Hubris

Hubris is a simple library for implementing inter-process communication (IPC) between separate Python processes. It's based on named pipes created by the POSIX mkfifo system call.

Hubris uses the "channel" metaphor. You can send and receive data on a with simple send and receive commands. Alternatively, you can have multiple processes subscribe to the a single channel on which data is published asynchronously or at a regular interval.

Sending Data

hub = Hub()
channel = hub['example']

# send data (which is serialized internally)
channel.send({'foo': 'bar'})

hub.close()

Receiving Data

Receiving data can be done synchronously or asynchronously.

hub = Hub()
channel = hub['example']

# block until we receive the data
data = channel.receive()

# ...or immediately return a future object
future = channel.receive(wait=False)
data = future.result(timeout=1)

hub.close()

Publishing Data

Data can be published at a regular interval.

hub = Hub()
channel = hub['example']

# send data (which is serialized internally)
channel.publish({'foo': 'bar'}, interval=timedelta(seconds=1))

# ...or generate data with a callback
channel.publish(generate, interval=timedelta(seconds=1))

# ...or just send data as normally to have it published to subscribers
channel.send({'foo': 'bar'})

hub.close()

Subscribing to Channels

You can subscribe a callback to receive and respond to data published on a channel.

hub = Hub()
channel = hub['example']

# subscribe this thread to the channel, triggering
# callback upon receipt of data
channel.subscribe(lambda: subscription, data: print(data))

hub.close()

Recommend Projects