Installation¶

Requirements¶

Calibre requires Python 3.12 or later.

Installing from PyPI¶

pip install calibre

This installs Calibre along with its runtime dependencies:

numpy >= 1.20.0
scipy >= 1.7.0
scikit-learn >= 1.0.0
cvxpy >= 1.2.0

That is the whole list. pandas and matplotlib were runtime dependencies before 0.7.0 and are not any more — nothing in the package imported them.

Plotting¶

Plotting needs matplotlib, which is an optional extra:

pip install 'calibre[plots]'

Everything else in calibre works without it. Importing calibre never imports matplotlib, so the core package stays light for anyone who only needs the numbers.

Installing from Source¶

git clone https://github.com/finite-sample/calibre.git
cd calibre
pip install -e .

Development Installation¶

git clone https://github.com/finite-sample/calibre.git
cd calibre
uv sync --all-extras --dev
uv run pytest

Important

Development dependencies are a PEP 735 [dependency-groups] entry, so pip install -e ".[dev]" does not work. Use uv sync --all-extras --dev.

Tooling is ruff (formatting and linting) plus pytest, pytest-cov and pyright. black, isort and flake8 were replaced by ruff in 0.4.1.

Documentation Dependencies¶

To build the documentation locally:

uv sync --all-extras --dev
make docs

The docs group pulls sphinx, furo, myst-parser, sphinx-copybutton, sphinx-autodoc-typehints and nbsphinx, plus matplotlib and pandas, which the example notebooks use.

Verifying Installation¶

import calibre

print(f"Calibre version: {calibre.__version__}")

Troubleshooting¶

ImportError: No module named ‘cvxpy’

CVXPY is a hard runtime dependency because nearly_isotonic.py imports it at module level. If it failed to install, try it on its own:

pip install cvxpy

On macOS, conda is often easier:

conda install -c conda-forge cvxpy
pip install calibre

Memory errors during installation

pip install --no-cache-dir calibre

Getting Help¶

If you hit an installation problem, check the GitHub Issues for similar reports, then open a new one including the output of pip --version and python --version.