Probability is a concept that many introductory Computer Science courses teach because of its frequent application in algorithms, data structures, and other mathematical fields. While numerous libraries for expressing probabilities exist (e.g. scipy, statistics, etc), the majority of them focus primarily on the application of these concepts rather than showcasing the mechanics of the mathematical theory.

The goal of this project is to leverage Python's built-in language features to expose an intuitive and expandable API for simple probabilitic expressions.

pip install probs

from probs import *

# define a normally-distributed random variable with
# mean = 0, variance = 1
X = Normal()

assert E(X) == 0
assert Var(X) == 1 / 12
assert X.pdf(0.5) == 2

# combine multiple random variables
u, v = Uniform(), Uniform()

assert E(u) == 1 / 2
assert Var(u) == 1 / 12

assert (u * v).pdf(0.5) == 39.0169
assert (1 * v).pdf(0.5) == 1

assert E(u + 1) == 1.5
assert E(u + v) == 1
assert E(u - v) == 0

assert Var(u + v) == 1 / 6

All issues and pull requests are much appreciated! To build the project:

• probs is actively developed using the lastest version of Python.
  • First, be sure to run pre-commit install.
  • To run all tests and use auto-formatting tools, use pre-commit run.
  • To only run unit tests, run pytest.

• Use ApproxFloat across all operations.

• Dataclasses are iffy, because:

  • Need to set super().init() in order to get the parent class's fields.

  • Need to set eq=False on all RandomVariables.

  • Need to set repr=False in order to get the parent's repr method.

  • However, clearer init function provided, other operators potentially builtin.

  • repr can be inherited without a rewrite

  • inheritance works so long as every parent is also a dataclass.

  • super short init syntax

• Figure out how to merge pmf and custom pmf functions.

• Convert all floats to Decimal? For precision.