Pioupiou 🐤

Introduction

Pioupiou is a nano probabilistic programming language, embedded into Python.

Use it to define probabilistic models :

import pioupiou as pp
a, b = 0.5, 1.0
X = pp.Uniform(0.0, 1.0)
E = pp.Normal(0.0, 0.01)
Y = a * X + b + E

and to simulate them

n = 1000 # number of samples
omega = pp.Omega(n)
x, y = X(omega), Y(omega)

The results are

>>> x # doctest: +ELLIPSIS
array([6.36961687e-01, 2.69786714e-01, 4.09735239e-02, ..., 3.80007897e-01])
>>> y # doctest: +ELLIPSIS
array([1.09588258, 1.22942954, 1.01954509, 0.99213115, ..., 1.14366864])

That's about it! Use this data as you see fit. For example:

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
data = pd.DataFrame({"x":x, "y": y})
_ = sns.jointplot(x="x", y="y", data=data, kind="reg", xlim=(0.0, 1.0), ylim=(0.75, 1.75))
plt.savefig("xy.svg")

Getting started

Install the latest version of pioupiou with:

$ pip install --upgrade git+https://github.com/boisgera/pioupiou.git

---

Last update: 2022-05-25