Installation¶
Requirements¶
- Python 3.10, 3.11, or 3.12 (CI-tested)
- pip 23+
From PyPI¶
pip install probviz
probviz app
From source (recommended for the web app)¶
git clone https://github.com/sanskarpan/probviz.git
cd probviz
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
streamlit run web/app.py
Editable install (Python API + CLI)¶
pip install -e ".[dev]"
probviz app
Docker¶
docker build -t probviz .
docker run -p 8501:8501 probviz
# or
docker compose up --build
The image runs as non-root appuser, exposes 8501, and includes a
/_stcore/health healthcheck.
Dependency groups¶
| File | Purpose |
|---|---|
requirements.txt |
runtime (numpy, scipy, pandas, matplotlib, seaborn, plotly, streamlit, scikit-learn, statsmodels, joblib) |
requirements-dev.txt |
lint/type/test tooling |
requirements-docs.txt |
MkDocs site |
Canonical metadata lives in pyproject.toml; requirements*.txt mirror it for
pip/Docker workflows. setup.py is a thin shim only.