kpu / kenlm

KenLM: Faster and Smaller Language Model Queries
http://kheafield.com/code/kenlm/
Other
2.46k stars 514 forks source link

Manually set include paths on macOS to find LZMA #414

Closed jacobkahn closed 1 year ago

jacobkahn commented 1 year ago

For whatever reason, when KenLM is built from pip (e.g. pip install git+https://github.com/kpu/kenlm.git), /usr/local/include isn't getting properly propagated on some macOS systems, and the resulting build can't find lzma.h. The test compilation with that header works when specifically passed with -include lzma.h, but fails when built by the setuptools Extension.

Failure example here: https://app.circleci.com/pipelines/github/flashlight/text/283/workflows/f4d5ff8f-2201-401b-a8d7-bf207fd7e64f/jobs/2492