GitLab

allow a custom regexp

Add recipe for building the package with conda

Use a set for the (lowercase) stopwords, instead of looking up every word via linear search

lower-case stopwords before using them.

fix handling of scale parameter & python 3 compatibility for cli

Add a command line interface

Update font API & merge upper and lower case words when stemming

Has the advantage of producing fewer checks, as most words will be multiples.

downgrade numpy req 

Frequency scaling

fix sorting and normalization: wasn't applied if frequencies were given directly

hopefully fixing cython built