New in version 3.4.
This module allows high-level and efficient I/O multiplexing, built upon the select module primitives. Users are encouraged to use this module instead, unless they want precise control over the OS-level primitives used.
It defines a BaseSelector abstract base class, along with several concrete implementations (KqueueSelector, EpollSelector...), that can be used to wait for I/O readiness notification on multiple file objects. In the following, “file object” refers to any object with a fileno() method, or a raw file descriptor. See file object.
DefaultSelector is an alias to the most efficient implementation available on the current platform: this should be the default choice for most users.
The type of file objects supported depends on the platform: on Windows, sockets are supported, but not pipes, whereas on Unix, both are supported (some other types may be supported as well, such as fifos or special file devices).
BaseSelector +-- SelectSelector +-- PollSelector +-- EpollSelector +-- KqueueSelector
In the following, events is a bitwise mask indicating which I/O events should be waited for on a given file object. It can be a combination of the constants below:
Constant Meaning EVENT_READ Available for read EVENT_WRITE Available for write
File object registered.
Underlying file descriptor.
Events that must be waited for on this file object.
Optional opaque data associated to this file object: for example, this could be used to store a per-client session ID.
A BaseSelector is used to wait for I/O event readiness on multiple file objects. It supports file stream registration, unregistration, and a method to wait for I/O events on those streams, with an optional timeout. It’s an abstract base class, so cannot be instantiated. Use DefaultSelector instead, or one of SelectSelector, KqueueSelector etc. if you want to specifically use an implementation, and your platform supports it. BaseSelector and its concrete implementations support the context manager protocol.
Register a file object for selection, monitoring it for I/O events.
fileobj is the file object to monitor. It may either be an integer file descriptor or an object with a fileno() method. events is a bitwise mask of events to monitor. data is an opaque object.
Unregister a file object from selection, removing it from monitoring. A file object shall be unregistered prior to being closed.
fileobj must be a file object previously registered.
This returns the associated SelectorKey instance, or raises a KeyError if fileobj is not registered. It will raise ValueError if fileobj is invalid (e.g. it has no fileno() method or its fileno() method has an invalid return value).
Change a registered file object’s monitored events or attached data.
This is equivalent to BaseSelector.unregister(fileobj)() followed by BaseSelector.register(fileobj, events, data)(), except that it can be implemented more efficiently.
Wait until some registered file objects become ready, or the timeout expires.
If timeout > 0, this specifies the maximum wait time, in seconds. If timeout <= 0, the call won’t block, and will report the currently ready file objects. If timeout is None, the call will block until a monitored file object becomes ready.
This returns a list of (key, events) tuples, one for each ready file object.
key is the SelectorKey instance corresponding to a ready file object. events is a bitmask of events ready on this file object.
This method can return before any file object becomes ready or the timeout has elapsed if the current process receives a signal: in this case, an empty list will be returned.
Close the selector.
This must be called to make sure that any underlying resource is freed. The selector shall not be used once it has been closed.
Return the key associated with a registered file object.
The default selector class, using the most efficient implementation available on the current platform. This should be the default choice for most users.
Here is a simple echo server implementation:
import selectors import socket sel = selectors.DefaultSelector() def accept(sock, mask): conn, addr = sock.accept() # Should be ready print('accepted', conn, 'from', addr) conn.setblocking(False) sel.register(conn, selectors.EVENT_READ, read) def read(conn, mask): data = conn.recv(1000) # Should be ready if data: print('echoing', repr(data), 'to', conn) conn.send(data) # Hope it won't block else: print('closing', conn) sel.unregister(conn) conn.close() sock = socket.socket() sock.bind(('localhost', 1234)) sock.listen(100) sock.setblocking(False) sel.register(sock, selectors.EVENT_READ, accept) while True: events = sel.select() for key, mask in events: callback = key.data callback(key.fileobj, mask)