fair-segmentation

FAIR organelle segmentation pipeline

Based on empanada-napari napari plugin, export with FAIR RO-Crate packaging

Installation

Requires Python 3.12 or later. All dependencies are defined in pyproject.toml; the napari extra adds napari, empanada-napari and the napari-meta-tiff reader.

empanada-napari is currently installed from the ccp-volume-em fork, which includes the open pull request volume-em/empanada-napari#56, rather than a release, see empanada-napari notes.

Conda

conda env create -f environment.yml
conda activate fair-segmentation-env

This installs the package in editable mode with the napari and dev extras. Conda only provides Python; everything else is installed with pip. On Windows, conda's matplotlib pulls in an ICU library that breaks the PyQt6 wheels napari uses, so avoid installing napari-related packages with conda into this environment.

pip

pip install -e ".[napari]"

pixi

pixi run -e napari python run_fair.py

Docker

docker build --build-arg IMAGE_TAG=<tag> -t ghcr.io/nl-bioimaging/empanada-napari-xpra:<tag> .
docker run -p 9876:9876 ghcr.io/nl-bioimaging/empanada-napari-xpra:<tag>

The tag given as IMAGE_TAG is recorded as the container of every packaged run.

Then open napari in the browser at http://localhost:9876.

Usage

python run_fair.py

When an output .ome.zarr is opened in napari afterwards (with napari-ome-zarr), the label layers are added hidden, as napari-ome-zarr does for all label images; switch them on with the eye icon in the layer list.