Simple and easy to use syncronous WebSocket client, no async.
Usable with Jupyter Notebook.
Install using pip
pip install nwebsocket
Python 3.7+ is required.
A simple example without classes
# example_minimal.py
import time
from nwebsocket import WebSocket
wscn = WebSocket("wss://ws.postman-echo.com/raw")
wscn.onmessage = lambda m: print(m)
wscn.onopen = lambda: print("Opened connection")
wscn.onclose = lambda: print("Closed connection")
wscn.onerror = lambda e: print("Connection error", e)
while(not wscn.readyState):
time.sleep(1e-4)
wscn.send('text')
time.sleep(1.)
wscn.close()Example of extending the WebSocket class.
# example_class.py
import time
from nwebsocket import WebSocket
class WSProtocolLogic(WebSocket):
def __init__(self, url):
super().__init__(url)
self.messages = []
# wait for connection, close or error
while(not self.readyState):
time.sleep(1e-4)
def onopen(self):
print("Opened connection")
def onclose(self):
print("Closed connection")
def onerror(self, e):
print("Connection error", e)
def onmessage(self, m):
self.messages.append(m)
wscn = WSProtocolLogic("wss://ws.postman-echo.com/raw")
wscn.send('text')
time.sleep(1.)
print(wscn.messages)
wscn.close()There are many asynchronous Python WebSocket client packages out there, and almost of them require your code to use the async syntax. This is unfortunate, since it will lead to a run_until_complete call eventually, which will block the main thread from performing other operations in parrallel.
This package was inspired by the ultra-simple WebSocket API in the JavaScript language, which it replicates one-to-one.
https://developer.mozilla.org/en-US/docs/Web/API/WebSocket/url
Since the callback-style API is quite unusual when it comes to pythonicity, your task will be to:
- define onmessage, onopen, onclose and onerror
- handle reconnection/s
- implement the TX/RX specification for working with the endpoint
- isolate the callback pattern from the rest of your code
Take a chat service as an example, there are operations which are following the traditional request/response pattern such as posting messages, and there are other messages which are received without issuing a request (chat posting by other users).
These latter messages must be handled by your code as soon as they are received. Class instances can help with that, by storing the received information (chat posts). Keep your callbacks short, fast and serializable.
This library is not suitable for high throughput, as the queue mechanism in Python is notoriously slow due to serialization.
Copyright (C) 2022 Adapta Robotics | MATT Robot