Installation¶
Requirements¶
- Python 3.11
- FEniCS/dolfinx 0.10 (for the PDE solver and mesh-view pipeline)
- MPI (for dolfinx parallelism)
Option 1 — Conda (recommended)¶
The complete environment including FEniCS, PETSc, and MPI can be installed via Conda. This is the most reliable route because FEniCS requires native compiled libraries that are difficult to install with pip alone.
Verify the installation:
Option 2 — uv (Python-only dependencies)¶
If you do not need the PDE/FEniCS features, you can install the pure-Python stack with uv:
FEniCS not available via pip/uv
The tumorfits pde, tumorfits heatmap, and tumorfits mesh-view
subcommands require dolfinx, which must be installed via Conda or a
pre-built Docker image. Use the Conda route for the full pipeline.
Option 3 — Docker¶
A pre-built Docker image is available for reproducible execution without a local Conda/FEniCS installation:
See the Dockerfile for details.
Development install¶
conda env create -f environment.yml
conda activate cna-resist-dynamics
pip install -e ".[dev,docs]"
pre-commit install