Python React

A Reactive Programming core for Python

Started as a sample on how Python's attibute models allow for intelligent objects in a presentation, the aim of this module is to become a simple, yet powerful helper for reactive programming in Python.

Check https://en.wikipedia.org/wiki/Reactive_programming for more on reactive programing.

The simple principle is to leverage on Python's '''setattr''' hook to both stablish propagating rules and apply rules on attribute change.

Usage

Simply instantiate a react.Reactor class - any assignment of a function or react.Rule class to an attribute on this class will create a new Rule - each time an attribute with the same name as the parameters on this function (or attrs parameter for a Rule) is assigned a new (non-Rule) value, the Rule action is called with that value (and the values for other parameters it depends on).

For a fast start on quick hacks, there is an "R" reactor instance ready to use:

For example:

>>> from react import R
>>>
>>> R.greeting = lambda name, age: "Hello {}! How  nice you are {} years old!".format(name, age)
>>> R.name = "João"
>>> R.age = 40
>>> R.greeting
'Hello João! How  nice you are 40 years old!'
>>> R.age = 25
>>> R.greeting
'Hello João! How  nice you are 25 years old!'

A couple things to take note for now:

Although the simple rule examples are with lambda functions, the Rule callable can be an arbitrary Python callable - even making I/O or network calls with any values.
Any created rules are independent of the actual value stored in the attributes. They are stored, and can be read as they are set by their name, using getattr or the instance's __dict__. The rules are kept in a _rules data structure on the instance. This is much like a spreadsheet where one does see the calculated values, but not the underlying formulas for each cell.
The propagation on the current Reactor class is synchronous: when an attribute is set on the instance, all rules are checked and calculated before the execution flow comes back to the frame where the assignment statement took place.

Compatibility:

The code so far has been tested with cPython 2.7, cPython 3.4 and pypy 2.4.0 - shour work fine for other Python implementations.

There are simple doctests (which also serve as examples) in the main file - run them with python -m doctest react.py

TODO:

This is still an incipient project - if you see have any possible usage for this, please jump in for colaboration.

Next steps are: better and more complete examples, possibility to watch over structured values like lists and dicts (probably using proxies for those), asynchronous value propagation, and the possibility to do interprocess value propagation (with a to-be-decided trasnport channel like memcached, celery, etc...)

COPYING: