It is our pleasure to announce the somewhat official alpha release of GeNePy3D, a python library for quantitative geometry. Its aim is to ease the construction of workflows analyzing 3D geometrical objects in python, a.k.a what do I do now that I a deepnet gave 1000 axons traces/segmented cell nuclei/? out of my image(s). In particular, it tries to link out to various libraries as needed to have everything in one place, to allow non-expert access to more specialized algorithms, including computational geometry (through CGAL and scipy), trees or mesh manipulation (etc…, see figure below), with hopefully more coming up.
The main page with the documentation is here, with the code living here, spreads into two repos for licensing reasons. It is easily installable via pip; we provide docker containers to ease use and deployment. In particular, some simple tutorials are available on the docs and more thorough example (to be expanded) are here.
There is a preprint describing the library with a first example reanalyzing a previousely published whole brain neuron traces dataset here.
The project is being developed by Minh-Son Phan (@msphan) and myself from the Laboratory of Optics for Biosciences at Ecole polytechnique near Paris. It’s pretty early in the project, which so far focuses on things we needed but we think it might be useful already. Please do not hesitate to try it, ask/comment etc… If you think it could be useful for you but some features are missing, ask here or on the github page and we’ll see what we can do, or better yet, think of contributing
Anatole and Son