clbarnes / nblast-py

Pure-python implementation of NBLAST, for experimentation purposes

nblast-py

Pure(ish)-python implementation of NBLAST (as published here and implemented here by the Jefferis lab), for experimentation purposes. Written for python 3.8.

For a faster, more feature-filled NBLAST library with python bindings, see nblast-rs.

Test data

Generated (in R) with

library(nat.nblast)
library(nat)
kcscores <- nblast_allbyall(kcs20)
write.csv(kcscores, file='data/kcscores.csv')

This depends on point clouds with tangents pre-calculated by the nat library, using a k=5. These point clouds/ tangent vectors/ alpha values are included in data/dotprops.

The score matrix data/smat_fcwb.csv is also pre-calculated, extracted from the nat.nblast github repository.