Cell Point Cloud Preprocessing (CPCP)
A python library for cell point cloud data preprocessing. The library currently includes utilities that supports:
- cross-time ST data alignment and evaluation using optimal transport
- cross-time training plan construction for machine learning models
- June 4th, 2025: mini-batch solver for entropic semi-relaxed FGW based on POT
- June 16th, 2025: full-batch solver for entropic semi-relaxed FGW based on POT
- June 16th, 2025: add transport plan evaluation functions: Locality preservation error, Spateo-CLC error, Label consistency, and number of points transported
- June 18th, 2025: implement algorithms for finding starting cells and ending cells for dataset curation