raisin 3.0.1

Simple parallel, distributed and cluster computing



Raisin: To perform cluster work easily!

Project Philosophy

The main aim of projectraisin is to share physical resources of your laptop with a community.

In counterpart, you can benefit from the community resources.

There are 2 sides in this project:

1. Resources usage

The raisin API wants to be as close as possible to the ‘threading’ and ‘multiprocessing’ python APIs.

The advantage in using raisin rather than ‘threading’ or ‘multiprocessing’ is that the computing power is greatly increased (depending on the number of connected resources).

Though raisin is based on ‘multiprocessing’ module - that splits tasks among the resources of a single computer - it also shares the load over the different machines in the network. Everything is automatically and intelligently orchestrated relying on code analysis and graph theory.

raisin wants to be as simple as possible. That’s why the code analysis and the resources management are automated. It also uses a bunch of classes and functions default parameters that are suitable for most usages.

However, you can tune raisin behavior as you want since all these parameters are fully customizable.

raisin is a multi-OS module 100% written in python in order to keep installation reliable and simple.

Although raisin uses powerful modules such as ‘sympy’, ‘numpy’, ‘giacpy’, ‘pycryptodomex’, ‘tkinter’… these modules are not required (they are sometimes not easy to install). This will just lead to less efficiency, but no failure!

In a future version, raisin will be able to perform automatic parallelization, a little like ‘pydron’.

2. Resources sharing

To be able to use community resources, you must give in return!

That’s why, when raisin is installed as a python package, you have to install the ‘application’ part.

To do this, execute the python3 raisin/communication/server.py command.

Examples

server code

import raisin

with raisin.Server(20001) as server:
    server.serve_forever()

client code

import raisin
from raisin.encapsulation import make_tasks

def my_function(x):
    """just a test"""
    return x**2

tasks_gen = make_tasks((f, f, f), ([2], [3], [4]))

with raisin.TasksSharer(tasks_gen) as tasks_sharer:
    tasks_sharer.start()
    for res in tasks_sharer:
        print(res)

• See the documentation for more details and examples.