For those of you interested in probabilities and probabilistic programming, I’m happy to announce that Lea 2.2.0 is now under beta-test.
What is Lea? ------------ Lea is a Python package aiming at working with discrete probability distributions in an intuitive way. It allows you to model a broad range of random phenomenons, like dice throwing, coin tossing, gambling, weather, etc. It offers several high-level modelling features for probabilistic programming, including bayesian inference and Markov chains. Lea is open-source (LGPL) and runs on Python 2 or 3. See project page below for more information (installation, tutorials, examples, etc).
What’s new? ----------- Compared to latest version (2.1.2), many things have been made in 2.2.0 to improve ease-of-use and overall performance, without breaking backward compatibility. Maybe one of the most notable feature is that you can now get individual probability very easily, as a fraction or float, thanks to the new 'P' and 'Pf' functions, e.g.
P(dice <= 5)
Pf(dice <= 5)
New methods allow you to read a CSV file or Pandas dataframe, then build the corresponding joint probability distribution. Also, Monte-Carlo sampling estimation is now available, should Lea’s default exact evaluation is intractable. Most of the new features are documented in a new tutorial on Lea's wiki (https://bitbucket.org/piedenis/lea/wiki/LeaPyTutorial3).
The latest version, Lea 2.2.0-beta.4, is fairly stable (no known bug) so you can start to use it and report any problem or dislike, if any.
Lea project page ---------------- https://bitbucket.org/piedenis/lea
Download Lea (PyPI) ------------------- http://pypi.python.org/pypi/lea
With the hope that Lea can make the Force less uncertain,