Axon.background

The background class makes it easy to run an Axon system in a separate thread (in effect: in the background).

This simplifies integration of Axon/Kamaelia code into other python code. See also Axon.Handle for a simple way to wrap a component in a thread safe way to access its inboxes and outboxes.

Example Usage

At its simplest, you could run a Kamaelia task independently in the background - such as a simple network connection, that dumps received data into a thread safe queue, after de-chunking it into lines of text.

NOTE: This example can be achieved more simply by using Axon.Handle. See the documentation of Axon.Handle to find out more.

1. We implement a simple component to collect the data:

   from Axon.background import background
   from Axon.Component import component
   
   class Receiver(component):
       def __init__(self, queue):
           super(Bucket,self).__init__()
           self.q = queue
   
       def main(self):
           while 1:
               while self.dataReady("inbox"):
                   self.q.put(self.recv("inbox"))
               self.pause()
               yield 1

2. Then we create a background object and call its start() method:

   from Axon.background import background
   
   background().start()

3. Finally, we create and activate our Kamaelia pipeline of components, including the receiver component we've just written, passing it a thread-safe queue to put the data into:

   from Kamaelia.Chassis.Pipeline import Pipeline
   from Kamaelia.Internet.TCPClient import TCPClient
   from Kamaelia.Visualisation.PhysicsGraph import chunks_to_lines
   from Queue import Queue
   
   queue = Queue()
   
   Pipeline(
       TCPClient("my.server.com", 1234),
       chunks_to_lines(),
       Receiver(queue)
   ).activate()

We can now fetch items of data, from the queue when they arrive:

received_line = queue.get()

Behavour

Create one of these and start it running by calling its start() method.

After that, any components you activate will default to using this scheduler.

Only one instance can be used within a given python interpreter.

The background thread is set as a "daemon" thread. This means that if your program exits, this background thread will be killed too. If it were not a daemon, then it would prevent the python interpreter terminating until the components running in it had all terminated too.

A python thread which runs the Axon Scheduler. Takes the same arguments at creation that Axon.Scheduler.scheduler.run.runThreads accepts.

Create one of these and start it running by calling its start() method.

After that, any components you activate will default to using this scheduler.

Only one instance can be used within a given python interpreter.