PR: Implement support for *HyAB* large color difference.

[mesvam]

Summary

Implement HyAB color difference and resolves #610

Preflight

Code Style and Quality

• Unit tests have been implemented and passed.
• Pyright static checking has been run and passed.
• Pre-commit hooks have been run and passed.
• New transformations have been added to the Automatic Colour Conversion Graph.
• New transformations have been exported to the relevant namespaces, e.g. colour, colour.models.

Documentation

• New features are documented along with examples if relevant.
• The documentation is Sphinx and numpydoc compliant.

[mesvam]

I think I got the implementation correct, but could somebody double check the other stuff: documentation, tests, citation, formatting, etc.? I'm not sure I got all the tests to run correctly.

[KelSolaar]

Thank you @mesvam,

I added a few important comments, what about also adding cbLAB? Would be great!

Cheers,

Thomas

[mesvam]

I just noticed there are functions for manhattan_distance and euclidean_distance in colour.algebra. Would you prefer I use those functions to implement these difference metrics instead?

[mesvam]

what about also adding cbLAB? Would be great!

I don't mind adding cbLAB if you like, as well as some of the other metrics in that paper, but the authors recommended HyAB and HyCH as the top candidates. We can also add all four formulas, but maybe that's too much or too confusing.

Also, would you prefer that in a separate PR, or just tack it onto this one?

[mesvam]

Ok, I'm stumped. How do I pass Codacy? I can't seem to find the error messages

[KelSolaar]

I just noticed there are functions for manhattan_distance and euclidean_distance in colour.algebra. Would you prefer I use those functions to implement these difference metrics instead?

If you can wherever possible, this would be fantastic!

I would probably try to get HyCH actually if you please don't mind:

The HyAB has the best consistency with CIELAB for
all categories of the second dataset. The HyCH has a good
consistency with CIEDE2000 for the second dataset as
well. So, the HyAB and HyCH could be used for small to
large color difference calculation. However, the HyAB
has a slightly better performance and the advantage of
simpler computation.
It is thus concluded that, for very large color differ-
ences, the effect of lightness difference should be consid-
ered perceptually separable, and the HyAB color
difference formula is the most practical and accurate
solution.

Ok, I'm stumped. How do I pass Codacy? I can't seem to find the error messages

Do not worry, it is all good atm!

[mesvam]

I submitted a separate PR for HyCH #1307

[mesvam]

On second thought, I think using manhattan_distance and euclidean_distance would just slow it down and make it more complicated. I'm pretty happy with this as is right now.

[KelSolaar]

Thank you! Merging this one!