A basic implementation of Go inspired channels.
Use these channels to communicate across Python threads, processes, or hosts, independent of network location. Channels are first class objects that can be passed themselves through the channels.
Start a RabbitMQ instance by executing the following command in the root directory of this repository:
docker-compose up -d
Create a channel with:
>>> from channelpy import Channel >>> ch = Channel()
Put and get objects to and from the channel:
>>> ch.put('foo') >>> ch.get() 'foo' >>> ch.get(timeout=1) # raises ChannelTimeoutException since channel is empty
Pass channels into channels:
>>> ch1 = Channel() >>> ch1.put(5) >>> ch2 = Channel() >>> ch2.put(ch1) >>> x = ch2.get() >>> x.get() 5
Channels can be instantiated by name:
>>> ch = Channel() >>> ch.name 'daa0a490f9254c69883335c9f925d74f' >>> another = Channel(name=ch.name) >>> another.put('foo') >>> ch.get() 'foo'
Or create them with a specific name:
>>> ch = Channel(name='my_channel')
Multiple consumers and producers can be attached to a channel. Messages are delivered in round-robin fashion to consumers. However, a
close_allcan be delivered to all the consumers to signal a full termination of the channel.
>>> ch = Channel() >>> def f(c): ... while True: ... print(c.get()) ... >>> threading.Thread(target=f, args=(ch,)).start() >>> threading.Thread(target=f, args=(ch,)).start() >>> ch.put(4) # one of the threads will print 4 >>> ch.close_all() # ChannelClosedException is raised in all threads
The broker to use can be configured at instantiation time or by using the config file
~/.channelpy.yml. For example:
connection: RabbitConnection uri: amqp://192.168.35.10:5672
Run the tests with:
$ BROKER='amqp://localhost:5672' py.test -v
To be written.