Skip to content

libertininick/bayesian-stats

Repository files navigation

bayesian-stats

This is a sandbox repository for documenting my journey learning probabilistic programming & Bayesian statistics.

I've chosen to build this library on top of PyTorch and Pyro given my experience with torch from my deep learning adventures.

Installation

To create a conda environment and install bayesian-stats with poetry:

# Clone repository
git clone [email protected]:libertininick/bayesian-stats.git

# Navigate to local repo directory
cd bayesian-stats

# Update base conda environment
conda update -y -n base -c defaults conda

# Install conda-lock
# conda-lock is used to generate fully reproducible conda environments via a lock file
conda install --channel=conda-forge --name=base conda-lock

# Create conda environment from `conda-lock.yml`
conda-lock install --name bayesian_stats conda-lock.yml

# Activate conda environment
conda activate bayesian_stats

# Install `bayesian-stats` w/ poetry
poetry install --with dev,jupyter

Running Linting and Type Checking

To format code to adhere to our style and run type checking run the following:

ruff check . --fix
mypy src/

Running Tests

To run tests and test coverage, run the following:

coverage erase \
    && coverage run -m pytest \
    && coverage report

To skip slow tests

coverage run -m pytest -m "not slow"

Authors

About

Sandbox repo for Bayesian statistics and modeling

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published