Quickstart

Declaring a pyfase compatible class

from pyfase import MicroService

class Class_1(MicroService):
    def __init__(self):
        super(Class_1, self).__init__(self, 
                                    sender_endpoint='ipc:///tmp/sender', 
                                    receiver_endpoint='ipc:///tmp/receiver')

    @MicroService.action
    def do_something_amazing(self, service, data):
        print('doing something amazing!')
        # [...]
        self.request_action('handle_amazingness', {'amazing': '10/10'})

Class_1().execute()

Here Class_1 inherits from MicroService, your window of comunication with other processes. That first line in __init__ is what communicates to core.py that your process exists and interfaces it with other ones.

But it alone won’t do any good, it still needs to have it’s @MicroService.action called from another process:

from pyfase import MicroService
from time import sleep

class Class_2(MicroService):
    def __init__(self):
        super(Class_2, self).__init__(self, 
                                    sender_endpoint='ipc:///tmp/sender', 
                                    receiver_endpoint='ipc:///tmp/receiver')

    @MicroService.action
    def handle_amazingness(self, service, data):
        print('amazing response: {}'.format(data['amazing']))

    @MicroService.task
    def request_amazingness():
        while True:
            sleep(10)
            message = dict(how_much_amazing='really amazing')
            self.request_action('do_something_amazing', message)

Class_2().execute(enable_tasks=True)

Having done that, all that is left is to run core.py, the script that acts as the broker for those processes.

Once all three processes are running, every 10 seconds Class_2 will request a action called do_something_amazing and send message as data parameter. Every process that is connected to this broker will listen to this request, and if it has a function by this name and that function is decorated by @MicroService.action, it will then execute the action.

Head over to the examples folder if you want to see more.

Known Limitations

• The data parameter for sending and receiving messages has to be JSON serializable
• For Windows users the ipc isn’t available. This can be solved by replacing those two parameters in __init__ method, as well as on the core.py:

  sender_endpoint='tcp://localhost:8001/'
  receiver_endpoint='tcp://localhost:8002/'
  

  Or any other ip:port available for you.

Next: Decorators

Home