solved by passing ‘*’ as HTTP method. Class based views require ‘*’ method also usually.

Implementation

The change introduces resource as first class citizen:

resource = router.add_resource('/path/{to}', name='name')

Resource has a path (dynamic or constant) and optional name.

The name is unique in router context.

Resource has routes.

Route corresponds to HTTP method and web-handler for the method:

route = resource.add_route('GET', handler)

User still may use wildcard for accepting all HTTP methods (maybe we will add something like resource.add_wildcard(handler) later).

Since names belongs to resources now app.router['name'] returns a resource instance instead of aiohttp.web.Route.

resource has .url() method, so app.router['name'].url(parts={'a': 'b'}, query={'arg': 'param'}) still works as usual.

The change allows to rewrite static file handling and implement nested applications as well.

Decoupling of HTTP location and HTTP method makes life easier.

Backward compatibility

The refactoring is 99% compatible with previous implementation.

99% means all example and the most of current code works without modifications but we have subtle API backward incompatibles.

app.router['name'] returns a aiohttp.web.BaseResource instance instead of aiohttp.web.Route but resource has the same resource.url(...) most useful method, so end user should feel no difference.

route.match(...) is not supported anymore, use aiohttp.web.AbstractResource.resolve() instead.

app.router.add_route(method, path, handler, name='name') now is just shortcut for:

resource = app.router.add_resource(path, name=name)
route = resource.add_route(method, handler)
return route

app.router.register_route(...) is still supported, it creates aiohttp.web.ResourceAdapter for every call (but it’s deprecated now).

What’s new in aiohttp 1.1
YARL and URL encoding

Since aiohttp 1.1 the library uses yarl for URL processing.

New API

yarl.URL gives handy methods for URL operations etc.

Client API still accepts str everywhere url is used, e.g. session.get('http://example.com') works as well as session.get(yarl.URL('http://example.com')).

Internal API has been switched to yarl.URL. aiohttp.CookieJar accepts URL instances only.

On server side has added web.Request.url and web.Request.rel_url properties for representing relative and absolute request’s URL.

URL using is the recommended way, already existed properties for retrieving URL parts are deprecated and will be eventually removed.

Redirection web exceptions accepts yarl.URL as location parameter. str is still supported and will be supported forever.

Reverse URL processing for router has been changed.

The main API is aiohttp.web.Request.url_for(name, **kwargs) which returns a yarl.URL instance for named resource. It does not support query args but adding args is trivial: request.url_for('named_resource', param='a').with_query(arg='val').

The method returns a relative URL, absolute URL may be constructed by request.url.join(request.url_for(...) call.

URL encoding

YARL encodes all non-ASCII symbols on yarl.URL creation.

Thus URL('https://www.python.org/путь') becomes 'https://www.python.org/%D0%BF%D1%83%D1%82%D1%8C'.

On filling route table it’s possible to use both non-ASCII and percent encoded paths:

app.router.add_get('/путь', handler)

and:

app.router.add_get('/%D0%BF%D1%83%D1%82%D1%8C', handler)

are the same. Internally '/путь' is converted into percent-encoding representation.

Route matching also accepts both URL forms: raw and encoded by converting the route pattern to canonical (encoded) form on route registration.

Sub-Applications

Sub applications are designed for solving the problem of the big monolithic code base. Let’s assume we have a project with own business logic and tools like administration panel and debug toolbar.

Administration panel is a separate application by its own nature but all toolbar URLs are served by prefix like /admin.

Thus we’ll create a totally separate application named admin and connect it to main app with prefix:

admin = web.Application()
# setup admin routes, signals and middlewares

app.add_subapp('/admin/', admin)

Middlewares and signals from app and admin are chained.

It means that if URL is '/admin/something' middlewares from app are applied first and admin.middlewares are the next in the call chain.

The same is going for on_response_prepare signal – the signal is delivered to both top level app and admin if processing URL is routed to admin sub-application.

Common signals like on_startup, on_shutdown and on_cleanup are delivered to all registered sub-applications. The passed parameter is sub-application instance, not top-level application.

Third level sub-applications can be nested into second level ones – there are no limitation for nesting level.

Url reversing

Url reversing for sub-applications should generate urls with proper prefix.

But for getting URL sub-application’s router should be used:

admin = web.Application()
admin.add_get('/resource', handler, name='name')

app.add_subapp('/admin/', admin)

url = admin.router['name'].url_for()

The generated url from example will have a value URL('/admin/resource').

Application freezing

Application can be used either as main app (app.make_handler()) or as sub-application – not both cases at the same time.

After connecting application by .add_subapp() call or starting serving web-server as toplevel application the application is frozen.

It means that registering new routes, signals and middlewares is forbidden. Changing state (app['name'] = 'value') of frozen application is deprecated and will be eventually removed.

Migration to 2.x
Client
chunking

aiohttp does not support custom chunking sizes. It is up to the developer to decide how to chunk data streams. If chunking is enabled, aiohttp encodes the provided chunks in the “Transfer-encoding: chunked” format.

aiohttp does not enable chunked encoding automatically even if a transfer-encoding header is supplied: chunked has to be set explicitly. If chunked is set, then the Transfer-encoding and content-length headers are disallowed.

compression

Compression has to be enabled explicitly with the compress parameter. If compression is enabled, adding a content-encoding header is not allowed. Compression also enables the chunked transfer-encoding. Compression can not be combined with a Content-Length header.

Client Connector

  1. By default a connector object manages a total number of concurrent connections. This limit was a per host rule in version 1.x. In 2.x, the limit parameter defines how many concurrent connection connector can open and a new limit_per_host parameter defines the limit per host. By default there is no per-host limit.

  2. BaseConnector.close is now a normal function as opposed to coroutine in version 1.x

  3. BaseConnector.conn_timeout was moved to ClientSession

ClientResponse.release

Internal implementation was significantly redesigned. It is not required to call release on the response object. When the client fully receives the payload, the underlying connection automatically returns back to pool. If the payload is not fully read, the connection is closed

Client exceptions

Exception hierarchy has been significantly modified. aiohttp now defines only exceptions that covers connection handling and server response misbehaviors. For developer specific mistakes, aiohttp uses python standard exceptions like ValueError or TypeError.

Reading a response content may raise a ClientPayloadError exception. This exception indicates errors specific to the payload encoding. Such as invalid compressed data, malformed chunked-encoded chunks or not enough data that satisfy the content-length header.

All exceptions are moved from aiohttp.errors module to top level aiohttp module.

New hierarchy of exceptions:

  • ClientError - Base class for all client specific exceptions

    • ClientResponseError - exceptions that could happen after we get response from server

      • WSServerHandshakeError - web socket server response error

        • ClientHttpProxyError - proxy response

    • ClientConnectionError - exceptions related to low-level connection problems

      • ClientOSError - subset of connection errors that are initiated by an OSError exception

        • ClientConnectorError - connector related exceptions

          • ClientProxyConnectionError - proxy connection initialization error

            • ServerConnectionError - server connection related errors

          • ServerDisconnectedError - server disconnected

          • ServerTimeoutError - server operation timeout, (read timeout, etc)

          • ServerFingerprintMismatch - server fingerprint mismatch

    • ClientPayloadError - This exception can only be raised while reading the response payload if one of these errors occurs: invalid compression, malformed chunked encoding or not enough data that satisfy content-length header.

Client payload (form-data)

To unify form-data/payload handling a new Payload system was introduced. It handles customized handling of existing types and provide implementation for user-defined types.

  1. FormData.__call__ does not take an encoding arg anymore and its return value changes from an iterator or bytes to a Payload instance. aiohttp provides payload adapters for some standard types like str, byte, io.IOBase, StreamReader or DataQueue.

  2. a generator is not supported as data provider anymore, streamer can be used instead. For example, to upload data from file:

    @aiohttp.streamer
def file_sender(writer, file_name=None):
      with open(file_name, 'rb') as f:
          chunk = f.read(2**16)
          while chunk:
              yield from writer.write(chunk)
              chunk = f.read(2**16)

# Then you can use `file_sender` like this:

async with session.post('http://httpbin.org/post',
                        data=file_sender(file_name='huge_file')) as resp:
       print(await resp.text())
Various

  1. the encoding parameter is deprecated in ClientSession.request(). Payload encoding is controlled at the payload level. It is possible to specify an encoding for each payload instance.

  2. the version parameter is removed in ClientSession.request() client version can be specified in the ClientSession constructor.

  3. aiohttp.MsgType dropped, use aiohttp.WSMsgType instead.

  4. ClientResponse.url is an instance of yarl.URL class (url_obj is deprecated)

  5. ClientResponse.raise_for_status() raises aiohttp.ClientResponseError exception

  6. ClientResponse.json() is strict about response’s content type. if content type does not match, it raises aiohttp.ClientResponseError exception. To disable content type check you can pass None as content_type parameter.

Server
ServerHttpProtocol and low-level details

Internal implementation was significantly redesigned to provide better performance and support HTTP pipelining. ServerHttpProtocol is dropped, implementation is merged with RequestHandler a lot of low-level api’s are dropped.

Application

  1. Constructor parameter loop is deprecated. Loop is get configured by application runner, run_app function for any of gunicorn workers.

  2. Application.router.add_subapp is dropped, use Application.add_subapp instead

  3. Application.finished is dropped, use Application.cleanup instead

WebRequest and WebResponse

  1. the GET and POST attributes no longer exist. Use the query attribute instead of GET

  2. Custom chunking size is not support WebResponse.chunked - developer is responsible for actual chunking.

  3. Payloads are supported as body. So it is possible to use client response’s content object as body parameter for WebResponse

  4. FileSender api is dropped, it is replaced with more general FileResponse class:

    async def handle(request):
    return web.FileResponse('path-to-file.txt')

  5. WebSocketResponse.protocol is renamed to WebSocketResponse.ws_protocol. WebSocketResponse.protocol is instance of RequestHandler class.

RequestPayloadError

Reading request’s payload may raise a RequestPayloadError exception. The behavior is similar to ClientPayloadError.

WSGI

WSGI support has been dropped, as well as gunicorn wsgi support. We still provide default and uvloop gunicorn workers for web.Application

What’s new in aiohttp 3.0
async/await everywhere

The main change is dropping yield from support and using async/await everywhere. Farewell, Python 3.4.

The minimal supported Python version is 3.5.3 now.

Why not 3.5.0? Because 3.5.3 has a crucial change: asyncio.get_event_loop() returns the running loop instead of default, which may be different, e.g.:

loop = asyncio.new_event_loop()
loop.run_until_complete(f())

Note, asyncio.set_event_loop() was not called and default loop is not equal to actually executed one.

Application Runners

People constantly asked about ability to run aiohttp servers together with other asyncio code, but aiohttp.web.run_app() is blocking synchronous call.

aiohttp had support for starting the application without run_app but the API was very low-level and cumbersome.

Now application runners solve the task in a few lines of code, see Application runners for details.

Client Tracing

Other long awaited feature is tracing client request life cycle to figure out when and why client request spends a time waiting for connection establishment, getting server response headers etc.

Now it is possible by registering special signal handlers on every request processing stage. Client Tracing provides more info about the feature.

HTTPS support

Unfortunately asyncio has a bug with checking SSL certificates for non-ASCII site DNS names, e.g. https://историк.рф or https://雜草工作室.香港.

The bug has been fixed in upcoming Python 3.7 only (the change requires breaking backward compatibility in ssl API).

aiohttp installs a fix for older Python versions (3.5 and 3.6).

Dropped obsolete API

A switch to new major version is a great chance for dropping already deprecated features.

The release dropped a lot, see Changelog for details.

All removals was already marked as deprecated or related to very low level implementation details.

If user code did not raise DeprecationWarning it is compatible with aiohttp 3.0 most likely.

Summary

Enjoy aiohttp 3.0 release!

The full change log is here: Changelog.

Glossary

aiodns

DNS resolver for asyncio.

https://pypi.python.org/pypi/aiodns

asyncio

The library for writing single-threaded concurrent code using coroutines, multiplexing I/O access over sockets and other resources, running network clients and servers, and other related primitives.

Reference implementation of PEP 3156

https://pypi.python.org/pypi/asyncio/

callable

Any object that can be called. Use callable() to check that.

cchardet

cChardet is high speed universal character encoding detector - binding to charsetdetect.

https://pypi.python.org/pypi/cchardet/

chardet

The Universal Character Encoding Detector

https://pypi.python.org/pypi/chardet/

gunicorn

Gunicorn ‘Green Unicorn’ is a Python WSGI HTTP Server for UNIX.

http://gunicorn.org/

IDNA

An Internationalized Domain Name in Applications (IDNA) is an industry standard for encoding Internet Domain Names that contain in whole or in part, in a language-specific script or alphabet, such as Arabic, Chinese, Cyrillic, Tamil, Hebrew or the Latin alphabet-based characters with diacritics or ligatures, such as French. These writing systems are encoded by computers in multi-byte Unicode. Internationalized domain names are stored in the Domain Name System as ASCII strings using Punycode transcription.

keep-alive

A technique for communicating between HTTP client and server when connection is not closed after sending response but kept open for sending next request through the same socket.

It makes communication faster by getting rid of connection establishment for every request.

nginx

Nginx [engine x] is an HTTP and reverse proxy server, a mail proxy server, and a generic TCP/UDP proxy server.

https://nginx.org/en/

percent-encoding

A mechanism for encoding information in a Uniform Resource Locator (URL) if URL parts don’t fit in safe characters space.

requests

Currently the most popular synchronous library to make HTTP requests in Python.

https://requests.readthedocs.io

requoting

Applying percent-encoding to non-safe symbols and decode percent encoded safe symbols back.

According to RFC 3986 allowed path symbols are:

allowed       = unreserved / pct-encoded / sub-delims
                / ":" / "@" / "/"

pct-encoded   = "%" HEXDIG HEXDIG

unreserved    = ALPHA / DIGIT / "-" / "." / "_" / "~"

sub-delims    = "!" / "$" / "&" / "'" / "(" / ")"
                / "*" / "+" / "," / ";" / "="
resource

A concept reflects the HTTP path, every resource corresponds to URI.

May have a unique name.

Contains route's for different HTTP methods.

route

A part of resource, resource’s path coupled with HTTP method.

web-handler

An endpoint that returns HTTP response.

websocket

A protocol providing full-duplex communication channels over a single TCP connection. The WebSocket protocol was standardized by the IETF as RFC 6455

yarl

A library for operating with URL objects.

https://pypi.python.org/pypi/yarl

Changelog

3.7.4 (2021-02-25)
Bugfixes
3.7.3 (2020-11-18)
Features

  • Use Brotli instead of brotlipy #3803

  • Made exceptions pickleable. Also changed the repr of some exceptions. #4077

Bugfixes

  • Raise a ClientResponseError instead of an AssertionError for a blank HTTP Reason Phrase. #3532

  • Fix web_middlewares.normalize_path_middleware behavior for patch without slash. #3669

  • Fix overshadowing of overlapped sub-applications prefixes. #3701

  • Make BaseConnector.close() a coroutine and wait until the client closes all connections. Drop deprecated “with Connector():” syntax. #3736

  • Reset the sock_read timeout each time data is received for a aiohttp.client response. #3808

  • Fixed type annotation for add_view method of UrlDispatcher to accept any subclass of View #3880

  • Fixed querying the address families from DNS that the current host supports. #5156

  • Change return type of MultipartReader.__aiter__() and BodyPartReader.__aiter__() to AsyncIterator. #5163

  • Provide x86 Windows wheels. #5230

Improved Documentation

  • Add documentation for aiohttp.web.FileResponse. #3958

  • Removed deprecation warning in tracing example docs #3964

  • Fixed wrong “Usage” docstring of aiohttp.client.request. #4603

  • Add aiohttp-pydantic to third party libraries #5228

Misc
3.7.2 (2020-10-27)
Bugfixes

  • Fixed static files handling for loops without .sendfile() support #5149

3.7.1 (2020-10-25)
Bugfixes

  • Fixed a type error caused by the conditional import of Protocol. #5111

  • Server doesn’t send Content-Length for 1xx or 204 #4901

  • Fix run_app typing #4957

  • Always require typing_extensions library. #5107

  • Fix a variable-shadowing bug causing ThreadedResolver.resolve to return the resolved IP as the hostname in each record, which prevented validation of HTTPS connections. #5110

  • Added annotations to all public attributes. #5115

  • Fix flaky test_when_timeout_smaller_second #5116

  • Ensure sending a zero byte file does not throw an exception #5124

  • Fix a bug in web.run_app() about Python version checking on Windows #5127

3.7.0 (2020-10-24)
Features

  • Response headers are now prepared prior to running on_response_prepare hooks, directly before headers are sent to the client. #1958

  • Add a quote_cookie option to CookieJar, a way to skip quotation wrapping of cookies containing special characters. #2571

  • Call AccessLogger.log with the current exception available from sys.exc_info(). #3557

  • web.UrlDispatcher.add_routes and web.Application.add_routes return a list of registered AbstractRoute instances. AbstractRouteDef.register (and all subclasses) return a list of registered resources registered resource. #3866

  • Added properties of default ClientSession params to ClientSession class so it is available for introspection #3882

  • Don’t cancel web handler on peer disconnection, raise OSError on reading/writing instead. #4080

  • Implement BaseRequest.get_extra_info() to access a protocol transports’ extra info. #4189

  • Added ClientSession.timeout property. #4191

  • allow use of SameSite in cookies. #4224

  • Use loop.sendfile() instead of custom implementation if available. #4269

  • Apply SO_REUSEADDR to test server’s socket. #4393

  • Use .raw_host instead of slower .host in client API #4402

  • Allow configuring the buffer size of input stream by passing read_bufsize argument. #4453

  • Pass tests on Python 3.8 for Windows. #4513

  • Add method and url attributes to TraceRequestChunkSentParams and TraceResponseChunkReceivedParams. #4674

  • Add ClientResponse.ok property for checking status code under 400. #4711

  • Don’t ceil timeouts that are smaller than 5 seconds. #4850

  • TCPSite now listens by default on all interfaces instead of just IPv4 when None is passed in as the host. #4894

  • Bump http_parser to 2.9.4 #5070

Bugfixes

  • Fix keepalive connections not being closed in time #3296

  • Fix failed websocket handshake leaving connection hanging. #3380

  • Fix tasks cancellation order on exit. The run_app task needs to be cancelled first for cleanup hooks to run with all tasks intact. #3805

  • Don’t start heartbeat until _writer is set #4062

  • Fix handling of multipart file uploads without a content type. #4089

  • Preserve view handler function attributes across middlewares #4174

  • Fix the string representation of ServerDisconnectedError. #4175

  • Raising RuntimeError when trying to get encoding from not read body #4214

  • Remove warning messages from noop. #4282

  • Raise ClientPayloadError if FormData re-processed. #4345

  • Fix a warning about unfinished task in web_protocol.py #4408

  • Fixed ‘deflate’ compression. According to RFC 2616 now. #4506

  • Fixed OverflowError on platforms with 32-bit time_t #4515

  • Fixed request.body_exists returns wrong value for methods without body. #4528

  • Fix connecting to link-local IPv6 addresses. #4554

  • Fix a problem with connection waiters that are never awaited. #4562

  • Always make sure transport is not closing before reuse a connection.

    Reuse a protocol based on keepalive in headers is unreliable. For example, uWSGI will not support keepalive even it serves a HTTP 1.1 request, except explicitly configure uWSGI with a --http-keepalive option.

    Servers designed like uWSGI could cause aiohttp intermittently raise a ConnectionResetException when the protocol poll runs out and some protocol is reused. #4587

  • Handle the last CRLF correctly even if it is received via separate TCP segment. #4630

  • Fix the register_resource function to validate route name before splitting it so that route name can include python keywords. #4691

  • Improve typing annotations for web.Request, aiohttp.ClientResponse and multipart module. #4736

  • Fix resolver task is not awaited when connector is cancelled #4795

  • Fix a bug “Aiohttp doesn’t return any error on invalid request methods” #4798

  • Fix HEAD requests for static content. #4809

  • Fix incorrect size calculation for memoryview #4890

  • Add HTTPMove to _all__. #4897

  • Fixed the type annotations in the tracing module. #4912

  • Fix typing for multipart __aiter__. #4931

  • Fix for race condition on connections in BaseConnector that leads to exceeding the connection limit. #4936

  • Add forced UTF-8 encoding for application/rdap+json responses. #4938

  • Fix inconsistency between Python and C http request parsers in parsing pct-encoded URL. #4972

  • Fix connection closing issue in HEAD request. #5012

  • Fix type hint on BaseRunner.addresses (from List[str] to List[Any]) #5086

  • Make web.run_app() more responsive to Ctrl+C on Windows for Python < 3.8. It slightly increases CPU load as a side effect. #5098

Improved Documentation

  • Fix example code in client quick-start #3376

  • Updated the docs so there is no contradiction in ttl_dns_cache default value #3512

  • Add ‘Deploy with SSL’ to docs. #4201

  • Change typing of the secure argument on StreamResponse.set_cookie from Optional[str] to Optional[bool] #4204

  • Changes ttl_dns_cache type from int to Optional[int]. #4270

  • Simplify README hello word example and add a documentation page for people coming from requests. #4272

  • Improve some code examples in the documentation involving websockets and starting a simple HTTP site with an AppRunner. #4285

  • Fix typo in code example in Multipart docs #4312

  • Fix code example in Multipart section. #4314

  • Update contributing guide so new contributors read the most recent version of that guide. Update command used to create test coverage reporting. #4810

  • Spelling: Change “canonize” to “canonicalize”. #4986

  • Add aiohttp-sse-client library to third party usage list. #5084

3.6.3 (2020-10-12)
Bugfixes

  • Pin yarl to <1.6.0 to avoid buggy behavior that will be fixed by the next aiohttp release.

3.6.2 (2019-10-09)
Features

  • Made exceptions pickleable. Also changed the repr of some exceptions. #4077

  • Use Iterable type hint instead of Sequence for Application middleware parameter. #4125

Bugfixes

  • Reset the sock_read timeout each time data is received for a aiohttp.ClientResponse. #3808

  • Fix handling of expired cookies so they are not stored in CookieJar. #4063

  • Fix misleading message in the string representation of ClientConnectorError; self.ssl == None means default SSL context, not SSL disabled #4097

  • Don’t clobber HTTP status when using FileResponse. #4106

Improved Documentation

  • Added minimal required logging configuration to logging documentation. #2469

  • Update docs to reflect proxy support. #4100

  • Fix typo in code example in testing docs. #4108

Misc
3.6.1 (2019-09-19)
Features

  • Compatibility with Python 3.8. #4056

Bugfixes

  • correct some exception string format #4068

  • Emit a warning when ssl.OP_NO_COMPRESSION is unavailable because the runtime is built against an outdated OpenSSL. #4052

  • Update multidict requirement to >= 4.5 #4057

Improved Documentation

  • Provide pytest-aiohttp namespace for pytest fixtures in docs. #3723

3.6.0 (2019-09-06)
Features

  • Add support for Named Pipes (Site and Connector) under Windows. This feature requires Proactor event loop to work. #3629

  • Removed Transfer-Encoding: chunked header from websocket responses to be compatible with more http proxy servers. #3798

  • Accept non-GET request for starting websocket handshake on server side. #3980

Bugfixes

  • Raise a ClientResponseError instead of an AssertionError for a blank HTTP Reason Phrase. #3532

  • Fix an issue where cookies would sometimes not be set during a redirect. #3576

  • Change normalize_path_middleware to use 308 redirect instead of 301.

    This behavior should prevent clients from being unable to use PUT/POST methods on endpoints that are redirected because of a trailing slash. #3579

  • Drop the processed task from all_tasks() list early. It prevents logging about a task with unhandled exception when the server is used in conjunction with asyncio.run(). #3587

  • Signal type annotation changed from Signal[Callable[['TraceConfig'], Awaitable[None]]] to Signal[Callable[ClientSession, SimpleNamespace, ...]. #3595

  • Use sanitized URL as Location header in redirects #3614

  • Improve typing annotations for multipart.py along with changes required by mypy in files that references multipart.py. #3621

  • Close session created inside aiohttp.request when unhandled exception occurs #3628

  • Cleanup per-chunk data in generic data read. Memory leak fixed. #3631

  • Use correct type for add_view and family #3633

  • Fix _keepalive field in __slots__ of RequestHandler. #3644

  • Properly handle ConnectionResetError, to silence the “Cannot write to closing transport” exception when clients disconnect uncleanly. #3648

  • Suppress pytest warnings due to test_utils classes #3660

  • Fix overshadowing of overlapped sub-application prefixes. #3701

  • Fixed return type annotation for WSMessage.json() #3720

  • Properly expose TooManyRedirects publicly as documented. #3818

  • Fix missing brackets for IPv6 in proxy CONNECT request #3841

  • Make the signature of aiohttp.test_utils.TestClient.request match asyncio.ClientSession.request according to the docs #3852

  • Use correct style for re-exported imports, makes mypy --strict mode happy. #3868

  • Fixed type annotation for add_view method of UrlDispatcher to accept any subclass of View #3880

  • Made cython HTTP parser set Reason-Phrase of the response to an empty string if it is missing. #3906

  • Add URL to the string representation of ClientResponseError. #3959

  • Accept istr keys in LooseHeaders type hints. #3976

  • Fixed race conditions in _resolve_host caching and throttling when tracing is enabled. #4013

  • For URLs like “unix://localhost/…” set Host HTTP header to “localhost” instead of “localhost:None”. #4039

Improved Documentation

  • Modify documentation for Background Tasks to remove deprecated usage of event loop. #3526

  • use if __name__ == '__main__': in server examples. #3775

  • Update documentation reference to the default access logger. #3783

  • Improve documentation for web.BaseRequest.path and web.BaseRequest.raw_path. #3791

  • Removed deprecation warning in tracing example docs #3964

3.5.4 (2019-01-12)
Bugfixes

  • Fix stream .read() / .readany() / .iter_any() which used to return a partial content only in case of compressed content #3525

3.5.3 (2019-01-10)
Bugfixes

  • Fix type stubs for aiohttp.web.run_app(access_log=True) and fix edge case of access_log=True and the event loop being in debug mode. #3504

  • Fix aiohttp.ClientTimeout type annotations to accept None for fields #3511

  • Send custom per-request cookies even if session jar is empty #3515

  • Restore Linux binary wheels publishing on PyPI

3.5.2 (2019-01-08)
Features

  • FileResponse from web_fileresponse.py uses a ThreadPoolExecutor to work with files asynchronously. I/O based payloads from payload.py uses a ThreadPoolExecutor to work with I/O objects asynchronously. #3313

  • Internal Server Errors in plain text if the browser does not support HTML. #3483

Bugfixes

  • Preserve MultipartWriter parts headers on write. Refactor the way how Payload.headers are handled. Payload instances now always have headers and Content-Type defined. Fix Payload Content-Disposition header reset after initial creation. #3035

  • Log suppressed exceptions in GunicornWebWorker. #3464

  • Remove wildcard imports. #3468

  • Use the same task for app initialization and web server handling in gunicorn workers. It allows to use Python3.7 context vars smoothly. #3471

  • Fix handling of chunked+gzipped response when first chunk does not give uncompressed data #3477

  • Replace collections.MutableMapping with collections.abc.MutableMapping to avoid a deprecation warning. #3480

  • Payload.size type annotation changed from Optional[float] to Optional[int]. #3484

  • Ignore done tasks when cancels pending activities on web.run_app finalization. #3497

Improved Documentation

  • Add documentation for aiohttp.web.HTTPException. #3490

Misc
3.5.1 (2018-12-24)

  • Fix a regression about ClientSession._requote_redirect_url modification in debug mode.

3.5.0 (2018-12-22)
Features

  • The library type annotations are checked in strict mode now.

  • Add support for setting cookies for individual request (#2387)

  • Application.add_domain implementation (#2809)

  • The default app in the request returned by test_utils.make_mocked_request can now have objects assigned to it and retrieved using the [] operator. (#3174)

  • Make request.url accessible when transport is closed. (#3177)

  • Add zlib_executor_size argument to Response constructor to allow compression to run in a background executor to avoid blocking the main thread and potentially triggering health check failures. (#3205)

  • Enable users to set ClientTimeout in aiohttp.request (#3213)

  • Don’t raise a warning if NETRC environment variable is not set and ~/.netrc file doesn’t exist. (#3267)

  • Add default logging handler to web.run_app If the Application.debug` flag is set and the default logger aiohttp.access is used, access logs will now be output using a stderr StreamHandler if no handlers are attached. Furthermore, if the default logger has no log level set, the log level will be set to DEBUG. (#3324)

  • Add method argument to session.ws_connect(). Sometimes server API requires a different HTTP method for WebSocket connection establishment. For example, Docker exec needs POST. (#3378)

  • Create a task per request handling. (#3406)

Bugfixes

  • Enable passing access_log_class via handler_args (#3158)

  • Return empty bytes with end-of-chunk marker in empty stream reader. (#3186)

  • Accept CIMultiDictProxy instances for headers argument in web.Response constructor. (#3207)

  • Don’t uppercase HTTP method in parser (#3233)

  • Make method match regexp RFC-7230 compliant (#3235)

  • Add app.pre_frozen state to properly handle startup signals in sub-applications. (#3237)

  • Enhanced parsing and validation of helpers.BasicAuth.decode. (#3239)

  • Change imports from collections module in preparation for 3.8. (#3258)

  • Ensure Host header is added first to ClientRequest to better replicate browser (#3265)

  • Fix forward compatibility with Python 3.8: importing ABCs directly from the collections module will not be supported anymore. (#3273)

  • Keep the query string by normalize_path_middleware. (#3278)

  • Fix missing parameter raise_for_status for aiohttp.request() (#3290)

  • Bracket IPv6 addresses in the HOST header (#3304)

  • Fix default message for server ping and pong frames. (#3308)

  • Fix tests/test_connector.py typo and tests/autobahn/server.py duplicate loop def. (#3337)

  • Fix false-negative indicator end_of_HTTP_chunk in StreamReader.readchunk function (#3361)

  • Release HTTP response before raising status exception (#3364)

  • Fix task cancellation when sendfile() syscall is used by static file handling. (#3383)

  • Fix stack trace for asyncio.TimeoutError which was not logged, when it is caught in the handler. (#3414)

Improved Documentation

  • Improve documentation of Application.make_handler parameters. (#3152)

  • Fix BaseRequest.raw_headers doc. (#3215)

  • Fix typo in TypeError exception reason in web.Application._handle (#3229)

  • Make server access log format placeholder %b documentation reflect behavior and docstring. (#3307)

Deprecations and Removals

  • Deprecate modification of session.requote_redirect_url (#2278)

  • Deprecate stream.unread_data() (#3260)

  • Deprecated use of boolean in resp.enable_compression() (#3318)

  • Encourage creation of aiohttp public objects inside a coroutine (#3331)

  • Drop dead Connection.detach() and Connection.writer. Both methods were broken for more than 2 years. (#3358)

  • Deprecate app.loop, request.loop, client.loop and connector.loop properties. (#3374)

  • Deprecate explicit debug argument. Use asyncio debug mode instead. (#3381)

  • Deprecate body parameter in HTTPException (and derived classes) constructor. (#3385)

  • Deprecate bare connector close, use async with connector: and await connector.close() instead. (#3417)

  • Deprecate obsolete read_timeout and conn_timeout in ClientSession constructor. (#3438)

Misc

  • #3341, #3351

3.4.4 (2018-09-05)

  • Fix installation from sources when compiling toolkit is not available (#3241)

3.4.3 (2018-09-04)

  • Add app.pre_frozen state to properly handle startup signals in sub-applications. (#3237)

3.4.2 (2018-09-01)

  • Fix iter_chunks type annotation (#3230)

3.4.1 (2018-08-28)

  • Fix empty header parsing regression. (#3218)

  • Fix BaseRequest.raw_headers doc. (#3215)

  • Fix documentation building on ReadTheDocs (#3221)

3.4.0 (2018-08-25)
Features

  • Add type hints (#3049)

  • Add raise_for_status request parameter (#3073)

  • Add type hints to HTTP client (#3092)

  • Minor server optimizations (#3095)

  • Preserve the cause when HTTPException is raised from another exception. (#3096)

  • Add close_boundary option in MultipartWriter.write method. Support streaming (#3104)

  • Added a remove_slash option to the normalize_path_middleware factory. (#3173)

  • The class AbstractRouteDef is importable from aiohttp.web. (#3183)

Bugfixes

  • Prevent double closing when client connection is released before the last data_received() callback. (#3031)

  • Make redirect with normalize_path_middleware work when using url encoded paths. (#3051)

  • Postpone web task creation to connection establishment. (#3052)

  • Fix sock_read timeout. (#3053)

  • When using a server-request body as the data= argument of a client request, iterate over the content with readany instead of readline to avoid Line too long errors. (#3054)

  • fix UrlDispatcher has no attribute add_options, add web.options (#3062)

  • correct filename in content-disposition with multipart body (#3064)

  • Many HTTP proxies has buggy keepalive support. Let’s not reuse connection but close it after processing every response. (#3070)

  • raise 413 “Payload Too Large” rather than raising ValueError in request.post() Add helpful debug message to 413 responses (#3087)

  • Fix StreamResponse equality, now that they are MutableMapping objects. (#3100)

  • Fix server request objects comparison (#3116)

  • Do not hang on 206 Partial Content response with Content-Encoding: gzip (#3123)

  • Fix timeout precondition checkers (#3145)

Improved Documentation

  • Add a new FAQ entry that clarifies that you should not reuse response objects in middleware functions. (#3020)

  • Add FAQ section “Why is creating a ClientSession outside of an event loop dangerous?” (#3072)

  • Fix link to Rambler (#3115)

  • Fix TCPSite documentation on the Server Reference page. (#3146)

  • Fix documentation build configuration file for Windows. (#3147)

  • Remove no longer existing lingering_timeout parameter of Application.make_handler from documentation. (#3151)

  • Mention that app.make_handler is deprecated, recommend to use runners API instead. (#3157)

Deprecations and Removals

  • Drop loop.current_task() from helpers.current_task() (#2826)

  • Drop reader parameter from request.multipart(). (#3090)

3.3.2 (2018-06-12)

  • Many HTTP proxies has buggy keepalive support. Let’s not reuse connection but close it after processing every response. (#3070)

  • Provide vendor source files in tarball (#3076)

3.3.1 (2018-06-05)

  • Fix sock_read timeout. (#3053)

  • When using a server-request body as the data= argument of a client request, iterate over the content with readany instead of readline to avoid Line too long errors. (#3054)

3.3.0 (2018-06-01)
Features

  • Raise ConnectionResetError instead of CancelledError on trying to write to a closed stream. (#2499)

  • Implement ClientTimeout class and support socket read timeout. (#2768)

  • Enable logging when aiohttp.web is used as a program (#2956)

  • Add canonical property to resources (#2968)

  • Forbid reading response BODY after release (#2983)

  • Implement base protocol class to avoid a dependency from internal asyncio.streams.FlowControlMixin (#2986)

  • Cythonize @helpers.reify, 5% boost on macro benchmark (#2995)

  • Optimize HTTP parser (#3015)

  • Implement runner.addresses property. (#3036)

  • Use bytearray instead of a list of bytes in websocket reader. It improves websocket message reading a little. (#3039)

  • Remove heartbeat on closing connection on keepalive timeout. The used hack violates HTTP protocol. (#3041)

  • Limit websocket message size on reading to 4 MB by default. (#3045)

Bugfixes

  • Don’t reuse a connection with the same URL but different proxy/TLS settings (#2981)

  • When parsing the Forwarded header, the optional port number is now preserved. (#3009)

Improved Documentation

  • Make Change Log more visible in docs (#3029)

  • Make style and grammar improvements on the FAQ page. (#3030)

  • Document that signal handlers should be async functions since aiohttp 3.0 (#3032)

Deprecations and Removals

  • Deprecate custom application’s router. (#3021)

Misc

  • #3008, #3011

3.2.1 (2018-05-10)

  • Don’t reuse a connection with the same URL but different proxy/TLS settings (#2981)

3.2.0 (2018-05-06)
Features

  • Raise TooManyRedirects exception when client gets redirected too many times instead of returning last response. (#2631)

  • Extract route definitions into separate web_routedef.py file (#2876)

  • Raise an exception on request body reading after sending response. (#2895)

  • ClientResponse and RequestInfo now have real_url property, which is request url without fragment part being stripped (#2925)

  • Speed up connector limiting (#2937)

  • Added and links property for ClientResponse object (#2948)

  • Add request.config_dict for exposing nested applications data. (#2949)

  • Speed up HTTP headers serialization, server micro-benchmark runs 5% faster now. (#2957)

  • Apply assertions in debug mode only (#2966)

Bugfixes

  • expose property app for TestClient (#2891)

  • Call on_chunk_sent when write_eof takes as a param the last chunk (#2909)

  • A closing bracket was added to __repr__ of resources (#2935)

  • Fix compression of FileResponse (#2942)

  • Fixes some bugs in the limit connection feature (#2964)

Improved Documentation

  • Drop async_timeout usage from documentation for client API in favor of timeout parameter. (#2865)

  • Improve Gunicorn logging documentation (#2921)

  • Replace multipart writer .serialize() method with .write() in documentation. (#2965)

Deprecations and Removals

  • Deprecate Application.make_handler() (#2938)

Misc

  • #2958

3.1.3 (2018-04-12)

  • Fix cancellation broadcast during DNS resolve (#2910)

3.1.2 (2018-04-05)

  • Make LineTooLong exception more detailed about actual data size (#2863)

  • Call on_chunk_sent when write_eof takes as a param the last chunk (#2909)

3.1.1 (2018-03-27)

  • Support asynchronous iterators (and asynchronous generators as well) in both client and server API as request / response BODY payloads. (#2802)

3.1.0 (2018-03-21)

Welcome to aiohttp 3.1 release.

This is an incremental release, fully backward compatible with aiohttp 3.0.

But we have added several new features.

The most visible one is app.add_routes() (an alias for existing app.router.add_routes(). The addition is very important because all aiohttp docs now uses app.add_routes() call in code snippets. All your existing code still do register routes / resource without any warning but you’ve got the idea for a favorite way: noisy app.router.add_get() is replaced by app.add_routes().

The library does not make a preference between decorators:

routes = web.RouteTableDef()

@routes.get('/')
async def hello(request):
    return web.Response(text="Hello, world")

app.add_routes(routes)

and route tables as a list:

async def hello(request):
    return web.Response(text="Hello, world")

app.add_routes([web.get('/', hello)])

Both ways are equal, user may decide basing on own code taste.

Also we have a lot of minor features, bug fixes and documentation updates, see below.

Features

  • Relax JSON content-type checking in the ClientResponse.json() to allow “application/xxx+json” instead of strict “application/json”. (#2206)

  • Bump C HTTP parser to version 2.8 (#2730)

  • Accept a coroutine as an application factory in web.run_app and gunicorn worker. (#2739)

  • Implement application cleanup context (app.cleanup_ctx property). (#2747)

  • Make writer.write_headers a coroutine. (#2762)

  • Add tracking signals for getting request/response bodies. (#2767)

  • Deprecate ClientResponseError.code in favor of .status to keep similarity with response classes. (#2781)

  • Implement app.add_routes() method. (#2787)

  • Implement web.static() and RouteTableDef.static() API. (#2795)

  • Install a test event loop as default by asyncio.set_event_loop(). The change affects aiohttp test utils but backward compatibility is not broken for 99.99% of use cases. (#2804)

  • Refactor ClientResponse constructor: make logically required constructor arguments mandatory, drop _post_init() method. (#2820)

  • Use app.add_routes() in server docs everywhere (#2830)

  • Websockets refactoring, all websocket writer methods are converted into coroutines. (#2836)

  • Provide Content-Range header for Range requests (#2844)

Bugfixes

  • Fix websocket client return EofStream. (#2784)

  • Fix websocket demo. (#2789)

  • Property BaseRequest.http_range now returns a python-like slice when requesting the tail of the range. It’s now indicated by a negative value in range.start rather then in range.stop (#2805)

  • Close a connection if an unexpected exception occurs while sending a request (#2827)

  • Fix firing DNS tracing events. (#2841)

Improved Documentation

  • Document behavior when cchardet detects encodings that are unknown to Python. (#2732)

  • Add diagrams for tracing request life style. (#2748)

  • Drop removed functionality for passing StreamReader as data at client side. (#2793)

3.0.9 (2018-03-14)

  • Close a connection if an unexpected exception occurs while sending a request (#2827)

3.0.8 (2018-03-12)

  • Use asyncio.current_task() on Python 3.7 (#2825)

3.0.7 (2018-03-08)

  • Fix SSL proxy support by client. (#2810)

  • Restore an imperative check in setup.py for python version. The check works in parallel to environment marker. As effect an error about unsupported Python versions is raised even on outdated systems with very old setuptools version installed. (#2813)

3.0.6 (2018-03-05)

  • Add _reuse_address and _reuse_port to web_runner.TCPSite.__slots__. (#2792)

3.0.5 (2018-02-27)

  • Fix InvalidStateError on processing a sequence of two RequestHandler.data_received calls on web server. (#2773)

3.0.4 (2018-02-26)

  • Fix IndexError in HTTP request handling by server. (#2752)

  • Fix MultipartWriter.append* no longer returning part/payload. (#2759)

3.0.3 (2018-02-25)

  • Relax attrs dependency to minimal actually supported version 17.0.3 The change allows to avoid version conflicts with currently existing test tools.

3.0.2 (2018-02-23)
Security Fix

  • Prevent Windows absolute URLs in static files. Paths like /static/D:\path and /static/\\hostname\drive\path are forbidden.

3.0.1

  • Technical release for fixing distribution problems.

3.0.0 (2018-02-12)
Features

  • Speed up the PayloadWriter.write method for large request bodies. (#2126)

  • StreamResponse and Response are now MutableMappings. (#2246)

  • ClientSession publishes a set of signals to track the HTTP request execution. (#2313)

  • Content-Disposition fast access in ClientResponse (#2455)

  • Added support to Flask-style decorators with class-based Views. (#2472)

  • Signal handlers (registered callbacks) should be coroutines. (#2480)

  • Support async with test_client.ws_connect(...) (#2525)

  • Introduce site and application runner as underlying API for web.run_app implementation. (#2530)

  • Only quote multipart boundary when necessary and sanitize input (#2544)

  • Make the aiohttp.ClientResponse.get_encoding method public with the processing of invalid charset while detecting content encoding. (#2549)

  • Add optional configurable per message compression for ClientWebSocketResponse and WebSocketResponse. (#2551)

  • Add hysteresis to StreamReader to prevent flipping between paused and resumed states too often. (#2555)

  • Support .netrc by trust_env (#2581)

  • Avoid to create a new resource when adding a route with the same name and path of the last added resource (#2586)

  • MultipartWriter.boundary is str now. (#2589)

  • Allow a custom port to be used by TestServer (and associated pytest fixtures) (#2613)

  • Add param access_log_class to web.run_app function (#2615)

  • Add ssl parameter to client API (#2626)

  • Fixes performance issue introduced by #2577. When there are no middlewares installed by the user, no additional and useless code is executed. (#2629)

  • Rename PayloadWriter to StreamWriter (#2654)

  • New options reuse_port, reuse_address are added to run_app and TCPSite. (#2679)

  • Use custom classes to pass client signals parameters (#2686)

  • Use attrs library for data classes, replace namedtuple. (#2690)

  • Pytest fixtures renaming, add aiohttp_ prefix (#2578)

  • Add aiohttp- prefix for pytest-aiohttp command line parameters (#2578)

Bugfixes

  • Correctly process upgrade request from server to HTTP2. aiohttp does not support HTTP2 yet, the protocol is not upgraded but response is handled correctly. (#2277)

  • Fix ClientConnectorSSLError and ClientProxyConnectionError for proxy connector (#2408)

  • Fix connector convert OSError to ClientConnectorError (#2423)

  • Fix connection attempts for multiple dns hosts (#2424)

  • Fix writing to closed transport by raising asyncio.CancelledError (#2499)

  • Fix warning in ClientSession.__del__ by stopping to try to close it. (#2523)

  • Fixed race-condition for iterating addresses from the DNSCache. (#2620)

  • Fix default value of access_log_format argument in web.run_app (#2649)

  • Freeze sub-application on adding to parent app (#2656)

  • Do percent encoding for .url_for() parameters (#2668)

  • Correctly process request start time and multiple request/response headers in access log extra (#2641)

Improved Documentation

  • Improve tutorial docs, using literalinclude to link to the actual files. (#2396)

  • Small improvement docs: better example for file uploads. (#2401)

  • Rename from_env to trust_env in client reference. (#2451)

  • Fixed mistype in Proxy Support section where trust_env parameter was used in session.get(“http://python.org”, trust_env=True) method instead of aiohttp.ClientSession constructor as follows: aiohttp.ClientSession(trust_env=True). (#2688)

  • Fix issue with unittest example not compiling in testing docs. (#2717)

Deprecations and Removals

  • Simplify HTTP pipelining implementation (#2109)

  • Drop StreamReaderPayload and DataQueuePayload. (#2257)

  • Drop md5 and sha1 finger-prints (#2267)

  • Drop WSMessage.tp (#2321)

  • Drop Python 3.4 and Python 3.5.0, 3.5.1, 3.5.2. Minimal supported Python versions are 3.5.3 and 3.6.0. yield from is gone, use async/await syntax. (#2343)

  • Drop aiohttp.Timeout and use async_timeout.timeout instead. (#2348)

  • Drop resolve param from TCPConnector. (#2377)

  • Add DeprecationWarning for returning HTTPException (#2415)

  • send_str(), send_bytes(), send_json(), ping() and pong() are genuine async functions now. (#2475)

  • Drop undocumented app.on_pre_signal and app.on_post_signal. Signal handlers should be coroutines, support for regular functions is dropped. (#2480)

  • StreamResponse.drain() is not a part of public API anymore, just use await StreamResponse.write(). StreamResponse.write is converted to async function. (#2483)

  • Drop deprecated slow_request_timeout param and **kwargs` from RequestHandler. (#2500)

  • Drop deprecated resource.url(). (#2501)

  • Remove %u and %l format specifiers from access log format. (#2506)

  • Drop deprecated request.GET property. (#2547)

  • Simplify stream classes: drop ChunksQueue and FlowControlChunksQueue, merge FlowControlStreamReader functionality into StreamReader, drop FlowControlStreamReader name. (#2555)

  • Do not create a new resource on router.add_get(…, allow_head=True) (#2585)

  • Drop access to TCP tuning options from PayloadWriter and Response classes (#2604)

  • Drop deprecated encoding parameter from client API (#2606)

  • Deprecate verify_ssl, ssl_context and fingerprint parameters in client API (#2626)

  • Get rid of the legacy class StreamWriter. (#2651)

  • Forbid non-strings in resource.url_for() parameters. (#2668)

  • Deprecate inheritance from ClientSession and web.Application and custom user attributes for ClientSession, web.Request and web.Application (#2691)

  • Drop resp = await aiohttp.request(…) syntax for sake of async with aiohttp.request(…) as resp:. (#2540)

  • Forbid synchronous context managers for ClientSession and test server/client. (#2362)

Misc

  • #2552

2.3.10 (2018-02-02)

  • Fix 100% CPU usage on HTTP GET and websocket connection just after it (#1955)

  • Patch broken ssl.match_hostname() on Python<3.7 (#2674)

2.3.9 (2018-01-16)

  • Fix colon handing in path for dynamic resources (#2670)

2.3.8 (2018-01-15)

  • Do not use yarl.unquote internal function in aiohttp. Fix incorrectly unquoted path part in URL dispatcher (#2662)

  • Fix compatibility with yarl==1.0.0 (#2662)

2.3.7 (2017-12-27)

  • Fixed race-condition for iterating addresses from the DNSCache. (#2620)

  • Fix docstring for request.host (#2591)

  • Fix docstring for request.remote (#2592)

2.3.6 (2017-12-04)

  • Correct request.app context (for handlers not just middlewares). (#2577)

2.3.5 (2017-11-30)

  • Fix compatibility with pytest 3.3+ (#2565)

2.3.4 (2017-11-29)

  • Make request.app point to proper application instance when using nested applications (with middlewares). (#2550)

  • Change base class of ClientConnectorSSLError to ClientSSLError from ClientConnectorError. (#2563)

  • Return client connection back to free pool on error in connector.connect(). (#2567)

2.3.3 (2017-11-17)

  • Having a ; in Response content type does not assume it contains a charset anymore. (#2197)

  • Use getattr(asyncio, ‘async’) for keeping compatibility with Python 3.7. (#2476)

  • Ignore NotImplementedError raised by set_child_watcher from uvloop. (#2491)

  • Fix warning in ClientSession.__del__ by stopping to try to close it. (#2523)

  • Fixed typo’s in Third-party libraries page. And added async-v20 to the list (#2510)

2.3.2 (2017-11-01)

  • Fix passing client max size on cloning request obj. (#2385)

  • Fix ClientConnectorSSLError and ClientProxyConnectionError for proxy connector. (#2408)

  • Drop generated _http_parser shared object from tarball distribution. (#2414)

  • Fix connector convert OSError to ClientConnectorError. (#2423)

  • Fix connection attempts for multiple dns hosts. (#2424)

  • Fix ValueError for AF_INET6 sockets if a preexisting INET6 socket to the aiohttp.web.run_app function. (#2431)

  • _SessionRequestContextManager closes the session properly now. (#2441)

  • Rename from_env to trust_env in client reference. (#2451)

2.3.1 (2017-10-18)

  • Relax attribute lookup in warning about old-styled middleware (#2340)

2.3.0 (2017-10-18)
Features

  • Add SSL related params to ClientSession.request (#1128)

  • Make enable_compression work on HTTP/1.0 (#1828)

  • Deprecate registering synchronous web handlers (#1993)

  • Switch to multidict 3.0. All HTTP headers preserve casing now but compared in case-insensitive way. (#1994)

  • Improvement for normalize_path_middleware. Added possibility to handle URLs with query string. (#1995)

  • Use towncrier for CHANGES.txt build (#1997)

  • Implement trust_env=True param in ClientSession. (#1998)

  • Added variable to customize proxy headers (#2001)

  • Implement router.add_routes and router decorators. (#2004)

  • Deprecated BaseRequest.has_body in favor of BaseRequest.can_read_body Added BaseRequest.body_exists attribute that stays static for the lifetime of the request (#2005)

  • Provide BaseRequest.loop attribute (#2024)

  • Make _CoroGuard awaitable and fix ClientSession.close warning message (#2026)

  • Responses to redirects without Location header are returned instead of raising a RuntimeError (#2030)

  • Added get_client, get_server, setUpAsync and tearDownAsync methods to AioHTTPTestCase (#2032)

  • Add automatically a SafeChildWatcher to the test loop (#2058)

  • add ability to disable automatic response decompression (#2110)

  • Add support for throttling DNS request, avoiding the requests saturation when there is a miss in the DNS cache and many requests getting into the connector at the same time. (#2111)

  • Use request for getting access log information instead of message/transport pair. Add RequestBase.remote property for accessing to IP of client initiated HTTP request. (#2123)

  • json() raises a ContentTypeError exception if the content-type does not meet the requirements instead of raising a generic ClientResponseError. (#2136)

  • Make the HTTP client able to return HTTP chunks when chunked transfer encoding is used. (#2150)

  • add append_version arg into StaticResource.url and StaticResource.url_for methods for getting an url with hash (version) of the file. (#2157)

  • Fix parsing the Forwarded header. * commas and semicolons are allowed inside quoted-strings; * empty forwarded-pairs (as in for=_1;;by=_2) are allowed; * non-standard parameters are allowed (although this alone could be easily done in the previous parser). (#2173)

  • Don’t require ssl module to run. aiohttp does not require SSL to function. The code paths involved with SSL will only be hit upon SSL usage. Raise RuntimeError if HTTPS protocol is required but ssl module is not present. (#2221)

  • Accept coroutine fixtures in pytest plugin (#2223)

  • Call shutdown_asyncgens before event loop closing on Python 3.6. (#2227)

  • Speed up Signals when there are no receivers (#2229)

  • Raise InvalidURL instead of ValueError on fetches with invalid URL. (#2241)

  • Move DummyCookieJar into cookiejar.py (#2242)

  • run_app: Make print=None disable printing (#2260)

  • Support brotli encoding (generic-purpose lossless compression algorithm) (#2270)

  • Add server support for WebSockets Per-Message Deflate. Add client option to add deflate compress header in WebSockets request header. If calling ClientSession.ws_connect() with compress=15 the client will support deflate compress negotiation. (#2273)

  • Support verify_ssl, fingerprint, ssl_context and proxy_headers by client.ws_connect. (#2292)

  • Added aiohttp.ClientConnectorSSLError when connection fails due ssl.SSLError (#2294)

  • aiohttp.web.Application.make_handler support access_log_class (#2315)

  • Build HTTP parser extension in non-strict mode by default. (#2332)

Bugfixes

  • Clear auth information on redirecting to other domain (#1699)

  • Fix missing app.loop on startup hooks during tests (#2060)

  • Fix issue with synchronous session closing when using ClientSession as an asynchronous context manager. (#2063)

  • Fix issue with CookieJar incorrectly expiring cookies in some edge cases. (#2084)

  • Force use of IPv4 during test, this will make tests run in a Docker container (#2104)

  • Warnings about unawaited coroutines now correctly point to the user’s code. (#2106)

  • Fix issue with IndexError being raised by the StreamReader.iter_chunks() generator. (#2112)

  • Support HTTP 308 Permanent redirect in client class. (#2114)

  • Fix FileResponse sending empty chunked body on 304. (#2143)

  • Do not add Content-Length: 0 to GET/HEAD/TRACE/OPTIONS requests by default. (#2167)

  • Fix parsing the Forwarded header according to RFC 7239. (#2170)

  • Securely determining remote/scheme/host #2171 (#2171)

  • Fix header name parsing, if name is split into multiple lines (#2183)

  • Handle session close during connection, KeyError: <aiohttp.connector._TransportPlaceholder> (#2193)

  • Fixes uncaught TypeError in helpers.guess_filename if name is not a string (#2201)

  • Raise OSError on async DNS lookup if resolved domain is an alias for another one, which does not have an A or CNAME record. (#2231)

  • Fix incorrect warning in StreamReader. (#2251)

  • Properly clone state of web request (#2284)

  • Fix C HTTP parser for cases when status line is split into different TCP packets. (#2311)

  • Fix web.FileResponse overriding user supplied Content-Type (#2317)

Improved Documentation

  • Add a note about possible performance degradation in await resp.text() if charset was not provided by Content-Type HTTP header. Pass explicit encoding to solve it. (#1811)

  • Drop disqus widget from documentation pages. (#2018)

  • Add a graceful shutdown section to the client usage documentation. (#2039)

  • Document connector_owner parameter. (#2072)

  • Update the doc of web.Application (#2081)

  • Fix mistake about access log disabling. (#2085)

  • Add example usage of on_startup and on_shutdown signals by creating and disposing an aiopg connection engine. (#2131)

  • Document encoded=True for yarl.URL, it disables all yarl transformations. (#2198)

  • Document that all app’s middleware factories are run for every request. (#2225)

  • Reflect the fact that default resolver is threaded one starting from aiohttp 1.1 (#2228)

Deprecations and Removals

  • Drop deprecated Server.finish_connections (#2006)

  • Drop %O format from logging, use %b instead. Drop %e format from logging, environment variables are not supported anymore. (#2123)

  • Drop deprecated secure_proxy_ssl_header support (#2171)

  • Removed TimeService in favor of simple caching. TimeService also had a bug where it lost about 0.5 seconds per second. (#2176)

  • Drop unused response_factory from static files API (#2290)

Misc

  • #2013, #2014, #2048, #2094, #2149, #2187, #2214, #2225, #2243, #2248

2.2.5 (2017-08-03)

  • Don’t raise deprecation warning on loop.run_until_complete(client.close()) (#2065)

2.2.4 (2017-08-02)

  • Fix issue with synchronous session closing when using ClientSession as an asynchronous context manager. (#2063)

2.2.3 (2017-07-04)

  • Fix _CoroGuard for python 3.4

2.2.2 (2017-07-03)

  • Allow await session.close() along with yield from session.close()

2.2.1 (2017-07-02)

  • Relax yarl requirement to 0.11+

  • Backport #2026: session.close is a coroutine (#2029)

2.2.0 (2017-06-20)

  • Add doc for add_head, update doc for add_get. (#1944)

  • Fixed consecutive calls for Response.write_eof.

  • Retain method attributes (e.g. __doc__) when registering synchronous handlers for resources. (#1953)

  • Added signal TERM handling in run_app to gracefully exit (#1932)

  • Fix websocket issues caused by frame fragmentation. (#1962)

  • Raise RuntimeError is you try to set the Content Length and enable chunked encoding at the same time (#1941)

  • Small update for unittest_run_loop

  • Use CIMultiDict for ClientRequest.skip_auto_headers (#1970)

  • Fix wrong startup sequence: test server and run_app() are not raise DeprecationWarning now (#1947)

  • Make sure cleanup signal is sent if startup signal has been sent (#1959)

  • Fixed server keep-alive handler, could cause 100% cpu utilization (#1955)

  • Connection can be destroyed before response get processed if await aiohttp.request(..) is used (#1981)

  • MultipartReader does not work with -OO (#1969)

  • Fixed ClientPayloadError with blank Content-Encoding header (#1931)

  • Support deflate encoding implemented in httpbin.org/deflate (#1918)

  • Fix BadStatusLine caused by extra CRLF after POST data (#1792)

  • Keep a reference to ClientSession in response object (#1985)

  • Deprecate undocumented app.on_loop_available signal (#1978)

2.1.0 (2017-05-26)

  • Added support for experimental async-tokio event loop written in Rust https://github.com/PyO3/tokio

  • Write to transport \r\n before closing after keepalive timeout, otherwise client can not detect socket disconnection. (#1883)

  • Only call loop.close in run_app if the user did not supply a loop. Useful for allowing clients to specify their own cleanup before closing the asyncio loop if they wish to tightly control loop behavior

  • Content disposition with semicolon in filename (#917)

  • Added request_info to response object and ClientResponseError. (#1733)

  • Added history to ClientResponseError. (#1741)

  • Allow to disable redirect url re-quoting (#1474)

  • Handle RuntimeError from transport (#1790)

  • Dropped “%O” in access logger (#1673)

  • Added args and kwargs to unittest_run_loop. Useful with other decorators, for example @patch. (#1803)

  • Added iter_chunks to response.content object. (#1805)

  • Avoid creating TimerContext when there is no timeout to allow compatibility with Tornado. (#1817) (#1180)

  • Add proxy_from_env to ClientRequest to read from environment variables. (#1791)

  • Add DummyCookieJar helper. (#1830)

  • Fix assertion errors in Python 3.4 from noop helper. (#1847)

  • Do not unquote + in match_info values (#1816)

  • Use Forwarded, X-Forwarded-Scheme and X-Forwarded-Host for better scheme and host resolution. (#1134)

  • Fix sub-application middlewares resolution order (#1853)

  • Fix applications comparison (#1866)

  • Fix static location in index when prefix is used (#1662)

  • Make test server more reliable (#1896)

  • Extend list of web exceptions, add HTTPUnprocessableEntity, HTTPFailedDependency, HTTPInsufficientStorage status codes (#1920)

2.0.7 (2017-04-12)

  • Fix pypi distribution

  • Fix exception description (#1807)

  • Handle socket error in FileResponse (#1773)

  • Cancel websocket heartbeat on close (#1793)

2.0.6 (2017-04-04)

  • Keeping blank values for request.post() and multipart.form() (#1765)

  • TypeError in data_received of ResponseHandler (#1770)

  • Fix web.run_app not to bind to default host-port pair if only socket is passed (#1786)

2.0.5 (2017-03-29)

  • Memory leak with aiohttp.request (#1756)

  • Disable cleanup closed ssl transports by default.

  • Exception in request handling if the server responds before the body is sent (#1761)

2.0.4 (2017-03-27)

  • Memory leak with aiohttp.request (#1756)

  • Encoding is always UTF-8 in POST data (#1750)

  • Do not add “Content-Disposition” header by default (#1755)

2.0.3 (2017-03-24)

  • Call https website through proxy will cause error (#1745)

  • Fix exception on multipart/form-data post if content-type is not set (#1743)

2.0.2 (2017-03-21)

  • Fixed Application.on_loop_available signal (#1739)

  • Remove debug code

2.0.1 (2017-03-21)

  • Fix allow-head to include name on route (#1737)

  • Fixed AttributeError in WebSocketResponse.can_prepare (#1736)

2.0.0 (2017-03-20)

  • Added json to ClientSession.request() method (#1726)

  • Added session’s raise_for_status parameter, automatically calls raise_for_status() on any request. (#1724)

  • response.json() raises ClientReponseError exception if response’s content type does not match (#1723)

    • Cleanup timer and loop handle on any client exception.

  • Deprecate loop parameter for Application’s constructor

2.0.0rc1 (2017-03-15)

  • Properly handle payload errors (#1710)

  • Added ClientWebSocketResponse.get_extra_info() (#1717)

  • It is not possible to combine Transfer-Encoding and chunked parameter, same for compress and Content-Encoding (#1655)

  • Connector’s limit parameter indicates total concurrent connections. New limit_per_host added, indicates total connections per endpoint. (#1601)

  • Use url’s raw_host for name resolution (#1685)

  • Change ClientResponse.url to yarl.URL instance (#1654)

  • Add max_size parameter to web.Request reading methods (#1133)

  • Web Request.post() stores data in temp files (#1469)

  • Add the allow_head=True keyword argument for add_get (#1618)

  • run_app and the Command Line Interface now support serving over Unix domain sockets for faster inter-process communication.

  • run_app now supports passing a preexisting socket object. This can be useful e.g. for socket-based activated applications, when binding of a socket is done by the parent process.

  • Implementation for Trailer headers parser is broken (#1619)

  • Fix FileResponse to not fall on bad request (range out of file size)

  • Fix FileResponse to correct stream video to Chromes

  • Deprecate public low-level api (#1657)

  • Deprecate encoding parameter for ClientSession.request() method

  • Dropped aiohttp.wsgi (#1108)

  • Dropped version from ClientSession.request() method

  • Dropped websocket version 76 support (#1160)

  • Dropped: aiohttp.protocol.HttpPrefixParser (#1590)

  • Dropped: Servers response’s .started, .start() and .can_start() method (#1591)

  • Dropped: Adding sub app via app.router.add_subapp() is deprecated use app.add_subapp() instead (#1592)

  • Dropped: Application.finish() and Application.register_on_finish() (#1602)

  • Dropped: web.Request.GET and web.Request.POST

  • Dropped: aiohttp.get(), aiohttp.options(), aiohttp.head(), aiohttp.post(), aiohttp.put(), aiohttp.patch(), aiohttp.delete(), and aiohttp.ws_connect() (#1593)

  • Dropped: aiohttp.web.WebSocketResponse.receive_msg() (#1605)

  • Dropped: ServerHttpProtocol.keep_alive_timeout attribute and keep-alive, keep_alive_on, timeout, log constructor parameters (#1606)

  • Dropped: TCPConnector’s` .resolve, .resolved_hosts, .clear_resolved_hosts() attributes and resolve constructor parameter (#1607)

  • Dropped ProxyConnector (#1609)

1.3.5 (2017-03-16)

  • Fixed None timeout support (#1720)

1.3.4 (2017-03-14)

  • Revert timeout handling in client request

  • Fix StreamResponse representation after eof

  • Fix file_sender to not fall on bad request (range out of file size)

  • Fix file_sender to correct stream video to Chromes

  • Fix NotImplementedError server exception (#1703)

  • Clearer error message for URL without a host name. (#1691)

  • Silence deprecation warning in __repr__ (#1690)

  • IDN + HTTPS = ssl.CertificateError (#1685)

1.3.3 (2017-02-19)

  • Fixed memory leak in time service (#1656)

1.3.2 (2017-02-16)

  • Awaiting on WebSocketResponse.send_* does not work (#1645)

  • Fix multiple calls to client ws_connect when using a shared header dict (#1643)

  • Make CookieJar.filter_cookies() accept plain string parameter. (#1636)

1.3.1 (2017-02-09)

  • Handle CLOSING in WebSocketResponse.__anext__

  • Fixed AttributeError ‘drain’ for server websocket handler (#1613)

1.3.0 (2017-02-08)

  • Multipart writer validates the data on append instead of on a request send (#920)

  • Multipart reader accepts multipart messages with or without their epilogue to consistently handle valid and legacy behaviors (#1526) (#1581)

  • Separate read + connect + request timeouts # 1523

  • Do not swallow Upgrade header (#1587)

  • Fix polls demo run application (#1487)

  • Ignore unknown 1XX status codes in client (#1353)

  • Fix sub-Multipart messages missing their headers on serialization (#1525)

  • Do not use readline when reading the content of a part in the multipart reader (#1535)

  • Add optional flag for quoting FormData fields (#916)

  • 416 Range Not Satisfiable if requested range end > file size (#1588)

  • Having a : or @ in a route does not work (#1552)

  • Added receive_timeout timeout for websocket to receive complete message. (#1325)

  • Added heartbeat parameter for websocket to automatically send ping message. (#1024) (#777)

  • Remove web.Application dependency from web.UrlDispatcher (#1510)

  • Accepting back-pressure from slow websocket clients (#1367)

  • Do not pause transport during set_parser stage (#1211)

  • Lingering close does not terminate before timeout (#1559)

  • setsockopt may raise OSError exception if socket is closed already (#1595)

  • Lots of CancelledError when requests are interrupted (#1565)

  • Allow users to specify what should happen to decoding errors when calling a responses text() method (#1542)

  • Back port std module http.cookies for python3.4.2 (#1566)

  • Maintain url’s fragment in client response (#1314)

  • Allow concurrently close WebSocket connection (#754)

  • Gzipped responses with empty body raises ContentEncodingError (#609)

  • Return 504 if request handle raises TimeoutError.

  • Refactor how we use keep-alive and close lingering timeouts.

  • Close response connection if we can not consume whole http message during client response release

  • Abort closed ssl client transports, broken servers can keep socket open un-limit time (#1568)

  • Log warning instead of RuntimeError is websocket connection is closed.

  • Deprecated: aiohttp.protocol.HttpPrefixParser will be removed in 1.4 (#1590)

  • Deprecated: Servers response’s .started, .start() and .can_start() method will be removed in 1.4 (#1591)

  • Deprecated: Adding sub app via app.router.add_subapp() is deprecated use app.add_subapp() instead, will be removed in 1.4 (#1592)

  • Deprecated: aiohttp.get(), aiohttp.options(), aiohttp.head(), aiohttp.post(), aiohttp.put(), aiohttp.patch(), aiohttp.delete(), and aiohttp.ws_connect() will be removed in 1.4 (#1593)

  • Deprecated: Application.finish() and Application.register_on_finish() will be removed in 1.4 (#1602)

1.2.0 (2016-12-17)

  • Extract BaseRequest from web.Request, introduce web.Server (former RequestHandlerFactory), introduce new low-level web server which is not coupled with web.Application and routing (#1362)

  • Make TestServer.make_url compatible with yarl.URL (#1389)

  • Implement range requests for static files (#1382)

  • Support task attribute for StreamResponse (#1410)

  • Drop TestClient.app property, use TestClient.server.app instead (BACKWARD INCOMPATIBLE)

  • Drop TestClient.handler property, use TestClient.server.handler instead (BACKWARD INCOMPATIBLE)

  • TestClient.server property returns a test server instance, was asyncio.AbstractServer (BACKWARD INCOMPATIBLE)

  • Follow gunicorn’s signal semantics in Gunicorn[UVLoop]WebWorker (#1201)

  • Call worker_int and worker_abort callbacks in Gunicorn[UVLoop]WebWorker (#1202)

  • Has functional tests for client proxy (#1218)

  • Fix bugs with client proxy target path and proxy host with port (#1413)

  • Fix bugs related to the use of unicode hostnames (#1444)

  • Preserve cookie quoting/escaping (#1453)

  • FileSender will send gzipped response if gzip version available (#1426)

  • Don’t override Content-Length header in web.Response if no body was set (#1400)

  • Introduce router.post_init() for solving (#1373)

  • Fix raise error in case of multiple calls of TimeServive.stop()

  • Allow to raise web exceptions on router resolving stage (#1460)

  • Add a warning for session creation outside of coroutine (#1468)

  • Avoid a race when application might start accepting incoming requests but startup signals are not processed yet e98e8c6

  • Raise a RuntimeError when trying to change the status of the HTTP response after the headers have been sent (#1480)

  • Fix bug with https proxy acquired cleanup (#1340)

  • Use UTF-8 as the default encoding for multipart text parts (#1484)

1.1.6 (2016-11-28)

  • Fix BodyPartReader.read_chunk bug about returns zero bytes before EOF (#1428)

1.1.5 (2016-11-16)

  • Fix static file serving in fallback mode (#1401)

1.1.4 (2016-11-14)

  • Make TestServer.make_url compatible with yarl.URL (#1389)

  • Generate informative exception on redirects from server which does not provide redirection headers (#1396)

1.1.3 (2016-11-10)

  • Support root resources for sub-applications (#1379)

1.1.2 (2016-11-08)

  • Allow starting variables with an underscore (#1379)

  • Properly process UNIX sockets by gunicorn worker (#1375)

  • Fix ordering for FrozenList

  • Don’t propagate pre and post signals to sub-application (#1377)

1.1.1 (2016-11-04)

  • Fix documentation generation (#1120)

1.1.0 (2016-11-03)

  • Drop deprecated WSClientDisconnectedError (BACKWARD INCOMPATIBLE)

  • Use yarl.URL in client API. The change is 99% backward compatible but ClientResponse.url is an yarl.URL instance now. (#1217)

  • Close idle keep-alive connections on shutdown (#1222)

  • Modify regex in AccessLogger to accept underscore and numbers (#1225)

  • Use yarl.URL in web server API. web.Request.rel_url and web.Request.url are added. URLs and templates are percent-encoded now. (#1224)

  • Accept yarl.URL by server redirections (#1278)

  • Return yarl.URL by .make_url() testing utility (#1279)

  • Properly format IPv6 addresses by aiohttp.web.run_app (#1139)

  • Use yarl.URL by server API (#1288)

    • Introduce resource.url_for(), deprecate resource.url().

    • Implement StaticResource.

    • Inherit SystemRoute from AbstractRoute

    • Drop old-style routes: Route, PlainRoute, DynamicRoute, StaticRoute, ResourceAdapter.

  • Revert resp.url back to str, introduce resp.url_obj (#1292)

  • Raise ValueError if BasicAuth login has a “:” character (#1307)

  • Fix bug when ClientRequest send payload file with opened as open(‘filename’, ‘r+b’) (#1306)

  • Enhancement to AccessLogger (pass extra dict) (#1303)

  • Show more verbose message on import errors (#1319)

  • Added save and load functionality for CookieJar (#1219)

  • Added option on StaticRoute to follow symlinks (#1299)

  • Force encoding of application/json content type to utf-8 (#1339)

  • Fix invalid invocations of errors.LineTooLong (#1335)

  • Websockets: Stop async for iteration when connection is closed (#1144)

  • Ensure TestClient HTTP methods return a context manager (#1318)

  • Raise ClientDisconnectedError to FlowControlStreamReader read function if ClientSession object is closed by client when reading data. (#1323)

  • Document deployment without Gunicorn (#1120)

  • Add deprecation warning for MD5 and SHA1 digests when used for fingerprint of site certs in TCPConnector. (#1186)

  • Implement sub-applications (#1301)

  • Don’t inherit web.Request from dict but implement MutableMapping protocol.

  • Implement frozen signals

  • Don’t inherit web.Application from dict but implement MutableMapping protocol.

  • Support freezing for web applications

  • Accept access_log parameter in web.run_app, use None to disable logging

  • Don’t flap tcp_cork and tcp_nodelay in regular request handling. tcp_nodelay is still enabled by default.

  • Improve performance of web server by removing premature computing of Content-Type if the value was set by web.Response constructor.

    While the patch boosts speed of trivial web.Response(text=’OK’, content_type=’text/plain) very well please don’t expect significant boost of your application – a couple DB requests and business logic is still the main bottleneck.

  • Boost performance by adding a custom time service (#1350)

  • Extend ClientResponse with content_type and charset properties like in web.Request. (#1349)

  • Disable aiodns by default (#559)

  • Don’t flap tcp_cork in client code, use TCP_NODELAY mode by default.

  • Implement web.Request.clone() (#1361)

1.0.5 (2016-10-11)

  • Fix StreamReader._read_nowait to return all available data up to the requested amount (#1297)

1.0.4 (2016-09-22)

  • Fix FlowControlStreamReader.read_nowait so that it checks whether the transport is paused (#1206)

1.0.2 (2016-09-22)

  • Make CookieJar compatible with 32-bit systems (#1188)

  • Add missing WSMsgType to web_ws.__all__, see (#1200)

  • Fix CookieJar ctor when called with loop=None (#1203)

  • Fix broken upper-casing in wsgi support (#1197)

1.0.1 (2016-09-16)

  • Restore aiohttp.web.MsgType alias for aiohttp.WSMsgType for sake of backward compatibility (#1178)

  • Tune alabaster schema.

  • Use text/html content type for displaying index pages by static file handler.

  • Fix AssertionError in static file handling (#1177)

  • Fix access log formats %O and %b for static file handling

  • Remove debug setting of GunicornWorker, use app.debug to control its debug-mode instead

1.0.0 (2016-09-16)

  • Change default size for client session’s connection pool from unlimited to 20 (#977)

  • Add IE support for cookie deletion. (#994)

  • Remove deprecated WebSocketResponse.wait_closed method (BACKWARD INCOMPATIBLE)

  • Remove deprecated force parameter for ClientResponse.close method (BACKWARD INCOMPATIBLE)

  • Avoid using of mutable CIMultiDict kw param in make_mocked_request (#997)

  • Make WebSocketResponse.close a little bit faster by avoiding new task creating just for timeout measurement

  • Add proxy and proxy_auth params to client.get() and family, deprecate ProxyConnector (#998)

  • Add support for websocket send_json and receive_json, synchronize server and client API for websockets (#984)

  • Implement router shourtcuts for most useful HTTP methods, use app.router.add_get(), app.router.add_post() etc. instead of app.router.add_route() (#986)

  • Support SSL connections for gunicorn worker (#1003)

  • Move obsolete examples to legacy folder

  • Switch to multidict 2.0 and title-cased strings (#1015)

  • {FOO}e logger format is case-sensitive now

  • Fix logger report for unix socket 8e8469b

  • Rename aiohttp.websocket to aiohttp._ws_impl

  • Rename aiohttp.MsgType tp aiohttp.WSMsgType

  • Introduce aiohttp.WSMessage officially

  • Rename Message -> WSMessage

  • Remove deprecated decode param from resp.read(decode=True)

  • Use 5min default client timeout (#1028)

  • Relax HTTP method validation in UrlDispatcher (#1037)

  • Pin minimal supported asyncio version to 3.4.2+ (loop.is_close() should be present)

  • Remove aiohttp.websocket module (BACKWARD INCOMPATIBLE) Please use high-level client and server approaches

  • Link header for 451 status code is mandatory

  • Fix test_client fixture to allow multiple clients per test (#1072)

  • make_mocked_request now accepts dict as headers (#1073)

  • Add Python 3.5.2/3.6+ compatibility patch for async generator protocol change (#1082)

  • Improvement test_client can accept instance object (#1083)

  • Simplify ServerHttpProtocol implementation (#1060)

  • Add a flag for optional showing directory index for static file handling (#921)

  • Define web.Application.on_startup() signal handler (#1103)

  • Drop ChunkedParser and LinesParser (#1111)

  • Call Application.startup in GunicornWebWorker (#1105)

  • Fix client handling hostnames with 63 bytes when a port is given in the url (#1044)

  • Implement proxy support for ClientSession.ws_connect (#1025)

  • Return named tuple from WebSocketResponse.can_prepare (#1016)

  • Fix access_log_format in GunicornWebWorker (#1117)

  • Setup Content-Type to application/octet-stream by default (#1124)

  • Deprecate debug parameter from app.make_handler(), use Application(debug=True) instead (#1121)

  • Remove fragment string in request path (#846)

  • Use aiodns.DNSResolver.gethostbyname() if available (#1136)

  • Fix static file sending on uvloop when sendfile is available (#1093)

  • Make prettier urls if query is empty dict (#1143)

  • Fix redirects for HEAD requests (#1147)

  • Default value for StreamReader.read_nowait is -1 from now (#1150)

  • aiohttp.StreamReader is not inherited from asyncio.StreamReader from now (BACKWARD INCOMPATIBLE) (#1150)

  • Streams documentation added (#1150)

  • Add multipart coroutine method for web Request object (#1067)

  • Publish ClientSession.loop property (#1149)

  • Fix static file with spaces (#1140)

  • Fix piling up asyncio loop by cookie expiration callbacks (#1061)

  • Drop Timeout class for sake of async_timeout external library. aiohttp.Timeout is an alias for async_timeout.timeout

  • use_dns_cache parameter of aiohttp.TCPConnector is True by default (BACKWARD INCOMPATIBLE) (#1152)

  • aiohttp.TCPConnector uses asynchronous DNS resolver if available by default (BACKWARD INCOMPATIBLE) (#1152)

  • Conform to RFC3986 - do not include url fragments in client requests (#1174)

  • Drop ClientSession.cookies (BACKWARD INCOMPATIBLE) (#1173)

  • Refactor AbstractCookieJar public API (BACKWARD INCOMPATIBLE) (#1173)

  • Fix clashing cookies with have the same name but belong to different domains (BACKWARD INCOMPATIBLE) (#1125)

  • Support binary Content-Transfer-Encoding (#1169)

0.22.5 (08-02-2016)

  • Pin miltidict version to >=1.2.2

0.22.3 (07-26-2016)

  • Do not filter cookies if unsafe flag provided (#1005)

0.22.2 (07-23-2016)

  • Suppress CancelledError when Timeout raises TimeoutError (#970)

  • Don’t expose aiohttp.__version__

  • Add unsafe parameter to CookieJar (#968)

  • Use unsafe cookie jar in test client tools

  • Expose aiohttp.CookieJar name

0.22.1 (07-16-2016)

  • Large cookie expiration/max-age does not break an event loop from now (fixes (#967))

0.22.0 (07-15-2016)

  • Fix bug in serving static directory (#803)

  • Fix command line arg parsing (#797)

  • Fix a documentation chapter about cookie usage (#790)

  • Handle empty body with gzipped encoding (#758)

  • Support 451 Unavailable For Legal Reasons http status (#697)

  • Fix Cookie share example and few small typos in docs (#817)

  • UrlDispatcher.add_route with partial coroutine handler (#814)

  • Optional support for aiodns (#728)

  • Add ServiceRestart and TryAgainLater websocket close codes (#828)

  • Fix prompt message for web.run_app (#832)

  • Allow to pass None as a timeout value to disable timeout logic (#834)

  • Fix leak of connection slot during connection error (#835)

  • Gunicorn worker with uvloop support aiohttp.worker.GunicornUVLoopWebWorker (#878)

  • Don’t send body in response to HEAD request (#838)

  • Skip the preamble in MultipartReader (#881)

  • Implement BasicAuth decode classmethod. (#744)

  • Don’t crash logger when transport is None (#889)

  • Use a create_future compatibility wrapper instead of creating Futures directly (#896)

  • Add test utilities to aiohttp (#902)

  • Improve Request.__repr__ (#875)

  • Skip DNS resolving if provided host is already an ip address (#874)

  • Add headers to ClientSession.ws_connect (#785)

  • Document that server can send pre-compressed data (#906)

  • Don’t add Content-Encoding and Transfer-Encoding if no body (#891)

  • Add json() convenience methods to websocket message objects (#897)

  • Add client_resp.raise_for_status() (#908)

  • Implement cookie filter (#799)

  • Include an example of middleware to handle error pages (#909)

  • Fix error handling in StaticFileMixin (#856)

  • Add mocked request helper (#900)

  • Fix empty ALLOW Response header for cls based View (#929)

  • Respect CONNECT method to implement a proxy server (#847)

  • Add pytest_plugin (#914)

  • Add tutorial

  • Add backlog option to support more than 128 (default value in “create_server” function) concurrent connections (#892)

  • Allow configuration of header size limits (#912)

  • Separate sending file logic from StaticRoute dispatcher (#901)

  • Drop deprecated share_cookies connector option (BACKWARD INCOMPATIBLE)

  • Drop deprecated support for tuple as auth parameter. Use aiohttp.BasicAuth instead (BACKWARD INCOMPATIBLE)

  • Remove deprecated request.payload property, use content instead. (BACKWARD INCOMPATIBLE)

  • Drop all mentions about api changes in documentation for versions older than 0.16

  • Allow to override default cookie jar (#963)

  • Add manylinux wheel builds

  • Dup a socket for sendfile usage (#964)

0.21.6 (05-05-2016)

  • Drop initial query parameters on redirects (#853)

0.21.5 (03-22-2016)

  • Fix command line arg parsing (#797)

0.21.4 (03-12-2016)

  • Fix ResourceAdapter: don’t add method to allowed if resource is not match (#826)

  • Fix Resource: append found method to returned allowed methods

0.21.2 (02-16-2016)

  • Fix a regression: support for handling ~/path in static file routes was broken (#782)

0.21.1 (02-10-2016)

  • Make new resources classes public (#767)

  • Add router.resources() view

  • Fix cmd-line parameter names in doc

0.21.0 (02-04-2016)

  • Introduce on_shutdown signal (#722)

  • Implement raw input headers (#726)

  • Implement web.run_app utility function (#734)

  • Introduce on_cleanup signal

  • Deprecate Application.finish() / Application.register_on_finish() in favor of on_cleanup.

  • Get rid of bare aiohttp.request(), aiohttp.get() and family in docs (#729)

  • Deprecate bare aiohttp.request(), aiohttp.get() and family (#729)

  • Refactor keep-alive support (#737):

    • Enable keepalive for HTTP 1.0 by default

    • Disable it for HTTP 0.9 (who cares about 0.9, BTW?)

    • For keepalived connections

      • Send Connection: keep-alive for HTTP 1.0 only

      • don’t send Connection header for HTTP 1.1

    • For non-keepalived connections

      • Send Connection: close for HTTP 1.1 only

      • don’t send Connection header for HTTP 1.0

  • Add version parameter to ClientSession constructor, deprecate it for session.request() and family (#736)

  • Enable access log by default (#735)

  • Deprecate app.router.register_route() (the method was not documented intentionally BTW).

  • Deprecate app.router.named_routes() in favor of app.router.named_resources()

  • route.add_static accepts pathlib.Path now (#743)

  • Add command line support: $ python -m aiohttp.web package.main (#740)

  • FAQ section was added to docs. Enjoy and fill free to contribute new topics

  • Add async context manager support to ClientSession

  • Document ClientResponse’s host, method, url properties

  • Use CORK/NODELAY in client API (#748)

  • ClientSession.close and Connector.close are coroutines now

  • Close client connection on exception in ClientResponse.release()

  • Allow to read multipart parts without content-length specified (#750)

  • Add support for unix domain sockets to gunicorn worker (#470)

  • Add test for default Expect handler (#601)

  • Add the first demo project

  • Rename loader keyword argument in web.Request.json method. (#646)

  • Add local socket binding for TCPConnector (#678)

0.20.2 (01-07-2016)

  • Enable use of await for a class based view (#717)

  • Check address family to fill wsgi env properly (#718)

  • Fix memory leak in headers processing (thanks to Marco Paolini) (#723)

0.20.1 (12-30-2015)

  • Raise RuntimeError is Timeout context manager was used outside of task context.

  • Add number of bytes to stream.read_nowait (#700)

  • Use X-FORWARDED-PROTO for wsgi.url_scheme when available

0.20.0 (12-28-2015)

  • Extend list of web exceptions, add HTTPMisdirectedRequest, HTTPUpgradeRequired, HTTPPreconditionRequired, HTTPTooManyRequests, HTTPRequestHeaderFieldsTooLarge, HTTPVariantAlsoNegotiates, HTTPNotExtended, HTTPNetworkAuthenticationRequired status codes (#644)

  • Do not remove AUTHORIZATION header by WSGI handler (#649)

  • Fix broken support for https proxies with authentication (#617)

  • Get REMOTE_* and SEVER_* http vars from headers when listening on unix socket (#654)

  • Add HTTP 308 support (#663)

  • Add Tf format (time to serve request in seconds, %06f format) to access log (#669)

  • Remove one and a half years long deprecated ClientResponse.read_and_close() method

  • Optimize chunked encoding: use a single syscall instead of 3 calls on sending chunked encoded data

  • Use TCP_CORK and TCP_NODELAY to optimize network latency and throughput (#680)

  • Websocket XOR performance improved (#687)

  • Avoid sending cookie attributes in Cookie header (#613)

  • Round server timeouts to seconds for grouping pending calls. That leads to less amount of poller syscalls e.g. epoll.poll(). (#702)

  • Close connection on websocket handshake error (#703)

  • Implement class based views (#684)

  • Add headers parameter to ws_connect() (#709)

  • Drop unused function parse_remote_addr() (#708)

  • Close session on exception (#707)

  • Store http code and headers in WSServerHandshakeError (#706)

  • Make some low-level message properties readonly (#710)

0.19.0 (11-25-2015)

  • Memory leak in ParserBuffer (#579)

  • Support gunicorn’s max_requests settings in gunicorn worker

  • Fix wsgi environment building (#573)

  • Improve access logging (#572)

  • Drop unused host and port from low-level server (#586)

  • Add Python 3.5 async for implementation to server websocket (#543)

  • Add Python 3.5 async for implementation to client websocket

  • Add Python 3.5 async with implementation to client websocket

  • Add charset parameter to web.Response constructor (#593)

  • Forbid passing both Content-Type header and content_type or charset params into web.Response constructor

  • Forbid duplicating of web.Application and web.Request (#602)

  • Add an option to pass Origin header in ws_connect (#607)

  • Add json_response function (#592)

  • Make concurrent connections respect limits (#581)

  • Collect history of responses if redirects occur (#614)

  • Enable passing pre-compressed data in requests (#621)

  • Expose named routes via UrlDispatcher.named_routes() (#622)

  • Allow disabling sendfile by environment variable AIOHTTP_NOSENDFILE (#629)

  • Use ensure_future if available

  • Always quote params for Content-Disposition (#641)

  • Support async for in multipart reader (#640)

  • Add Timeout context manager (#611)

0.18.4 (13-11-2015)

  • Relax rule for router names again by adding dash to allowed characters: they may contain identifiers, dashes, dots and columns

0.18.3 (25-10-2015)

  • Fix formatting for _RequestContextManager helper (#590)

0.18.2 (22-10-2015)

  • Fix regression for OpenSSL < 1.0.0 (#583)

0.18.1 (20-10-2015)

  • Relax rule for router names: they may contain dots and columns starting from now

0.18.0 (19-10-2015)

  • Use errors.HttpProcessingError.message as HTTP error reason and message (#459)

  • Optimize cythonized multidict a bit

  • Change repr’s of multidicts and multidict views

  • default headers in ClientSession are now case-insensitive

  • Make ‘=’ char and ‘wss://’ schema safe in urls (#477)

  • ClientResponse.close() forces connection closing by default from now (#479)

    N.B. Backward incompatible change: was .close(force=False) Using `force parameter for the method is deprecated: use .release() instead.

  • Properly requote URL’s path (#480)

  • add skip_auto_headers parameter for client API (#486)

  • Properly parse URL path in aiohttp.web.Request (#489)

  • Raise RuntimeError when chunked enabled and HTTP is 1.0 (#488)

  • Fix a bug with processing io.BytesIO as data parameter for client API (#500)

  • Skip auto-generation of Content-Type header (#507)

  • Use sendfile facility for static file handling (#503)

  • Default response_factory in app.router.add_static now is StreamResponse, not None. The functionality is not changed if default is not specified.

  • Drop ClientResponse.message attribute, it was always implementation detail.

  • Streams are optimized for speed and mostly memory in case of a big HTTP message sizes (#496)

  • Fix a bug for server-side cookies for dropping cookie and setting it again without Max-Age parameter.

  • Don’t trim redirect URL in client API (#499)

  • Extend precision of access log “D” to milliseconds (#527)

  • Deprecate StreamResponse.start() method in favor of StreamResponse.prepare() coroutine (#525)

    .start() is still supported but responses begun with .start() does not call signal for response preparing to be sent.

  • Add StreamReader.__repr__

  • Drop Python 3.3 support, from now minimal required version is Python 3.4.1 (#541)

  • Add async with support for ClientSession.request() and family (#536)

  • Ignore message body on 204 and 304 responses (#505)

  • TCPConnector processed both IPv4 and IPv6 by default (#559)

  • Add .routes() view for urldispatcher (#519)

  • Route name should be a valid identifier name from now (#567)

  • Implement server signals (#562)

  • Drop a year-old deprecated files parameter from client API.

  • Added async for support for aiohttp stream (#542)

0.17.4 (09-29-2015)

  • Properly parse URL path in aiohttp.web.Request (#489)

  • Add missing coroutine decorator, the client api is await-compatible now

0.17.3 (08-28-2015)

  • Remove Content-Length header on compressed responses (#450)

  • Support Python 3.5

  • Improve performance of transport in-use list (#472)

  • Fix connection pooling (#473)

0.17.2 (08-11-2015)

  • Don’t forget to pass data argument forward (#462)

  • Fix multipart read bytes count (#463)

0.17.1 (08-10-2015)

  • Fix multidict comparison to arbitrary abc.Mapping

0.17.0 (08-04-2015)

  • Make StaticRoute support Last-Modified and If-Modified-Since headers (#386)

  • Add Request.if_modified_since and Stream.Response.last_modified properties

  • Fix deflate compression when writing a chunked response (#395)

  • Request`s content-length header is cleared now after redirect from POST method (#391)

  • Return a 400 if server received a non HTTP content (#405)

  • Fix keep-alive support for aiohttp clients (#406)

  • Allow gzip compression in high-level server response interface (#403)

  • Rename TCPConnector.resolve and family to dns_cache (#415)

  • Make UrlDispatcher ignore quoted characters during url matching (#414) Backward-compatibility warning: this may change the url matched by your queries if they send quoted character (like %2F for /) (#414)

  • Use optional cchardet accelerator if present (#418)

  • Borrow loop from Connector in ClientSession if loop is not set

  • Add context manager support to ClientSession for session closing.

  • Add toplevel get(), post(), put(), head(), delete(), options(), patch() coroutines.

  • Fix IPv6 support for client API (#425)

  • Pass SSL context through proxy connector (#421)

  • Make the rule: path for add_route should start with slash

  • Don’t process request finishing by low-level server on closed event loop

  • Don’t override data if multiple files are uploaded with same key (#433)

  • Ensure multipart.BodyPartReader.read_chunk read all the necessary data to avoid false assertions about malformed multipart payload

  • Don’t send body for 204, 205 and 304 http exceptions (#442)

  • Correctly skip Cython compilation in MSVC not found (#453)

  • Add response factory to StaticRoute (#456)

  • Don’t append trailing CRLF for multipart.BodyPartReader (#454)

0.16.6 (07-15-2015)

  • Skip compilation on Windows if vcvarsall.bat cannot be found (#438)

0.16.5 (06-13-2015)

  • Get rid of all comprehensions and yielding in _multidict (#410)

0.16.4 (06-13-2015)

  • Don’t clear current exception in multidict’s __repr__ (cythonized versions) (#410)

0.16.3 (05-30-2015)

  • Fix StaticRoute vulnerability to directory traversal attacks (#380)

0.16.2 (05-27-2015)

  • Update python version required for __del__ usage: it’s actually 3.4.1 instead of 3.4.0

  • Add check for presence of loop.is_closed() method before call the former (#378)

0.16.1 (05-27-2015)

  • Fix regression in static file handling (#377)

0.16.0 (05-26-2015)

  • Unset waiter future after cancellation (#363)

  • Update request url with query parameters (#372)

  • Support new fingerprint param of TCPConnector to enable verifying SSL certificates via MD5, SHA1, or SHA256 digest (#366)

  • Setup uploaded filename if field value is binary and transfer encoding is not specified (#349)

  • Implement ClientSession.close() method

  • Implement connector.closed readonly property

  • Implement ClientSession.closed readonly property

  • Implement ClientSession.connector readonly property

  • Implement ClientSession.detach method

  • Add __del__ to client-side objects: sessions, connectors, connections, requests, responses.

  • Refactor connections cleanup by connector (#357)

  • Add limit parameter to connector constructor (#358)

  • Add request.has_body property (#364)

  • Add response_class parameter to ws_connect() (#367)

  • ProxyConnector does not support keep-alive requests by default starting from now (#368)

  • Add connector.force_close property

  • Add ws_connect to ClientSession (#374)

  • Support optional chunk_size parameter in router.add_static()

0.15.3 (04-22-2015)

  • Fix graceful shutdown handling

  • Fix Expect header handling for not found and not allowed routes (#340)

0.15.2 (04-19-2015)

  • Flow control subsystem refactoring

  • HTTP server performance optimizations

  • Allow to match any request method with *

  • Explicitly call drain on transport (#316)

  • Make chardet module dependency mandatory (#318)

  • Support keep-alive for HTTP 1.0 (#325)

  • Do not chunk single file during upload (#327)

  • Add ClientSession object for cookie storage and default headers (#328)

  • Add keep_alive_on argument for HTTP server handler.

0.15.1 (03-31-2015)

  • Pass Autobahn Testsuite tests

  • Fixed websocket fragmentation

  • Fixed websocket close procedure

  • Fixed parser buffer limits

  • Added timeout parameter to WebSocketResponse ctor

  • Added WebSocketResponse.close_code attribute

0.15.0 (03-27-2015)

  • Client WebSockets support

  • New Multipart system (#273)

  • Support for “Except” header (#287) (#267)

  • Set default Content-Type for post requests (#184)

  • Fix issue with construction dynamic route with regexps and trailing slash (#266)

  • Add repr to web.Request

  • Add repr to web.Response

  • Add repr for NotFound and NotAllowed match infos

  • Add repr for web.Application

  • Add repr to UrlMappingMatchInfo (#217)

  • Gunicorn 19.2.x compatibility

0.14.4 (01-29-2015)

  • Fix issue with error during constructing of url with regex parts (#264)

0.14.3 (01-28-2015)

  • Use path=’/’ by default for cookies (#261)

0.14.2 (01-23-2015)

  • Connections leak in BaseConnector (#253)

  • Do not swallow websocket reader exceptions (#255)

  • web.Request’s read, text, json are memorized (#250)

0.14.1 (01-15-2015)

  • HttpMessage._add_default_headers does not overwrite existing headers (#216)

  • Expose multidict classes at package level

  • add aiohttp.web.WebSocketResponse

  • According to RFC 6455 websocket subprotocol preference order is provided by client, not by server

  • websocket’s ping and pong accept optional message parameter

  • multidict views do not accept getall parameter anymore, it returns the full body anyway.

  • multidicts have optional Cython optimization, cythonized version of multidicts is about 5 times faster than pure Python.

  • multidict.getall() returns list, not tuple.

  • Backward incompatible change: now there are two mutable multidicts (MultiDict, CIMultiDict) and two immutable multidict proxies (MultiDictProxy and CIMultiDictProxy). Previous edition of multidicts was not a part of public API BTW.

  • Router refactoring to push Not Allowed and Not Found in middleware processing

  • Convert ConnectionError to aiohttp.DisconnectedError and don’t eat ConnectionError exceptions from web handlers.

  • Remove hop headers from Response class, wsgi response still uses hop headers.

  • Allow to send raw chunked encoded response.

  • Allow to encode output bytes stream into chunked encoding.

  • Allow to compress output bytes stream with deflate encoding.

  • Server has 75 seconds keepalive timeout now, was non-keepalive by default.

  • Application does not accept **kwargs anymore ((#243)).

  • Request is inherited from dict now for making per-request storage to middlewares ((#242)).

0.13.1 (12-31-2014)

  • Add aiohttp.web.StreamResponse.started property (#213)

  • HTML escape traceback text in ServerHttpProtocol.handle_error

  • Mention handler and middlewares in aiohttp.web.RequestHandler.handle_request on error ((#218))

0.13.0 (12-29-2014)

  • StreamResponse.charset converts value to lower-case on assigning.

  • Chain exceptions when raise ClientRequestError.

  • Support custom regexps in route variables (#204)

  • Fixed graceful shutdown, disable keep-alive on connection closing.

  • Decode HTTP message with utf-8 encoding, some servers send headers in utf-8 encoding (#207)

  • Support aiohtt.web middlewares (#209)

  • Add ssl_context to TCPConnector (#206)

0.12.0 (12-12-2014)

  • Deep refactoring of aiohttp.web in backward-incompatible manner. Sorry, we have to do this.

  • Automatically force aiohttp.web handlers to coroutines in UrlDispatcher.add_route() (#186)

  • Rename Request.POST() function to Request.post()

  • Added POST attribute

  • Response processing refactoring: constructor does not accept Request instance anymore.

  • Pass application instance to finish callback

  • Exceptions refactoring

  • Do not unquote query string in aiohttp.web.Request

  • Fix concurrent access to payload in RequestHandle.handle_request()

  • Add access logging to aiohttp.web

  • Gunicorn worker for aiohttp.web

  • Removed deprecated AsyncGunicornWorker

  • Removed deprecated HttpClient

0.11.0 (11-29-2014)

  • Support named routes in aiohttp.web.UrlDispatcher (#179)

  • Make websocket subprotocols conform to spec (#181)

0.10.2 (11-19-2014)

  • Don’t unquote environ[‘PATH_INFO’] in wsgi.py (#177)

0.10.1 (11-17-2014)

  • aiohttp.web.HTTPException and descendants now files response body with string like 404: NotFound

  • Fix multidict __iter__, the method should iterate over keys, not (key, value) pairs.

0.10.0 (11-13-2014)

  • Add aiohttp.web subpackage for highlevel HTTP server support.

  • Add reason optional parameter to aiohttp.protocol.Response ctor.

  • Fix aiohttp.client bug for sending file without content-type.

  • Change error text for connection closed between server responses from ‘Can not read status line’ to explicit ‘Connection closed by server’

  • Drop closed connections from connector (#173)

  • Set server.transport to None on .closing() (#172)

0.9.3 (10-30-2014)

  • Fix compatibility with asyncio 3.4.1+ (#170)

0.9.2 (10-16-2014)

  • Improve redirect handling (#157)

  • Send raw files as is (#153)

  • Better websocket support (#150)

0.9.1 (08-30-2014)

  • Added MultiDict support for client request params and data (#114).

  • Fixed parameter type for IncompleteRead exception (#118).

  • Strictly require ASCII headers names and values (#137)

  • Keep port in ProxyConnector (#128).

  • Python 3.4.1 compatibility (#131).

0.9.0 (07-08-2014)

  • Better client basic authentication support (#112).

  • Fixed incorrect line splitting in HttpRequestParser (#97).

  • Support StreamReader and DataQueue as request data.

  • Client files handling refactoring (#20).

  • Backward incompatible: Replace DataQueue with StreamReader for request payload (#87).

0.8.4 (07-04-2014)

  • Change ProxyConnector authorization parameters.

0.8.3 (07-03-2014)

  • Publish TCPConnector properties: verify_ssl, family, resolve, resolved_hosts.

  • Don’t parse message body for HEAD responses.

  • Refactor client response decoding.

0.8.2 (06-22-2014)

  • Make ProxyConnector.proxy immutable property.

  • Make UnixConnector.path immutable property.

  • Fix resource leak for aiohttp.request() with implicit connector.

  • Rename Connector’s reuse_timeout to keepalive_timeout.

0.8.1 (06-18-2014)

  • Use case insensitive multidict for server request/response headers.

  • MultiDict.getall() accepts default value.

  • Catch server ConnectionError.

  • Accept MultiDict (and derived) instances in aiohttp.request header argument.

  • Proxy ‘CONNECT’ support.

0.8.0 (06-06-2014)

  • Add support for utf-8 values in HTTP headers

  • Allow to use custom response class instead of HttpResponse

  • Use MultiDict for client request headers

  • Use MultiDict for server request/response headers

  • Store response headers in ClientResponse.headers attribute

  • Get rid of timeout parameter in aiohttp.client API

  • Exceptions refactoring

0.7.3 (05-20-2014)

  • Simple HTTP proxy support.

0.7.2 (05-14-2014)

  • Get rid of __del__ methods

  • Use ResourceWarning instead of logging warning record.

0.7.1 (04-28-2014)

  • Do not unquote client request urls.

  • Allow multiple waiters on transport drain.

  • Do not return client connection to pool in case of exceptions.

  • Rename SocketConnector to TCPConnector and UnixSocketConnector to UnixConnector.

0.7.0 (04-16-2014)

  • Connection flow control.

  • HTTP client session/connection pool refactoring.

  • Better handling for bad server requests.

0.6.5 (03-29-2014)

  • Added client session reuse timeout.

  • Better client request cancellation support.

  • Better handling responses without content length.

  • Added HttpClient verify_ssl parameter support.

0.6.4 (02-27-2014)

  • Log content-length missing warning only for put and post requests.

0.6.3 (02-27-2014)

  • Better support for server exit.

  • Read response body until EOF if content-length is not defined (#14)

0.6.2 (02-18-2014)

  • Fix trailing char in allowed_methods.

  • Start slow request timer for first request.

0.6.1 (02-17-2014)

  • Added utility method HttpResponse.read_and_close()

  • Added slow request timeout.

  • Enable socket SO_KEEPALIVE if available.

0.6.0 (02-12-2014)

  • Better handling for process exit.

0.5.0 (01-29-2014)

  • Allow to use custom HttpRequest client class.

  • Use gunicorn keepalive setting for asynchronous worker.

  • Log leaking responses.

  • python 3.4 compatibility

0.4.4 (11-15-2013)

  • Resolve only AF_INET family, because it is not clear how to pass extra info to asyncio.

0.4.3 (11-15-2013)

  • Allow to wait completion of request with HttpResponse.wait_for_close()

0.4.2 (11-14-2013)

  • Handle exception in client request stream.

  • Prevent host resolving for each client request.

0.4.1 (11-12-2013)

  • Added client support for expect: 100-continue header.

0.4 (11-06-2013)

  • Added custom wsgi application close procedure

  • Fixed concurrent host failure in HttpClient

0.3 (11-04-2013)

  • Added PortMapperWorker

  • Added HttpClient

  • Added TCP connection timeout to HTTP client

  • Better client connection errors handling

  • Gracefully handle process exit

0.2

  • Fix packaging

Indices and tables

Who uses aiohttp?

The list of aiohttp users: both libraries, big projects and web sites.

Please don’t hesitate to add your awesome project to the list by making a Pull Request on GitHub.

If you like the project – please go to GitHub and press Star button!

Third-Party libraries

aiohttp is not the library for making HTTP requests and creating WEB server only.

It is the grand basement for libraries built on top of aiohttp.

This page is a list of these tools.

Please feel free to add your open sourced library if it’s not enlisted yet by making Pull Request to https://github.com/aio-libs/aiohttp/

  • Why do you might want to include your awesome library into the list?

  • Just because the list increases your library visibility. People will have an easy way to find it.

Officially supported

This list contains libraries which are supported by aio-libs team and located on https://github.com/aio-libs

aiohttp extensions
Database drivers
Other tools

  • aiodocker Python Docker API client based on asyncio and aiohttp.

  • aiobotocore asyncio support for botocore library using aiohttp.

Approved third-party libraries

The libraries are not part of aio-libs but they are proven to be very well written and highly recommended for usage.

  • uvloop Ultra fast implementation of asyncio event loop on top of libuv.

    We are highly recommending to use it instead of standard asyncio.

Database drivers

  • asyncpg Another PostgreSQL async driver. It’s much faster than aiopg but it is not drop-in replacement – the API is different. Anyway please take a look on it – the driver is really incredible fast.

Others

The list of libraries which are exists but not enlisted in former categories.

They may be perfect or not – we don’t know.

Please add your library reference here first and after some time period ask to raise the status.

  • aiohttp-cache A cache system for aiohttp server.

  • aiocache Caching for asyncio with multiple backends (framework agnostic)

  • gain Web crawling framework based on asyncio for everyone.

  • aiohttp-swagger Swagger API Documentation builder for aiohttp server.

  • aiohttp-swaggerify Library to automatically generate swagger2.0 definition for aiohttp endpoints.

  • aiohttp-validate Simple library that helps you validate your API endpoints requests/responses with json schema.

  • aiohttp-pydantic An aiohttp.View to validate the HTTP request’s body, query-string, and headers regarding function annotations and generate Open API doc. Python 3.8+ required.

  • raven-aiohttp An aiohttp transport for raven-python (Sentry client).

  • webargs A friendly library for parsing HTTP request arguments, with built-in support for popular web frameworks, including Flask, Django, Bottle, Tornado, Pyramid, webapp2, Falcon, and aiohttp.

  • aioauth-client OAuth client for aiohttp.

  • aiohttpretty A simple asyncio compatible httpretty mock using aiohttp.

  • aioresponses a helper for mock/fake web requests in python aiohttp package.

  • aiohttp-transmute A transmute implementation for aiohttp.

  • aiohttp_apiset Package to build routes using swagger specification.

  • aiohttp-login Registration and authorization (including social) for aiohttp applications.

  • aiohttp_utils Handy utilities for building aiohttp.web applications.

  • aiohttpproxy Simple aiohttp HTTP proxy.

  • aiohttp_traversal Traversal based router for aiohttp.web.

  • aiohttp_autoreload Makes aiohttp server auto-reload on source code change.

  • gidgethub An async GitHub API library for Python.

  • aiohttp_jrpc aiohttp JSON-RPC service.

  • fbemissary A bot framework for the Facebook Messenger platform, built on asyncio and aiohttp.

  • aioslacker slacker wrapper for asyncio.

  • aioreloader Port of tornado reloader to asyncio.

  • aiohttp_babel Babel localization support for aiohttp.

  • python-mocket a socket mock framework - for all kinds of socket animals, web-clients included.

  • aioraft asyncio RAFT algorithm based on aiohttp.

  • home-assistant Open-source home automation platform running on Python 3.

  • discord.py Discord client library.

  • aiogram A fully asynchronous library for Telegram Bot API written with asyncio and aiohttp.

  • vk.py Extremely-fast Python 3.6+ toolkit for create applications work`s with VKAPI.

  • aiohttp-graphql GraphQL and GraphIQL interface for aiohttp.

  • aiohttp-sentry An aiohttp middleware for reporting errors to Sentry. Python 3.5+ is required.

  • aiohttp-datadog An aiohttp middleware for reporting metrics to DataDog. Python 3.5+ is required.

  • async-v20 Asynchronous FOREX client for OANDA’s v20 API. Python 3.6+

  • aiohttp-jwt An aiohttp middleware for JWT(JSON Web Token) support. Python 3.5+ is required.

  • AWS Xray Python SDK Native tracing support for Aiohttp applications.

  • GINO An asyncio ORM on top of SQLAlchemy core, delivered with an aiohttp extension.

  • aiohttp-apispec Build and document REST APIs with aiohttp and apispec.

  • eider-py Python implementation of the Eider RPC protocol.

  • asynapplicationinsights A client for Azure Application Insights implemented using aiohttp client, including a middleware for aiohttp servers to collect web apps telemetry.

  • aiogmaps Asynchronous client for Google Maps API Web Services. Python 3.6+ required.

  • DBGR Terminal based tool to test and debug HTTP APIs with aiohttp.

  • rororo Implement aiohtp.web OpenAPI 3 server applications with schema first approach. Python 3.6+ required.

  • aiohttp-middlewares Collection of useful middlewares for aiohttp.web applications. Python 3.6+ required.

  • aiohttp-tus tus.io protocol implementation for aiohttp.web applications. Python 3.6+ required.

  • aiohttp-sse-client A Server-Sent Event python client base on aiohttp. Python 3.6+ required.

Built with aiohttp

aiohttp is used to build useful libraries built on top of it, and there’s a page dedicated to list them: Third-Party libraries.

There are also projects that leverage the power of aiohttp to provide end-user tools, like command lines or software with full user interfaces.

This page aims to list those projects. If you are using aiohttp in your software and if it’s playing a central role, you can add it here in this list.

You can also add a Built with aiohttp link somewhere in your project, pointing to https://github.com/aio-libs/aiohttp.

Powered by aiohttp

Web sites powered by aiohttp.

Feel free to fork documentation on github, add a link to your site and make a Pull Request!

Contributing

Instructions for contributors

In order to make a clone of the GitHub repo: open the link and press the “Fork” button on the upper-right menu of the web page.

I hope everybody knows how to work with git and github nowadays :)

Workflow is pretty straightforward:

  1. Make sure you are reading the latest version of this document. It can be found in the GitHub repo in the docs subdirectory.

  2. Clone the GitHub repo using the --recurse-submodules argument

  3. Setup your machine with the required dev environment

  4. Make a change

  5. Make sure all tests passed

  6. Add a file into the CHANGES folder (see Changelog update for how).

  7. Commit changes to your own aiohttp clone

  8. Make a pull request from the github page of your clone against the master branch

  9. Optionally make backport Pull Request(s) for landing a bug fix into released aiohttp versions.

Note

The project uses Squash-and-Merge strategy for GitHub Merge button.

Basically it means that there is no need to rebase a Pull Request against master branch. Just git merge master into your working copy (a fork) if needed. The Pull Request is automatically squashed into the single commit once the PR is accepted.

Note

GitHub issue and pull request threads are automatically locked when there has not been any recent activity for one year. Please open a new issue for related bugs.

If you feel like there are important points in the locked discussions, please include those excerpts into that new issue.

Preconditions for running aiohttp test suite

We expect you to use a python virtual environment to run our tests.

There are several ways to make a virtual environment.

If you like to use virtualenv please run:

$ cd aiohttp
$ virtualenv --python=`which python3` venv
$ . venv/bin/activate

For standard python venv:

$ cd aiohttp
$ python3 -m venv venv
$ . venv/bin/activate

For virtualenvwrapper:

$ cd aiohttp
$ mkvirtualenv --python=`which python3` aiohttp

There are other tools like pyvenv but you know the rule of thumb now: create a python3 virtual environment and activate it.

After that please install libraries required for development:

$ pip install -r requirements/dev.txt

Note

For now, the development tooling depends on make and assumes an Unix OS If you wish to contribute to aiohttp from a Windows machine, the easiest way is probably to configure the WSL so you can use the same instructions. If it’s not possible for you or if it doesn’t work, please contact us so we can find a solution together.

Install pre-commit hooks:

$ pre-commit install

Warning

If you plan to use temporary print(), pdb or ipdb within the test suite, execute it with -s:

$ pytest tests -s

in order to run the tests without output capturing.

Congratulations, you are ready to run the test suite!

Run autoformatter

The project uses black + isort formatters to keep the source code style. Please run make fmt after every change before starting tests.

$ make fmt

Run aiohttp test suite

After all the preconditions are met you can run tests typing the next command:

$ make test

The command at first will run the linters (sorry, we don’t accept pull requests with pyflakes, black, isort, or mypy errors).

On lint success the tests will be run.

Please take a look on the produced output.

Any extra texts (print statements and so on) should be removed.

Tests coverage

We are trying hard to have good test coverage; please don’t make it worse.

Use:

$ make cov-dev

to run test suite and collect coverage information. Once the command has finished check your coverage at the file that appears in the last line of the output: open file:///.../aiohttp/htmlcov/index.html

Please go to the link and make sure that your code change is covered.

The project uses codecov.io for storing coverage results. Visit https://codecov.io/gh/aio-libs/aiohttp for looking on coverage of master branch, history, pull requests etc.

The browser extension https://docs.codecov.io/docs/browser-extension is highly recommended for analyzing the coverage just in Files Changed tab on GitHub Pull Request review page.

Documentation

We encourage documentation improvements.

Please before making a Pull Request about documentation changes run:

$ make doc

Once it finishes it will output the index html page open file:///.../aiohttp/docs/_build/html/index.html.

Go to the link and make sure your doc changes looks good.

Spell checking

We use pyenchant and sphinxcontrib-spelling for running spell checker for documentation:

$ make doc-spelling

Unfortunately there are problems with running spell checker on MacOS X.

To run spell checker on Linux box you should install it first:

$ sudo apt-get install enchant
$ pip install sphinxcontrib-spelling

Changelog update

The CHANGES.rst file is managed using towncrier tool and all non trivial changes must be accompanied by a news entry.

To add an entry to the news file, first you need to have created an issue describing the change you want to make. A Pull Request itself may function as such, but it is preferred to have a dedicated issue (for example, in case the PR ends up rejected due to code quality reasons).

Once you have an issue or pull request, you take the number and you create a file inside of the CHANGES/ directory named after that issue number with an extension of .removal, .feature, .bugfix, or .doc. Thus if your issue or PR number is 1234 and this change is fixing a bug, then you would create a file CHANGES/1234.bugfix. PRs can span multiple categories by creating multiple files (for instance, if you added a feature and deprecated/removed the old feature at the same time, you would create CHANGES/NNNN.feature and CHANGES/NNNN.removal). Likewise if a PR touches multiple issues/PRs you may create a file for each of them with the exact same contents and Towncrier will deduplicate them.

The contents of this file are reStructuredText formatted text that will be used as the content of the news file entry. You do not need to reference the issue or PR numbers here as towncrier will automatically add a reference to all of the affected issues when rendering the news file.

Making a Pull Request

After finishing all steps make a GitHub Pull Request with master base branch.

Backporting

All Pull Requests are created against master git branch.

If the Pull Request is not a new functionality but bug fixing backport to maintenance branch would be desirable.

aiohttp project committer may ask for making a backport of the PR into maintained branch(es), in this case he or she adds a github label like needs backport to 3.1.

Backporting is performed after main PR merging into master.

Please do the following steps:

  1. Find Pull Request’s commit for cherry-picking.

    aiohttp does squashing PRs on merging, so open your PR page on github and scroll down to message like asvetlov merged commit f7b8921 into master 9 days ago. f7b8921 is the required commit number.

  2. Run cherry_picker tool for making backport PR (the tool is already pre-installed from ./requirements/dev.txt), e.g. cherry_picker f7b8921 3.1.

  3. In case of conflicts fix them and continue cherry-picking by cherry_picker --continue.

    cherry_picker --abort stops the process.

    cherry_picker --status shows current cherry-picking status (like git status)

  4. After all conflicts are done the tool opens a New Pull Request page in a browser with pre-filed information. Create a backport Pull Request and wait for review/merging.

  5. aiohttp committer should remove backport Git label after merging the backport.

How to become an aiohttp committer

Contribute!

The easiest way is providing Pull Requests for issues in our bug tracker. But if you have a great idea for the library improvement – please make an issue and Pull Request.

The rules for committers are simple:

  1. No wild commits! Everything should go through PRs.

  2. Take a part in reviews. It’s very important part of maintainer’s activity.

  3. Pickup issues created by others, especially if they are simple.

  4. Keep test suite comprehensive. In practice it means leveling up coverage. 97% is not bad but we wish to have 100% someday. Well, 99% is good target too.

  5. Don’t hesitate to improve our docs. Documentation is very important thing, it’s the key for project success. The documentation should not only cover our public API but help newbies to start using the project and shed a light on non-obvious gotchas.

After positive answer aiohttp committer creates an issue on github with the proposal for nomination. If the proposal will collect only positive votes and no strong objection – you’ll be a new member in our team.