Sanic Server

sanic.http

class sanic.http.Http(protocol)

Bases: Stream

Internal helper for managing the HTTP/1.1 request/response cycle

Raises:

• ServerError –

• PayloadTooLarge –

• Exception –

• BadRequest –

• ExpectationFailed –

• RuntimeError –

• ServerError –

• ServerError –

• BadRequest –

• BadRequest –

• BadRequest –

• PayloadTooLarge –

• RuntimeError –

create_empty_request()

Current error handling code needs a request object that won’t exist if an error occurred during before a request was received. Create a bogus response for error handling use.

Return type:

None

async error_response(exception)

Handle response when exception encountered

Parameters:

exception (Exception) –

Return type:

None

head_response_ignored(data, end_stream)

HEAD response: body data silently ignored.

Parameters:

• data (bytes) –

• end_stream (bool) –

Return type:

None

async http1()

HTTP 1.1 connection handler

async http1_request_header()

Receive and parse request header into self.request.

async http1_response_chunked(data, end_stream)

Format a part of response body in chunked encoding.

Parameters:

• data (bytes) –

• end_stream (bool) –

Return type:

None

async http1_response_normal(data, end_stream)

Format / keep track of non-chunked response.

Parameters:

• data (bytes) –

• end_stream (bool) –

Return type:

None

init_for_request()

Init/reset all per-request variables.

log_response()

Helper method provided to enable the logging of responses in case if the HttpProtocol.access_log is enabled.

Return type:

None

async read()

Read some bytes of request body.

Return type:

bytes | None

respond(response)

Initiate new streaming response.

Nothing is sent until the first send() call on the returned object, and calling this function multiple times will just alter the response to be given.

Parameters:

response (BaseHTTPResponse) –

Return type:

BaseHTTPResponse

class sanic.http.Http3(protocol, transmit)

Bases: object

Internal helper for managing the HTTP/3 request/response cycle

Parameters:

• protocol (Http3Protocol) –

• transmit (Callable[[], None]) –

enum sanic.http.Stage(value)

Bases: Enum

Enum for representing the stage of the request/response cycle

IDLE Waiting for request

REQUEST Request headers being received

HANDLER Headers done, handler running

RESPONSE Response headers sent, body in progress

FAILED Unrecoverable state (error while sending response)

Valid values are as follows:

IDLE = <Stage.IDLE: 0>

REQUEST = <Stage.REQUEST: 1>

HANDLER = <Stage.HANDLER: 3>

RESPONSE = <Stage.RESPONSE: 4>

FAILED = <Stage.FAILED: 100>

sanic.server

class sanic.server.AsyncioServer(app, loop, serve_coro, connections)

Bases: object

Wraps an asyncio server with functionality that might be useful to a user who needs to manage the server lifecycle manually.

Parameters:

app (Sanic) –

after_start()

Trigger “after_server_start” events

after_stop()

Trigger “after_server_stop” events

before_start()

Trigger “before_server_start” events

before_stop()

Trigger “before_server_stop” events

startup()

Trigger “before_server_start” events

class sanic.server.ConnInfo(transport, unix=None)

Bases: object

Local and remote addresses and SSL status info.

Parameters:

transport (TransportProtocol) –

class sanic.server.HttpProtocol(*, loop, app, signal=None, connections=None, state=None, unix=None, **kwargs)

Bases: HttpProtocolMixin, SanicProtocol

This class provides implements the HTTP 1.1 protocol on top of our Sanic Server transport

Parameters:

app (Sanic) –

HTTP_CLASS

alias of Http

check_timeouts()

Runs itself periodically to enforce any expired timeouts.

close_if_idle()

Close the connection if a request is not being sent or received

Returns:

boolean - True if closed, false if staying open

Return type:

bool

connection_made(transport)

HTTP-protocol-specific new connection handler

async connection_task()

Run a HTTP connection.

Timeouts and some additional error handling occur here, while most of everything else happens in class Http or in code called from there.

data_received(data)

Called when some data is received.

The argument is a bytes object.

Parameters:

data (bytes) –

async send(data)

Writes HTTP data with backpressure control.

sanic.server.serve(host, port, app, ssl=None, sock=None, unix=None, reuse_port=False, loop=None, protocol=<class 'sanic.server.protocols.http_protocol.HttpProtocol'>, backlog=100, register_sys_signals=True, run_multiple=False, run_async=False, connections=None, signal=<sanic.models.server_types.Signal object>, state=None, asyncio_server_kwargs=None, version=HTTP.VERSION_1)

Start asynchronous HTTP Server on an individual process.

Parameters:

• host – Address to host on

• port – Port to host on

• before_start – function to be executed before the server starts listening. Takes arguments app instance and loop

• after_start – function to be executed after the server starts listening. Takes arguments app instance and loop

• before_stop – function to be executed when a stop signal is received before it is respected. Takes arguments app instance and loop

• after_stop – function to be executed when a stop signal is received after it is respected. Takes arguments app instance and loop

• ssl (Optional[SSLContext]) – SSLContext

• sock (Optional[socket.socket]) – Socket for the server to accept connections from

• unix (Optional[str]) – Unix socket to listen on instead of TCP port

• reuse_port (bool) – True for multiple workers

• loop – asyncio compatible event loop

• run_async (bool) – bool: Do not create a new event loop for the server, and return an AsyncServer object rather than running it

• asyncio_server_kwargs – key-value args for asyncio/uvloop create_server method

• app (Sanic) –

• protocol (Type[asyncio.Protocol]) –

• backlog (int) –

• register_sys_signals (bool) –

• run_multiple (bool) –

Returns:

Nothing

sanic.server.serve_multiple(server_settings, workers)

Start multiple server processes simultaneously. Stop on interrupt and terminate signals, and drain connections when complete.

Parameters:

• server_settings – kw arguments to be passed to the serve function

• workers – number of workers to launch

• stop_event – if provided, is used as a stop signal

Returns:

sanic.server.try_use_uvloop()

Use uvloop instead of the default asyncio loop.

Return type:

None