Massive refactor and optimization

[hadware]

Here's what's changing:

• Updated the continuum structure to better match the data format
• Added support for continuum construction from csv, textgrid, pyannote objects
• Massive speedup using numba for the disorder computations
• Removed sequence dissimilarity (not supported by the numba implem)
• Type hinted most of the API
• added a CLI command
• started writing a documentation

[Rachine]

Tests with 3by100 and 2by1000 should be added and we are good to go.

[Rachine]

It's missing the csv data

FileNotFoundError: [Errno 2] No such file or directory: 'data/2by1000.csv' could you push it?

[Rachine]

Seems good to me! 👍

This PR speeds up greatly the Gamma computations.

The MIP problem will need to be changed in the Future.