simpy-events

event system with simpy to decouple simulation code and increase reusability


License
MIT
Install
pip install simpy-events==0.0.2

Documentation

simpy-events

license python version build-status Documentation Status coverage pypi package

event system with SimPy to decouple simulation code and increase reusability

( >>>>>>> WORK IN PROGRESS <<<<<<< )

A basic example

Note

SimPy is a process-based discrete-event simulation framework based on standard Python.

  • Our simplified scenario is composed of:

    • satellites emitting signals
    • receivers receiving and processing signals
  • basic imports and creating the root namespace:

from simpy_events.manager import RootNameSpace
import simpy

root = RootNameSpace()
  • implementing a satellite model:
sat = root.ns('satellite')

class Satellite:
    chunk = 4

    def __init__(self, name, data):
        self.signal = sat.event('signal', sat=name)
        self.data = tuple(map(str, data))

    def process(self, env):
        signal = self.signal
        data = self.data
        chunk = self.chunk
        # slice data in chunks
        for chunk in [data[chunk*i:chunk*i+chunk]
                      for i in range(int(len(data) / chunk))]:
            event = env.timeout(1, ','.join(chunk))
            yield signal(event)
  • implementing a receiver model:
receiver = root.ns('receiver')
signals = receiver.topic('signals')

@signals.after
def receive_signal(context, event):
    env = event.env
    metadata = context.event.metadata
    header = str({key: val for key, val in metadata.items()
                  if key not in ('name', 'ns')})
    env.process(process_signal(env, header, event.value))

def process_signal(env, header, signal):
    receive = receiver.event('process')
    for data in signal.split(','):
        yield receive(env.timeout(0, f'{header}: {data}'))
  • creating code to analyse what's going on:
@root.enable('analyse')
def new_process(context, event):
    metadata = context.event.metadata
    context = {key: str(val) for key, val in metadata.items()}
    print(f'new signal process: {context}')

@root.after('analyse')
def signal(context, event):
    metadata = context.event.metadata
    ns = metadata['ns']
    print(f'signal: {ns.path}: {event.value}')
  • setting up our simulation:
root.topic('receiver::signals').extend([
    '::satellite::signal',
])
root.topic('analyse').extend([
    '::satellite::signal',
    '::receiver::process',
])

def run(env):
    # create some actors
    s1 = Satellite('sat1', range(8))
    s2 = Satellite('sat2', range(100, 108))
    env.process(s1.process(env))
    env.process(s2.process(env))

    # execute
    root.enabled = True
    env.run()
  • running the simulation

    new signal process: {'ns': '::satellite', 'name': 'signal', 'sat': 'sat1'}
    new signal process: {'ns': '::satellite', 'name': 'signal', 'sat': 'sat2'}
    signal: ::satellite: 0,1,2,3
    new signal process: {'ns': '::receiver', 'name': 'process'}
    signal: ::satellite: 100,101,102,103
    new signal process: {'ns': '::receiver', 'name': 'process'}
    signal: ::receiver: {'sat': 'sat1'}: 0
    signal: ::receiver: {'sat': 'sat2'}: 100
    signal: ::receiver: {'sat': 'sat1'}: 1
    signal: ::receiver: {'sat': 'sat2'}: 101
    signal: ::receiver: {'sat': 'sat1'}: 2
    signal: ::receiver: {'sat': 'sat2'}: 102
    signal: ::receiver: {'sat': 'sat1'}: 3
    signal: ::receiver: {'sat': 'sat2'}: 103
    signal: ::satellite: 4,5,6,7
    new signal process: {'ns': '::receiver', 'name': 'process'}
    signal: ::satellite: 104,105,106,107
    new signal process: {'ns': '::receiver', 'name': 'process'}
    signal: ::receiver: {'sat': 'sat1'}: 4
    signal: ::receiver: {'sat': 'sat2'}: 104
    signal: ::receiver: {'sat': 'sat1'}: 5
    signal: ::receiver: {'sat': 'sat2'}: 105
    signal: ::receiver: {'sat': 'sat1'}: 6
    signal: ::receiver: {'sat': 'sat2'}: 106
    signal: ::receiver: {'sat': 'sat1'}: 7
    signal: ::receiver: {'sat': 'sat2'}: 107
    

install and test

install from pypi

using pip:

$ pip install simpy-events

dev install

There is a makefile in the project root directory:

$ make dev

Using pip, the above is equivalent to:

$ pip install -r requirements-dev.txt
$ pip install -e .

run the tests

Use the makefile in the project root directory:

$ make test

This runs the tests generating a coverage html report

build the doc

The documentation is made with sphinx, you can use the makefile in the project root directory to build html doc:

$ make doc

Documentation

Documentation on Read The Docs.

Meta

loicpw - peronloic.us@gmail.com

Distributed under the MIT license. See LICENSE.txt for more information.

https://github.com/loicpw