API Reference

This reference is generated from the public namespace of each dit subpackage, so it lists every function and class you can import from it. The narrative pages elsewhere in the documentation explain what the measures mean and show worked examples; this section is the place to look up exact signatures.

Subpackages re-export the contents of their nested modules. For example, dit.multivariate includes the common informations and secret key agreement bounds, dit.pid includes every PID_* measure, and dit.rate_distortion includes the Gray–Wyner tools.

Core

`dit`	dit is a Python package for information theory on discrete random variables.
`dit.example_dists`	A selection of generic distributions which could be useful.
`dit.symbolic`	Symbolic (sympy-backed) distributions and helpers.
`dit.math`	Mathematical tools for dit.
`dit.utils`	Module providing miscellaneous functionality.

Measures

`dit.shannon`	The basic forms of Shannon's information measures.
`dit.multivariate`	Multivariate measures of information.
`dit.divergences`	Divergences, measuring the distance between distributions.
`dit.other`	Esoteric measures of information, typically fairly divorced from Shannon's measures.

Decompositions and Profiles

`dit.pid`	Import all the PIDs.
`dit.profiles`	This module contains various information "profiles", decomposing the total information into the information contained at different scales.
`dit.visualization`	Visualizations of multivariate distributions.

Channels, Coding, and Rate-Distortion

`dit.example_channels`	A catalog of canonical discrete memoryless channels.
`dit.channelorder`	Channel comparison: orderings and deficiencies.
`dit.coding`	Coding: building source and channel codes.
`dit.rate_distortion`	Various implementations of rate-distortion theory.

Estimation and Optimization

`dit.inference`	Module for basic inference tools.
`dit.algorithms`	Various algorithms related to manipulating or measuring properties of distributions.