Getting Started¶

This document explains how to get started using wsproto to connect to WebSocket servers as well as how to write your own.

We assume some level of familiarity with writing Python and networking code. If you’re not familiar with these we highly recommend you read up on these first. It may also be helpful to study Sans-I/O, which describes the ideas behind writing a network protocol library that doesn’t do any network I/O.

Connections¶

The main class you’ll be working with is the WSConnection object. This object represents a connection to a WebSocket client or server and contains all the state needed to communicate with the entity at the other end. Whether you’re connecting to a server or receiving a connection from a client, this is the object you’ll use.

wsproto provides two layers of abstractions. You need to write code that interfaces with both of these layers. The following diagram illustrates how your code is like a sandwich around wsproto.

Application

wsproto

Network Layer

wsproto does not do perform any network I/O, so <NETWORK GLUE> represents the code you need to write to glue wsproto to the actual network layer, i.e. code that can send and receive data over the network. The WSConnection class provides two methods for this purpose. When data has been received on a network socket, you feed this data into wsproto by calling receive_bytes. When wsproto has data that needs to be sent over the network, you retrieve that data by calling bytes_to_send, and your code is responsible for actually sending that data over the network.

Note

If the connection drops, a standard Python socket.recv() will return zero. You should call receive_bytes(None) to update the internal wsproto state to indicate that the connection has been closed.

Internally, wsproto process the raw network data you feed into it and turns it into higher level representations of WebSocket events. In <APPLICATION GLUE>, you need to write code to process these events. The WSConnection class contains a generator method events that yields WebSocket events. To send a message, you call the send_data method.

Connecting to a WebSocket server¶

Begin by instantiating a connection object. The host and resource arguments are required to instantiate a client. If the WebSocket server is located at http://myhost.com/foo, then you would instantiate the connection as follows:

ws = WSConnection(ConnectionType.CLIENT, host="myhost.com", resource='foo')

Now you need to provide the network glue. For the sake of example, we will use standard Python sockets here, but wsproto can be integrated with any network layer:

stream = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
stream.connect(("myhost", 8000))

To read from the network:

data = stream.recv(4096)
ws.receive_bytes(data)

You also need to check if wsproto has data to send to the network:

data = ws.bytes_to_send()
stream.send(data)

Note that bytes_to_send() will return zero bytes if the protocol has no pending data. You can either poll this method or call it only when you expect to have pending data.

A standard Python socket will block on the call to stream.recv(), so you will probably need to use a non-blocking socket or some form of concurrency like threading, greenlets, asyncio, etc.

You also need to provide the application glue. To send a WebSocket message:

ws.send_data("Hello world!")

And to receive WebSocket events:

for event in ws.events():
    if isinstance(event, ConnectionEstablished):
        print('Connection established')
    elif isinstance(event, ConnectionClosed):
        print('Connection closed: code={} reason={}'.format(
            event.code, event.reason))
    elif isinstance(event, TextReceived):
        print('Received message: {}'.format(event.data))
    else:
        print('Unknown event: {!r}'.format(event))

The method events() returns a generator which will yield events for all of the data currently in the wsproto internal buffer and then exit. Therefore, you should iterate over this generator after receiving new network data.

For a more complete example, see synchronous_client.py.

WebSocket Servers¶

A WebSocket server is similar to a client except that it uses a different constant:

ws = WSConnection(ConnectionType.SERVER)

A server also needs to explicitly call the accept method after it receives a ConnectionRequested event:

for event in ws.events():
    if isinstance(event, ConnectionRequested):
        print('Accepting connection request')
        ws.accept(event)
    elif isinstance(event, ConnectionClosed):
        print('Connection closed: code={} reason={}'.format(
            event.code, event.reason))
    elif isinstance(event, TextReceived):
        print('Received message: {}'.format(event.data))
    else:
        print('Unknown event: {!r}'.format(event))

For a more complete example, see synchronous_server.py.

Closing¶

WebSockets are closed with a handshake that requires each endpoint to send one frame and receive one frame. The close() method places a close frame in the send buffer. When a close frame is received, it yields a ConnectionClosed event, and it also places a reply frame in the send buffer. When that reply has been received by the initiator, it will also receive a ConnectionClosed event.

Regardless of which endpoint initiates the closing handshake, the server is responsible for tearing down the underlying connection. When the server receives a ConnectionClosed event, it should send pending wsproto data (if any) and then it can start tearing down the underlying connection.

Ping Pong¶

The WSConnection class supports sending WebSocket ping and pong frames via the methods ping and pong.

Note

When a ping is received, wsproto automatically places a pong frame in its outgoing buffer. You should only call pong() if you want to send an unsolicited pong frame.

Back-pressure¶

Back-pressure is an important concept to understand when implementing a client/server protocol. This section briefly explains the issue and then explains how to handle back-pressure when using wsproto.

Imagine that you have a WebSocket server that reads messages from the client, does some processing, and then sends a response. What happens if the client sends messages faster than the the server can process them? If the incoming messages are buffered in memory, then the server will slowly use more and more memory, until the OS eventually kills it. This scenario is directly applicable to wsproto, because every time you call receive_bytes(), it appends that data to an internal buffer.

The slow endpoint needs a way to signal the fast endpoint to stop sending messages until the slow endpoint can catch up. This signaling is called “back-pressure”. As a Sans-IO library, wsproto is not responsible for network concerns like back-pressure, so that responsibility belongs to your network glue code.

Fortunately, TCP has the ability to signal backpressure, and the operating system will do that for you automatically—if you follow a few rules! The OS buffers all incoming and outgoing network data. Standard Python socket methods like send() and recv() copy data to and from those OS buffers. For example, if the peer is sending data too quickly, then the OS receive buffere will start to get full, and the OS will signal the peer to stop transmitting. When recv() is called, the OS will copy data from its internal buffer into your process, free up space in its own buffer, and then signal to the peer to start transmitting again.

Therefore, you need to follow these two rules to implement back-pressure over TCP:

Do not receive from the socket faster than your code can process the messages. Your processing code may need to signal the receiving code when its ready to receive more data.
Do not store out-going messages in an unbounded collection. Ideally, out-going messages should be sent to the OS as soon as possible. If you need to buffer messages in memory, the buffer should be bounded so that it can not grow indefinitely.