phat

PHAT is the persistent-homology matrix-reduction engine — the algorithmic core several higher-level tools sit on top of, exposed to Python through pybind11. It is on no conda channel, so phat-environment stages this build and grades L1 for it.

LGPL-3.0 is redistributable, so nothing but the pull request stands between this and a channel. That makes it one of the cheaper entries on the upstreaming backlog: the build is a straightforward pybind11 compile with no vendored submodules and no network requirement, unlike giotto-ph-recipe and pyflagser-recipe.

Verified green under rattler-build on linux-64.