18.5.1. Event loops

The event loop is the central execution device provided by asyncio. It provides multiple facilities, amongst which:

  • Registering, executing and cancelling delayed calls (timeouts).
  • Creating client and server transports for various kinds of communication.
  • Launching subprocesses and the associated transports for communication with an external program.
  • Delegating costly function calls to a pool of threads.

18.5.1.1. Event loop policies and the default policy

Event loop management is abstracted with a policy pattern, to provide maximal flexibility for custom platforms and frameworks. Throughout the execution of a process, a single global policy object manages the event loops available to the process based on the calling context. A policy is an object implementing the AbstractEventLoopPolicy interface.

For most users of asyncio, policies never have to be dealt with explicitly, since the default global policy is sufficient.

The default policy defines context as the current thread, and manages an event loop per thread that interacts with asyncio. The module-level functions get_event_loop() and set_event_loop() provide convenient access to event loops managed by the default policy.

18.5.1.2. Event loop functions

The following functions are convenient shortcuts to accessing the methods of the global policy. Note that this provides access to the default policy, unless an alternative policy was set by calling set_event_loop_policy() earlier in the execution of the process.

asyncio.get_event_loop()

Equivalent to calling get_event_loop_policy().get_event_loop().

asyncio.set_event_loop(loop)

Equivalent to calling get_event_loop_policy().set_event_loop(loop).

asyncio.new_event_loop()

Equivalent to calling get_event_loop_policy().new_event_loop().

18.5.1.3. Event loop policy interface

An event loop policy must implement the following interface:

class asyncio.AbstractEventLoopPolicy
get_event_loop()

Get the event loop for current context. Returns an event loop object implementing BaseEventLoop interface, or raises an exception in case no event loop has been set for the current context and the current policy does not specify to create one. It should never return None.

set_event_loop(loop)

Set the event loop of the current context to loop.

new_event_loop()

Create and return a new event loop object according to this policy’s rules. If there’s need to set this loop as the event loop of the current context, set_event_loop() must be called explicitly.

18.5.1.4. Access to the global loop policy

asyncio.get_event_loop_policy()

Get the current event loop policy.

asyncio.set_event_loop_policy(policy)

Set the current event loop policy. If policy is None, the default policy is restored.

18.5.1.5. Run an event loop

BaseEventLoop.run_forever()

Run until stop() is called.

BaseEventLoop.run_until_complete(future)

Run until the Future is done.

If the argument is a coroutine, it is wrapped in a Task.

Return the Future’s result, or raise its exception.

BaseEventLoop.is_running()

Returns running status of event loop.

BaseEventLoop.stop()

Stop running the event loop.

Every callback scheduled before stop() is called will run. Callback scheduled after stop() is called won’t. However, those callbacks will run if run_forever() is called again later.

BaseEventLoop.close()

Close the event loop. The loop should not be running.

This clears the queues and shuts down the executor, but does not wait for the executor to finish.

This is idempotent and irreversible. No other methods should be called after this one.

18.5.1.6. Calls

BaseEventLoop.call_soon(callback, *args)

Arrange for a callback to be called as soon as possible.

This operates as a FIFO queue, callbacks are called in the order in which they are registered. Each callback will be called exactly once.

Any positional arguments after the callback will be passed to the callback when it is called.

An instance of asyncio.Handle is returned.

BaseEventLoop.call_soon_threadsafe(callback, *args)

Like call_soon(), but thread safe.

18.5.1.7. Delayed calls

The event loop has its own internal clock for computing timeouts. Which clock is used depends on the (platform-specific) event loop implementation; ideally it is a monotonic clock. This will generally be a different clock than time.time().

Note

Timeouts (relative delay or absolute when) should not exceed one day.

BaseEventLoop.call_later(delay, callback, *args)

Arrange for the callback to be called after the given delay seconds (either an int or float).

An instance of asyncio.Handle is returned.

callback will be called exactly once per call to call_later(). If two callbacks are scheduled for exactly the same time, it is undefined which will be called first.

The optional positional args will be passed to the callback when it is called. If you want the callback to be called with some named arguments, use a closure or functools.partial().

BaseEventLoop.call_at(when, callback, *args)

Arrange for the callback to be called at the given absolute timestamp when (an int or float), using the same time reference as time().

This method’s behavior is the same as call_later().

BaseEventLoop.time()

Return the current time, as a float value, according to the event loop’s internal clock.

See also

The asyncio.sleep() function.

18.5.1.8. Creating connections

BaseEventLoop.create_connection(protocol_factory, host=None, port=None, *, ssl=None, family=0, proto=0, flags=0, sock=None, local_addr=None, server_hostname=None)

Create a streaming transport connection to a given Internet host and port: socket family AF_INET or AF_INET6 depending on host (or family if specified), socket type SOCK_STREAM. protocol_factory must be a callable returning a protocol instance.

This method is a coroutine which will try to establish the connection in the background. When successful, the coroutine returns a (transport, protocol) pair.

The chronological synopsis of the underlying operation is as follows:

  1. The connection is established, and a transport is created to represent it.
  2. protocol_factory is called without arguments and must return a protocol instance.
  3. The protocol instance is tied to the transport, and its connection_made() method is called.
  4. The coroutine returns successfully with the (transport, protocol) pair.

The created transport is an implementation-dependent bidirectional stream.

Note

protocol_factory can be any kind of callable, not necessarily a class. For example, if you want to use a pre-created protocol instance, you can pass lambda: my_protocol.

Options allowing to change how the connection is created:

  • ssl: if given and not false, a SSL/TLS transport is created (by default a plain TCP transport is created). If ssl is a ssl.SSLContext object, this context is used to create the transport; if ssl is True, a context with some unspecified default settings is used.

  • server_hostname, is only for use together with ssl, and sets or overrides the hostname that the target server’s certificate will be matched against. By default the value of the host argument is used. If host is empty, there is no default and you must pass a value for server_hostname. If server_hostname is an empty string, hostname matching is disabled (which is a serious security risk, allowing for man-in-the-middle-attacks).

  • family, proto, flags are the optional address family, protocol and flags to be passed through to getaddrinfo() for host resolution. If given, these should all be integers from the corresponding socket module constants.

  • sock, if given, should be an existing, already connected socket.socket object to be used by the transport. If sock is given, none of host, port, family, proto, flags and local_addr should be specified.

  • local_addr, if given, is a (local_host, local_port) tuple used to bind the socket to locally. The local_host and local_port are looked up using getaddrinfo(), similarly to host and port.

See also

The open_connection() function can be used to get a pair of (StreamReader, StreamWriter) instead of a protocol.

BaseEventLoop.create_datagram_endpoint(protocol_factory, local_addr=None, remote_addr=None, *, family=0, proto=0, flags=0)

Create datagram connection: socket family AF_INET or AF_INET6 depending on host (or family if specified), socket type SOCK_DGRAM.

This method is a coroutine which will try to establish the connection in the background. When successful, the coroutine returns a (transport, protocol) pair.

See the BaseEventLoop.create_connection() method for parameters.

BaseEventLoop.create_unix_connection(protocol_factory, path, *, ssl=None, sock=None, server_hostname=None)

Create UNIX connection: socket family AF_UNIX, socket type SOCK_STREAM. The AF_UNIX socket family is used to communicate between processes on the same machine efficiently.

This method is a coroutine which will try to establish the connection in the background. When successful, the coroutine returns a (transport, protocol) pair.

See the BaseEventLoop.create_connection() method for parameters.

Availability: UNIX.

18.5.1.9. Creating listening connections

BaseEventLoop.create_server(protocol_factory, host=None, port=None, *, family=socket.AF_UNSPEC, flags=socket.AI_PASSIVE, sock=None, backlog=100, ssl=None, reuse_address=None)

A coroutine method which creates a TCP server bound to host and port.

The return value is a AbstractServer object which can be used to stop the service.

If host is an empty string or None all interfaces are assumed and a list of multiple sockets will be returned (most likely one for IPv4 and another one for IPv6).

family can be set to either AF_INET or AF_INET6 to force the socket to use IPv4 or IPv6. If not set it will be determined from host (defaults to AF_UNSPEC).

flags is a bitmask for getaddrinfo().

sock can optionally be specified in order to use a preexisting socket object.

backlog is the maximum number of queued connections passed to listen() (defaults to 100).

ssl can be set to an SSLContext to enable SSL over the accepted connections.

reuse_address 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.

See also

The function start_server() creates a (StreamReader, StreamWriter) pair and calls back a function with this pair.

BaseEventLoop.create_unix_server(protocol_factory, path=None, *, sock=None, backlog=100, ssl=None)

Similar to BaseEventLoop.create_server(), but specific to the socket family AF_UNIX.

Availability: UNIX.

18.5.1.10. Watch file descriptors

BaseEventLoop.add_reader(fd, callback, *args)

Start watching the file descriptor for read availability and then call the callback with specified arguments.

BaseEventLoop.remove_reader(fd)

Stop watching the file descriptor for read availability.

BaseEventLoop.add_writer(fd, callback, *args)

Start watching the file descriptor for write availability and then call the callback with specified arguments.

BaseEventLoop.remove_writer(fd)

Stop watching the file descriptor for write availability.

18.5.1.11. Low-level socket operations

BaseEventLoop.sock_recv(sock, nbytes)

Receive data from the socket. The return value is a bytes object representing the data received. The maximum amount of data to be received at once is specified by nbytes.

This method is a coroutine.

See also

The socket.socket.recv() method.

BaseEventLoop.sock_sendall(sock, data)

Send data to the socket. The socket must be connected to a remote socket. This method continues to send data from data until either all data has been sent or an error occurs. None is returned on success. On error, an exception is raised, and there is no way to determine how much data, if any, was successfully processed by the receiving end of the connection.

This method is a coroutine.

See also

The socket.socket.sendall() method.

BaseEventLoop.sock_connect(sock, address)

Connect to a remote socket at address.

The address must be already resolved to avoid the trap of hanging the entire event loop when the address requires doing a DNS lookup. For example, it must be an IP address, not an hostname, for AF_INET and AF_INET6 address families. Use getaddrinfo() to resolve the hostname asynchronously.

This method is a coroutine.

See also

The BaseEventLoop.create_connection() method, the open_connection() function and the socket.socket.connect() method.

BaseEventLoop.sock_accept(sock)

Accept a connection. The socket must be bound to an address and listening for connections. The return value is a pair (conn, address) where conn is a new socket object usable to send and receive data on the connection, and address is the address bound to the socket on the other end of the connection.

This method is a coroutine.

See also

The BaseEventLoop.create_server() method, the start_server() function and the socket.socket.accept() method.

18.5.1.12. Resolve host name

BaseEventLoop.getaddrinfo(host, port, *, family=0, type=0, proto=0, flags=0)

This method is a coroutine, similar to socket.getaddrinfo() function but non-blocking.

BaseEventLoop.getnameinfo(sockaddr, flags=0)

This method is a coroutine, similar to socket.getnameinfo() function but non-blocking.

18.5.1.13. Connect pipes

BaseEventLoop.connect_read_pipe(protocol_factory, pipe)

Register read pipe in eventloop.

protocol_factory should instantiate object with Protocol interface. pipe is file-like object already switched to nonblocking. Return pair (transport, protocol), where transport support ReadTransport interface.

This method is a coroutine.

BaseEventLoop.connect_write_pipe(protocol_factory, pipe)

Register write pipe in eventloop.

protocol_factory should instantiate object with BaseProtocol interface. Pipe is file-like object already switched to nonblocking. Return pair (transport, protocol), where transport support WriteTransport interface.

This method is a coroutine.

18.5.1.14. UNIX signals

Availability: UNIX only.

BaseEventLoop.add_signal_handler(signum, callback, *args)

Add a handler for a signal.

Raise ValueError if the signal number is invalid or uncatchable. Raise RuntimeError if there is a problem setting up the handler.

BaseEventLoop.remove_signal_handler(sig)

Remove a handler for a signal.

Return True if a signal handler was removed, False if not.

See also

The signal module.

18.5.1.15. Executor

Call a function in an Executor (pool of threads or pool of processes). By default, an event loop uses a thread pool executor (ThreadPoolExecutor).

BaseEventLoop.run_in_executor(executor, callback, *args)

Arrange for a callback to be called in the specified executor.

The executor argument should be an Executor instance. The default executor is used if executor is None.

This method is a coroutine.

BaseEventLoop.set_default_executor(executor)

Set the default executor used by run_in_executor().

18.5.1.16. Error Handling API

Allows to customize how exceptions are handled in the event loop.

BaseEventLoop.set_exception_handler(handler)

Set handler as the new event loop exception handler.

If handler is None, the default exception handler will be set.

If handler is a callable object, it should have a matching signature to (loop, context), where loop will be a reference to the active event loop, context will be a dict object (see call_exception_handler() documentation for details about context).

BaseEventLoop.default_exception_handler(context)

Default exception handler.

This is called when an exception occurs and no exception handler is set, and can be called by a custom exception handler that wants to defer to the default behavior.

context parameter has the same meaning as in call_exception_handler().

BaseEventLoop.call_exception_handler(context)

Call the current event loop exception handler.

context is a dict object containing the following keys (new keys may be introduced later):

  • ‘message’: Error message;
  • ‘exception’ (optional): Exception object;
  • ‘future’ (optional): asyncio.Future instance;
  • ‘handle’ (optional): asyncio.Handle instance;
  • ‘protocol’ (optional): Protocol instance;
  • ‘transport’ (optional): Transport instance;
  • ‘socket’ (optional): socket.socket instance.

Note

Note: this method should not be overloaded in subclassed event loops. For any custom exception handling, use set_exception_handler() method.

18.5.1.17. Debug mode

BaseEventLoop.get_debug()

Get the debug mode (bool) of the event loop, False by default.

BaseEventLoop.set_debug(enabled: bool)

Set the debug mode of the event loop.

See also

The Develop with asyncio section.

18.5.1.18. Server

class asyncio.AbstractServer

Abstract server returned by BaseEventLoop.create_server().

close()

Stop serving. This leaves existing connections open.

wait_closed()

A coroutine to wait until service is closed.

18.5.1.19. Handle

class asyncio.Handle

A callback wrapper object returned by BaseEventLoop.call_soon(), BaseEventLoop.call_soon_threadsafe(), BaseEventLoop.call_later(), and BaseEventLoop.call_at().

cancel()

Cancel the call.

18.5.1.20. Example: Hello World (callback)

Print Hello World every two seconds, using a callback:

import asyncio

def print_and_repeat(loop):
    print('Hello World')
    loop.call_later(2, print_and_repeat, loop)

loop = asyncio.get_event_loop()
loop.call_soon(print_and_repeat, loop)
loop.run_forever()

18.5.1.21. Example: Set signal handlers for SIGINT and SIGTERM

Register handlers for signals SIGINT and SIGTERM:

import asyncio
import functools
import os
import signal

def ask_exit(signame):
    print("got signal %s: exit" % signame)
    loop.stop()

loop = asyncio.get_event_loop()
for signame in ('SIGINT', 'SIGTERM'):
    loop.add_signal_handler(getattr(signal, signame),
                            functools.partial(ask_exit, signame))

print("Event loop running forever, press CTRL+c to interrupt.")
print("pid %s: send SIGINT or SIGTERM to exit." % os.getpid())
loop.run_forever()