Fixed tritanopia simulation

[williamjsdavis]

Summary: Fixed a typo causing incorrect simulation of tritanopic color vision deficiencies. Also added tests for this corrected behavior.

This is my first pull request, I apologize if I have done anything incorrectly. I have been working with simulating various color vision deficiencies, and I noticed that the simulation for tritanopia was not producing results that I expected. Looking into the code in the function tritanopic had a typo:

Colors.jl/src/algorithms.jl

Line 138 in b550b7b

(one(p) - p) * q.l + p * (-(a*q.l + b*q.m)/c))

It should be:

(one(p) - p) * q.s + p * (-(a*q.l + b*q.m)/c))

This error can be easily observed when inspecting the behavior of the function at p=0 (no photopigment loss):

This can also be seen in the RGB values:

rgb = RGB(0.44, 0.26, 0.8)
rgb |> dump
tritanopic(rgb, 0) |> dump

RGB{Float64}
  r: Float64 0.44
  g: Float64 0.26
  b: Float64 0.8
RGB{Float64}
  r: Float64 0.4547022003851215
  g: Float64 0.2596348656158941
  b: Float64 0.33903165107645483

As a counterexample, the function protanopic is working correctly:

rgb = RGB(0.44, 0.26, 0.8)
rgb |> dump
protanopic(rgb, 0) |> dump

RGB{Float64}
  r: Float64 0.44
  g: Float64 0.26
  b: Float64 0.8
RGB{Float64}
  r: Float64 0.4400000030003578
  g: Float64 0.26000003129854465
  b: Float64 0.8000000074911268

This pull request fixes this typo, and adds three tests to validate this correct behavior. Please let me know if there are any questions.

[codecov]

Codecov Report

Base: 94.33% // Head: 95.51% // Increases project coverage by +1.18% 🎉

Coverage data is based on head (d117720) compared to base (b550b7b).
Patch has no changes to coverable lines.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #525      +/-   ##
==========================================
+ Coverage   94.33%   95.51%   +1.18%     
==========================================
  Files           9        9              
  Lines        1270     1270              
==========================================
+ Hits         1198     1213      +15     
+ Misses         72       57      -15     

Impacted Files	Coverage Δ
src/algorithms.jl	91.96% <ø> (+16.07%)	⬆️
src/utilities.jl	95.60% <0.00%> (-0.74%)	⬇️
src/conversions.jl	99.13% <0.00%> (-0.29%)	⬇️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[kimikage]

It is ok, but why .≈?

Algorithms: Test Failed at ~\.julia\dev\Colors\test\algorithms.jl:61
  Expression: protanopic(rgb, 0) .≈ rgb

Stacktrace:

More informative

Algorithms: Test Failed at ~\.julia\dev\Colors\test\algorithms.jl:61
  Expression: protanopic(rgb, 0) ≈ rgb
   Evaluated: RGB{Float64}(0.4400000030003578,0.26000003129854465,0.8000000074911268) ≈ RGB{Float64}(0.44,0.26,0.8)

Stacktrace:

[williamjsdavis]

You're right in highlighting that the use of .≈ is not ideal. Instead, ≈ should be used with an appropriate tolerance accounting for the small floating point errors introduced in the calculation.