Skip to content

Latest commit

 

History

History
46 lines (34 loc) · 1.56 KB

manual.md

File metadata and controls

46 lines (34 loc) · 1.56 KB
Raw

Design philosophy

The idea is to design loosely connected modules. The package consists of different modules that can be used independently up to a point.

For installing the package, please install the requirement file

conda install --file requirements.txt or pip install -r requirements.txt

Overview

The main division can be found as follows:

  • Modeling
  • Methods
  • Plotting

Please note that a unittest is written for nearly every different main file

Modeling

Models form a general model that can be found in models.py. Through inheritance one can define their own modes. Note that only a minimal set of properties are required for the different models

Modules:

  • models.py : contains the main model mold [inherit your model from this main mold]
  • fastIsing.py: numpy orientated Ising model
  • XOR.py: contains XOR and AND functions for testing simple properties of methods -simulate.py: currently only contains simulate for simulation, the idea is to expand this or rework present functions for better cognitive separation

Methods

This module contains the methods for the different information theoretical measures. Currently, IDT for source and sinks are implemented.

Modules:

  • information.py: IDT and sampling methods (MCMC) and other miscellaneous functions

plotting

This part of the model is responsible for plotting various parts of the results.

Modules

  • plotting.py contains functions to show graph (including adjacency) next to fits for computing the IDT as well as computing the impact of the nudge on node distribution using the Hellinger distance.