Source code:
———
The socketserver module simplifies the task of writing network servers.
There are four basic concrete server classes:
classsocketserver.TCPServer(server_address, RequestHandlerClass, bind_and_activate=True)
This uses the internet TCP protocol, which provides for continuous streams of data between the client and server. If bind_and_activate is true, the constructor automatically attempts to invoke
and
. The other parameters are passed to the
base class.
classsocketserver.UDPServer(server_address, RequestHandlerClass, bind_and_activate=True)
This uses datagrams, which are discrete packets of information that may arrive out of order or be lost while in transit. The parameters are the same as for
.
classsocketserver.UnixStreamServer(server_address, RequestHandlerClass, bind_and_activate=True)
classsocketserver.UnixDatagramServer(server_address, RequestHandlerClass, bind_and_activate=True)
These more infrequently used classes are similar to the TCP and UDP classes, but use Unix domain sockets; they’re not available on non-Unix platforms. The parameters are the same as for
.
These four classes process requests synchronously; each request must be completed before the next request can be started. This isn’t suitable if each request takes a long time to complete, because it requires a lot of computation, or because it returns a lot of data which the client is slow to process. The solution is to create a separate process or thread to handle each request; the
and
mix-in classes can be used to support asynchronous behaviour.
Creating a server requires several steps. First, you must create a request handler class by subclassing the
class and overriding its
method; this method will process incoming requests. Second, you must instantiate one of the server classes, passing it the server’s address and the request handler class. It is recommended to use the server in a
statement. Then call the
or
method of the server object to process one or many requests. Finally, call
to close the socket (unless you used a with statement).
When inheriting from
for threaded connection behavior, you should explicitly declare how you want your threads to behave on an abrupt shutdown. The ThreadingMixIn class defines an attribute daemon_threads, which indicates whether or not the server should wait for thread termination. You should set the flag explicitly if you would like threads to behave autonomously; the default is
, meaning that Python will not exit until all threads created by ThreadingMixIn have exited.
Server classes have the same external methods and attributes, no matter what network protocol they use.
Server Creation Notes
There are five classes in an inheritance diagram, four of which represent synchronous servers of four types:
+------------+|BaseServer|+------------+|v+-----------++------------------+|TCPServer|------->|UnixStreamServer|+-----------++------------------+|v+-----------++--------------------+|UDPServer|------->|UnixDatagramServer|+-----------++--------------------+Note that
derives from
, not from
— the only difference between an IP and a Unix server is the address family.
classsocketserver.ForkingMixIn
classsocketserver.ThreadingMixIn
Forking and threading versions of each type of server can be created using these mix-in classes. For instance,
is created as follows:
classThreadingUDPServer(ThreadingMixIn,UDPServer):passThe mix-in class comes first, since it overrides a method defined in
. Setting the various attributes also changes the behavior of the underlying server mechanism.
ForkingMixIn and the Forking classes mentioned below are only available on POSIX platforms that support
.
block_on_close
waits until all child processes complete, except if
attribute is False.
waits until all non-daemon threads complete, except if
attribute is False.
max_children
Specify how many child processes will exist to handle requests at a time for ForkingMixIn. If the limit is reached, new requests will wait until one child process has finished.
daemon_threads
For ThreadingMixIn use daemonic threads by setting
to True to not wait until threads complete.
Changed in version 3.7:
and ThreadingMixIn.server_close now waits until all child processes and non-daemonic threads complete. Add a new
class attribute to opt-in for the pre-3.7 behaviour.
classsocketserver.ForkingTCPServer
classsocketserver.ForkingUDPServer
classsocketserver.ThreadingTCPServer
classsocketserver.ThreadingUDPServer
classsocketserver.ForkingUnixStreamServer
classsocketserver.ForkingUnixDatagramServer
classsocketserver.ThreadingUnixStreamServer
classsocketserver.ThreadingUnixDatagramServer
These classes are pre-defined using the mix-in classes.
Added in version 3.12: The ForkingUnixStreamServer and ForkingUnixDatagramServer classes were added.
To implement a service, you must derive a class from
and redefine its
method. You can then run various versions of the service by combining one of the server classes with your request handler class. The request handler class must be different for datagram or stream services. This can be hidden by using the handler subclasses
or
.
Of course, you still have to use your head! For instance, it makes no sense to use a forking server if the service contains state in memory that can be modified by different requests, since the modifications in the child process would never reach the initial state kept in the parent process and passed to each child. In this case, you can use a threading server, but you will probably have to use locks to protect the integrity of the shared data.
On the other hand, if you are building an HTTP server where all data is stored externally (for instance, in the file system), a synchronous class will essentially render the service “deaf” while one request is being handled – which may be for a very long time if a client is slow to receive all the data it has requested. Here a threading or forking server is appropriate.
In some cases, it may be appropriate to process part of a request synchronously, but to finish processing in a forked child depending on the request data. This can be implemented by using a synchronous server and doing an explicit fork in the request handler class
method.
Another approach to handling multiple simultaneous requests in an environment that supports neither threads nor
(or where these are too expensive or inappropriate for the service) is to maintain an explicit table of partially finished requests and to use
to decide which request to work on next (or whether to handle a new incoming request). This is particularly important for stream services where each client can potentially be connected for a long time (if threads or subprocesses cannot be used).
Server Objects
classsocketserver.BaseServer(server_address, RequestHandlerClass)
This is the superclass of all Server objects in the module. It defines the interface, given below, but does not implement most of the methods, which is done in subclasses. The two parameters are stored in the respective
and
attributes.
fileno()
Return an integer file descriptor for the socket on which the server is listening. This function is most commonly passed to
, to allow monitoring multiple servers in the same process.
handle_request()
Process a single request. This function calls the following methods in order:
,
, and
. If the user-provided
method of the handler class raises an exception, the server’s
method will be called. If no request is received within
seconds,
will be called and handle_request() will return.
serve_forever(poll_interval=0.5)
Handle requests until an explicit
request. Poll for shutdown every poll_interval seconds. Ignores the
attribute. It also calls
, which may be used by a subclass or mixin to provide actions specific to a given service. For example, the
class uses service_actions() to clean up zombie child processes.
Changed in version 3.3: Added service_actions call to the serve_forever method.
service_actions()
This is called in the
loop. This method can be overridden by subclasses or mixin classes to perform actions specific to a given service, such as cleanup actions.
Added in version 3.3.
shutdown()
Tell the
loop to stop and wait until it does. shutdown() must be called while serve_forever() is running in a different thread otherwise it will deadlock.
server_close()
Clean up the server. May be overridden.
address_family
The family of protocols to which the server’s socket belongs. Common examples are
,
, and
. Subclass the TCP or UDP server classes in this module with class attribute address_family=AF_INET6 set if you want IPv6 server classes.
RequestHandlerClass
The user-provided request handler class; an instance of this class is created for each request.
server_address
The address on which the server is listening. The format of addresses varies depending on the protocol family; see the documentation for the
module for details. For internet protocols, this is a tuple containing a string giving the address, and an integer port number: ('127.0.0.1',80), for example.
socket
The socket object on which the server will listen for incoming requests.
The server classes support the following class variables:
allow_reuse_address
Whether the server will allow the reuse of an address. This defaults to
, and can be set in subclasses to change the policy.
request_queue_size
The size of the request queue. If it takes a long time to process a single request, any requests that arrive while the server is busy are placed into a queue, up to
requests. Once the queue is full, further requests from clients will get a “Connection denied” error. The default value is usually 5, but this can be overridden by subclasses.
socket_type
The type of socket used by the server;
and
are two common values.
timeout
Timeout duration, measured in seconds, or
if no timeout is desired. If
receives no incoming requests within the timeout period, the
method is called.
There are various server methods that can be overridden by subclasses of base server classes like
; these methods aren’t useful to external users of the server object.
finish_request(request, client_address)
Actually processes the request by instantiating
and calling its
method.
get_request()
Must accept a request from the socket, and return a 2-tuple containing the new socket object to be used to communicate with the client, and the client’s address.
handle_error(request, client_address)
This function is called if the
method of a
instance raises an exception. The default action is to print the traceback to standard error and continue handling further requests.
Changed in version 3.6: Now only called for exceptions derived from the
class.
handle_timeout()
This function is called when the
attribute has been set to a value other than
and the timeout period has passed with no requests being received. The default action for forking servers is to collect the status of any child processes that have exited, while in threading servers this method does nothing.
process_request(request, client_address)
Calls
to create an instance of the
. If desired, this function can create a new process or thread to handle the request; the
and
classes do this.
server_activate()
Called by the server’s constructor to activate the server. The default behavior for a TCP server just invokes
on the server’s socket. May be overridden.
server_bind()
Called by the server’s constructor to bind the socket to the desired address. May be overridden.
verify_request(request, client_address)
Must return a Boolean value; if the value is
, the request will be processed, and if it’s
, the request will be denied. This function can be overridden to implement access controls for a server. The default implementation always returns True.
Changed in version 3.6: Support for the
protocol was added. Exiting the context manager is equivalent to calling
.
Request Handler Objects
classsocketserver.BaseRequestHandler
This is the superclass of all request handler objects. It defines the interface, given below. A concrete request handler subclass must define a new
method, and can override any of the other methods. A new instance of the subclass is created for each request.
setup()
Called before the
method to perform any initialization actions required. The default implementation does nothing.
handle()
This function must do all the work required to service a request. The default implementation does nothing. Several instance attributes are available to it; the request is available as
; the client address as
; and the server instance as
, in case it needs access to per-server information.
The type of
is different for datagram or stream services. For stream services, request is a socket object; for datagram services, request is a pair of string and socket.
finish()
Called after the
method to perform any clean-up actions required. The default implementation does nothing. If
raises an exception, this function will not be called.
request
The new
object to be used to communicate with the client.
client_address
Client address returned by
.
server
object used for handling the request.
classsocketserver.StreamRequestHandler
classsocketserver.DatagramRequestHandler
These
subclasses override the
and
methods, and provide
and
attributes.
rfile
A file object from which receives the request is read. Support the
readable interface.
wfile
A file object to which the reply is written. Support the
writable interface
Changed in version 3.6:
also supports the
writable interface.
Examples
Example
This is the server side:
importsocketserverclassMyTCPHandler(socketserver.BaseRequestHandler):""" The request handler class for our server. It is instantiated once per connection to the server, and must override the handle() method to implement communication to the client. """defhandle(self):# self.request is the TCP socket connected to the clientpieces=[b'']total=0whileb'\n'notinpieces[-1]andtotal<10_000:pieces.append(self.request.recv(2000))total+=len(pieces[-1])self.data=b''.join(pieces)print(f"Received from {self.client_address[0]}:")print(self.data.decode("utf-8"))# just send back the same data, but upper-casedself.request.sendall(self.data.upper())# after we return, the socket will be closed.if__name__=="__main__":HOST,PORT="localhost",9999# Create the server, binding to localhost on port 9999withsocketserver.TCPServer((HOST,PORT),MyTCPHandler)asserver:# Activate the server; this will keep running until you# interrupt the program with Ctrl-Cserver.serve_forever()An alternative request handler class that makes use of streams (file-like objects that simplify communication by providing the standard file interface):
classMyTCPHandler(socketserver.StreamRequestHandler):defhandle(self):# self.rfile is a file-like object created by the handler.# We can now use e.g. readline() instead of raw recv() calls.# We limit ourselves to 10000 bytes to avoid abuse by the sender.self.data=self.rfile.readline(10000).rstrip()print(f"{self.client_address[0]} wrote:")print(self.data.decode("utf-8"))# Likewise, self.wfile is a file-like object used to write back# to the clientself.wfile.write(self.data.upper())The difference is that the readline() call in the second handler will call recv() multiple times until it encounters a newline character, while the first handler had to use a recv() loop to accumulate data until a newline itself. If it had just used a single recv() without the loop it would just have returned what has been received so far from the client. TCP is stream based: data arrives in the order it was sent, but there is no correlation between client send() or sendall() calls and the number of recv() calls on the server required to receive it.
This is the client side:
importsocketimportsysHOST,PORT="localhost",9999data=" ".join(sys.argv[1:])# Create a socket (SOCK_STREAM means a TCP socket)withsocket.socket(socket.AF_INET,socket.SOCK_STREAM)assock:# Connect to server and send datasock.connect((HOST,PORT))sock.sendall(bytes(data,"utf-8"))sock.sendall(b"\n")# Receive data from the server and shut downreceived=str(sock.recv(1024),"utf-8")print("Sent: ",data)print("Received:",received)The output of the example should look something like this:
Server:
$ pythonTCPServer.py 127.0.0.1 wrote:b'hello world with TCP'127.0.0.1 wrote:b'python is nice'Client:
$ pythonTCPClient.pyhelloworldwithTCP Sent: hello world with TCPReceived: HELLO WORLD WITH TCP$ pythonTCPClient.pypythonisnice Sent: python is niceReceived: PYTHON IS NICE
Example
This is the server side:
importsocketserverclassMyUDPHandler(socketserver.BaseRequestHandler):""" This class works similar to the TCP handler class, except that self.request consists of a pair of data and client socket, and since there is no connection the client address must be given explicitly when sending data back via sendto(). """defhandle(self):data=self.request[0].strip()socket=self.request[1]print(f"{self.client_address[0]} wrote:")print(data)socket.sendto(data.upper(),self.client_address)if__name__=="__main__":HOST,PORT="localhost",9999withsocketserver.UDPServer((HOST,PORT),MyUDPHandler)asserver:server.serve_forever()This is the client side:
importsocketimportsysHOST,PORT="localhost",9999data=" ".join(sys.argv[1:])# SOCK_DGRAM is the socket type to use for UDP socketssock=socket.socket(socket.AF_INET,socket.SOCK_DGRAM)# As you can see, there is no connect() call; UDP has no connections.# Instead, data is directly sent to the recipient via sendto().sock.sendto(bytes(data+"\n","utf-8"),(HOST,PORT))received=str(sock.recv(1024),"utf-8")print("Sent: ",data)print("Received:",received)The output of the example should look exactly like for the TCP server example.
Asynchronous Mixins
To build asynchronous handlers, use the
and
classes.
An example for the
class:
importsocketimportthreadingimportsocketserverclassThreadedTCPRequestHandler(socketserver.BaseRequestHandler):defhandle(self):data=str(self.request.recv(1024),'ascii')cur_thread=threading.current_thread()response=bytes("{}: {}".format(cur_thread.name,data),'ascii')self.request.sendall(response)classThreadedTCPServer(socketserver.ThreadingMixIn,socketserver.TCPServer):passdefclient(ip,port,message):withsocket.socket(socket.AF_INET,socket.SOCK_STREAM)assock:sock.connect((ip,port))sock.sendall(bytes(message,'ascii'))response=str(sock.recv(1024),'ascii')print("Received: {}".format(response))if__name__=="__main__":# Port 0 means to select an arbitrary unused portHOST,PORT="localhost",0server=ThreadedTCPServer((HOST,PORT),ThreadedTCPRequestHandler)withserver:ip,port=server.server_address# Start a thread with the server -- that thread will then start one# more thread for each requestserver_thread=threading.Thread(target=server.serve_forever)# Exit the server thread when the main thread terminatesserver_thread.daemon=Trueserver_thread.start()print("Server loop running in thread:",server_thread.name)client(ip,port,"Hello World 1")client(ip,port,"Hello World 2")client(ip,port,"Hello World 3")server.shutdown()The output of the example should look something like this:
$ pythonThreadedTCPServer.py Server loop running in thread: Thread-1Received: Thread-2: Hello World 1Received: Thread-3: Hello World 2Received: Thread-4: Hello World 3The
class is used in the same way, except that the server will spawn a new process for each request. Available only on POSIX platforms that support
.