Metadata-Version: 2.3
Name: routing-pyramids
Version: 0.1.0a1
Summary: Unsupervised Learning of Cell Instances with Generative Routing Pyramids
Requires-Dist: numpy
Requires-Dist: torch
Requires-Dist: lightning
Requires-Dist: tensorboard>=2.20.0
Requires-Dist: tifffile[codecs]>=2026.1.14
Requires-Dist: scikit-image>=0.25.2
Requires-Dist: torchvision>=0.21.0
Requires-Dist: einops>=0.8.1
Requires-Dist: cmap>=0.7.2
Requires-Dist: zarr>=3.2.1
Requires-Dist: matplotlib>=3.10.8 ; extra == 'analysis'
Requires-Dist: pandas ; extra == 'analysis'
Requires-Dist: seaborn>=0.13.2 ; extra == 'analysis'
Requires-Dist: scikit-learn>=1.9.0 ; extra == 'analysis'
Requires-Dist: umap-learn>=0.5.12 ; extra == 'analysis'
Requires-Python: >=3.12
Provides-Extra: analysis
Description-Content-Type: text/markdown

# Generative Routing Pyramids

Official implementation of
_Unsupervised Learning of Cell Instances with Generative Routing Pyramids_
([arXiv](http://arxiv.org/abs/2608.16810)).

## Install

Install with pip:

```sh
pip install git+https://github.com/weigertlab/routing-pyramids.git
```

The visualization scripts require optional dependencies from the `[analysis]` extra.

Development installation requires [uv](https://docs.astral.sh/uv):

```sh
# in cloned repository
uv sync --extra analysis
```

## Examples

We provide example scripts for:

- [training](examples/train_fluohela.py)
- [prediction](examples/predict_psc.py)
- [visualizing pyramidal routing and feature maps](examples/visualize_flow_aics_object_posterior.py)
- [latent analysis and instance generation](examples/visualize_bbbc013_latents.py)

Also see the [dataset documentation](data/README.md).
