v0.3.0 beta - Presentation

So named because it is this version that is used for the second (and presumably final) submission for the project that initially inspired the development of this module.

Improvements

• Better encapsulation and modularity.
• Better cross-platform compatibility (os.path.join()).
• Less terrible calculation of statistical significance.

New Features

• make_dir(): Create directory if non-existent.
• most_common(): Rank most common values within specified constraints.
• most_assoc(): Rank most associated pairs within specified constraints.
• extremes(): Rank most extreme associations among all data.
• savefig(): More encapsulated figure saving.
• AsciiTable(): Table methods have been split off from Analysis() into its own class.

v0.2.1 - Initial Beta Release

I have decided to move from alpha to beta (which is a very arbitrary distinction for a project this size) since the state of the project has transitioned from "it works but has many obvious flaws" to "it works but may have several non-obvious flaws."

The following is the addition to the changelog since the previous release.

0.2.0 - Not Released (2017-01-01)

This version saw major improvements but was not released because there were some issues.

Improvements

Use general formula for association ratio. The internals of Associator().find() are completely new.

The following are deprecated by this change:

• Associator().overall_ratios()
• Associator().test()
• Associator().relevant
• Associator().memo

New Features

Further encapsulate Histogram() object by providing the following:

• Histogram().reduce() (this method is also used to shorten Histogram.simplify())
• Histogram().slice()
• Histogram().nonzeros()
• Histogram().nonzero_indices

0.2.1 - Second Release (2017-01-01)

Bug Fixes

Reintroduce support for Associator().notable and Associator().significant.

Improvements

Clean up excess code, remove old methods, rename new methods to take their place (as they should be according to 0.2.0 changelog).

Associations v0.1.0 alpha

This is a fully functioning version of the Associations Python module. It still needs some work to be optimized and prepared for general use. The first thing we need to do is see about implementing the general formula in associations.py. While working on that, there may be some breakages in commits coming after this release. So this release is being rolled out as a tentative working version that may be favorable over the latest commit until the next release.