fsmpy

Minimal state machine


Keywords
finite, state, machine, minimal, finite-state-machine, fsm, python, state-machine
License
MIT
Install
pip install fsmpy==2.1.0

Documentation

Overview

docs Read the Docs
tests
Travis-CI Build Status
Coverage Status
package PyPI Package latest release PyPI Wheel Supported versions Supported implementations

Minimal state machine

  • Free software: BSD license
import fsm

class MyTasks(fsm.FiniteStateMachineMixin):
    """An example to test the state machine.

    Contains transitions to everywhere, nowhere and specific states.
    """

    state_machine = {
        'created': '__all__',
        'pending': ('running',),
        'running': ('success', 'failed'),
        'success': None,
        'failed': ('retry',),
        'retry': ('pending', 'retry'),
    }

    def __init__(self, state):
        """Initialize setting a state."""
        self.state = state

    def on_before_pending(self):
        print("I'm going to a pending state")
In [4]: m = MyTasks(state='created')

In [5]: m.change_state('pending')
I'm going to a pending state
Out[5]: 'pending'
In [6]: m.change_state('failed')  # Let's try to transition to an invalid state
---------------------------------------------------------------------------
InvalidTransition                         Traceback (most recent call last)
<ipython-input-6-71d2461eee74> in <module>()
----> 1 m.change_state('failed')

~/pyfsm/src/fsm/fsm.py in change_state(self, next_state, **kwargs)
    90             msg = "The transition from {0} to {1} is not valid".format(previous_state,
    91                                                                        next_state)
---> 92             raise InvalidTransition(msg)
    93
    94         name = 'pre_{0}'.format(next_state)

InvalidTransition: The transition from pending to failed is not valid

Installation

pip install fsmpy

Usage

  1. Define in a class the state_machine
  2. Initialize state, either with a value, using __init__ or as a django field
  3. Add hooks:
Method Description
on_before_change_state Before transitioning to the state
on_change_state After transitioning to the state, if no failure, runs for every state
pre_<state_name> Runs before a particular state, where state_name is the specified name in the state_machine
post_<state_name> Runs after a particular state, where state_name is the specified name in the state_machine

This hooks will receive any extra argument given to change_state

E.g:

Running m.change_state('pending', name='john') will trigger pre_pending(name='john')

Django integration

import fsm
from django.db import models


class MyModel(models.Model, fsm.FiniteStateMachineMixin):
    """An example to test the state machine.

    Contains transitions to everywhere, nowhere and specific states.
    """

    CHOICES = (
        ('created', 'CREATED'),
        ('pending', 'PENDING'),
        ('running', 'RUNNING'),
        ('success', 'SUCCESS'),
        ('failed', 'FAILED'),
        ('retry', 'RETRY'),
    )

    state_machine = {
        'created': '__all__',
        'pending': ('running',),
        'running': ('success', 'failed'),
        'success': None,
        'failed': ('retry',),
        'retry': ('pending', 'retry'),
    }

    state = models.CharField(max_length=30, choices=CHOICES, default='created')

    def on_change_state(self, previous_state, next_state, **kwargs):
        self.save()

Django Rest Framework

If you are using serializers, they usually perform the save, so saving inside on_change_state is not necessary.

One simple solution is to do this:

class MySerializer(serializers.ModelSerializer):

    def update(self, instance, validated_data):
        new_state = validated_data.get('state', instance.state)
        try:
            instance.change_state(new_state)
        except fsm.InvalidTransition:
            raise serializers.ValidationError("Invalid transition")
        instance = super().update(instance, validated_data)
        return instance

Documentation

https://pyfsm.readthedocs.org/

Development

To run the tests run:

tox

Note, to combine the coverage data from all the tox environments run:

Windows
set PYTEST_ADDOPTS=--cov-append
tox
Other
PYTEST_ADDOPTS=--cov-append tox