Installation

PyPlanAnalysis uses a standard src-layout package with a pyproject.toml build configuration.

There are two supported ways to install it, depending on whether your environment has direct access to PyPI.

Standard install (PyPI access available)

If your machine has normal internet/PyPI access

git clone https://github.com/oncoray/PyPlanAnalysis.git
cd PyPlanAnalysis
pip install -e .

To also install development dependencies (pytest):

pip install -e ".[dev]"

To also install documentation dependencies (sphinx, furo, etc.):

pip install -e ".[docs]"

Conda-based install (restricted/firewalled networks)

If your network blocks direct access to PyPI (common on institutional or clinical workstations), use Environment.yml to install all dependencies via conda first, then install the package itself

git clone https://github.com/oncoray/PyPlanAnalysis.git
cd PyPlanAnalysis
conda env create -f Environment.yml
conda activate PyPlan_env
pip install -e . --no-deps --no-build-isolation

--no-deps skips installing runtime dependencies via pip, since they are already provided by the conda environment. --no-build-isolation avoids pip fetching build tools (setuptools, wheel) into an isolated build environment

Requirements

• Python >= 3.10

• Core dependencies (installed automatically via pip or provided via Environment.yml): numpy, pandas, scipy, matplotlib, pydicom, scikit-image, openpyxl, SimpleITK.

• For the conda-based install: setuptools and wheel must also be listed in Environment.yml.

The NTCP parameter workbook (NTCPModels_params.xlsx) is bundled inside the package (PyPlanAnalysis/data/) and is located automatically at runtime — no separate download or manual path configuration needed.