This is python package for performing Boolean matrix factorisation.
Install as pip install bfact-core, which should create CLI command bfact.
You can then get available options by running bfact --help.
Note, you will need a valid Gurobi license for bfact to run, academic Gurobi licenses are free.
--in-file: path to .h5ad or a tab separated file containing binary matric
--out-stub: The path + prefix for ouputs.
Any other input, see bfact --help for available options.
{out_stub}_best_{metric}.L.txt
{out_stub}_best_{metric}.R.txt
{out_stub}_best_{metric}.summary_info.json
{out_stub}_{num_considered_factors}_cols.pkl (only if --to-save-cols specified true). This file stores the candidate factors used in the restricted master problem. It is useful for large matrices to save time if provided in future runs (using the base as --saved-cols-stub).