lehins / color Goto Github PK

Hello, I am trying to find out if this library can already do the computation to find perceived lightness as described in this post. If I have a (Word8, Word8, Word8) RGB triple, how can I determine perceived lightness using this library? I assume this is possible, but can someone help?

I noticed that the test specifications for e.g. LAB use the same Abritrary instance as others:

instance (Elevator e, Random e, Illuminant i) => Arbitrary (Color (LAB (i :: k)) e) where
  arbitrary = ColorLAB <$> arbitraryElevator <*> arbitraryElevator <*> arbitraryElevator

where arbitraryElevator gives either the full range of values (for discrete quanta like Word8) or the range [0,1] for Float and Double. However, LAB at least (and possibly others) uses a different range: L can go from 0 to 100, while a and b vary in a similar range, depending on L, and can also be negative. Thus, right now the spec tests are only checking colors very close to black. I'm not sure how this can be handled within QuickCheck.

@ekmett mentioned on reddit that he'd like some color distance computation added to Color:

One thing I'd really like to have would be DIN99o color distance computations

It seems like a pretty good idea, I wanna record it before I forget about it. There are sample implementations available in:

• Python: https://github.com/colour-science/colour/blob/master/colour/difference/din99.py
• and JS: https://github.com/Evercoder/d3-color-difference

so it shouldn't be too hard to port it into Haskell.

SVG defines a set of named colours, I believe in sRGB, that are also used in CSS. It would be nice to have these (maybe as Graphics.Color.Standard.SVG).

I'm not sure if there's already an easy way to do this, but it would be nice to have some of the same functionality that colour has. Something like:

• blend :: (...) => e -> Color cs e -> Color cs e -> Color cs e
• over :: (...) => Color (Alpha cs) e -> Color cs e -> Color cs e
• overAlpha :: (...) => Color (Alpha cs) e -> Color (Alpha cs) e -> Color (Alpha cs) e

As I understand it, in colour, blends are always done in linear space.

For the Lab color space.

This is a request for adding the CIELUV colorspace, as well as the cylindrical views of LAB and LUV, LCh(ab) and LCh(uv).

I'd do this myself and make a PR, but I see that the modules for (most of) these are there, commented out, in the cabal file, and didn't want to duplicate work or clumsily trample tricky problems you had with implementation.