colour-science / colour / 23070242166

100%

develop: 100%

push

github

Refactor *Munsell* module into a package and add *ONNX*-based conversions.

Restructure `colour.notation.munsell` from a single module into a package, splitting the existing code into `centore2014.py` and `value.py`.

Add `onnx.py` with pre-trained neural network models from the *colour-science/learning-munsell* *HuggingFace* repository for *Munsell* <-> *CIE xyY* conversions.

Add `colour.utilities.models` for downloading and caching remote files and `colour.utilities.requirements` for optional dependency gating.

Performance improvements are as follows:

- xyY -> Munsell Colour: ~742x
- Munsell Colour -> xyY: ~27x

686 of 688 new or added lines in 11 files covered. (99.71%)

45317 of 45414 relevant lines covered (99.79%)

1.0 hits per line