gtaylor
Update color_diff.py
Default CIEDE200 changed to use unity 1 in Kl, Kc, Kh without passing as args.
delta_e_cie2000_nonunity(color1, color2, Kl, Kc, Kh) proposed as new function to allow for passing in Kl, Kc, Kh in instances which a different weighting factor, or nonunity weighting factors are desired.
I'm wondering if we should retain the old kwarg layout, but emit a deprecation warning so that we're not breaking people without notice. We could completely remove these in a future release.
What do you think?
I think a future release is the way to go.
Mark Brown Sent with Sparrow (http://www.sparrowmailapp.com/?sig)
On Monday, March 21, 2016 at 8:57 PM, Greg Taylor wrote:
I'm wondering if we should retain the old kwarg layout, but emit a deprecation warning so that we're not breaking people without notice. We could completely remove these in a future release. What do you think?
— You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub (https://github.com/gtaylor/python-colormath/pull/61#issuecomment-199628169)
Are there comments to the suggested edits? Is delta_e_cie2000_nonunity() sufficient in the warning / a good enough name? For the relatively short overhead of duplicate functionality, though with application-specific values people use, should amendments be made? i.e. (Argv[x]=some value) with fabric:linen, cotton, suede, paper:LWC-B72, N-ST58, SC-B65 as examples
I'm admittedly not actively using the module at this point, so I've got to lean heavily on other people giving this a shot and weighing in. If you are spectating, please hop in and try this PR out!
The PR I've made is seemingly arbitrary, but in my attempt to port the delta e function to gpu, I've found that if this package can get away from treating LabColor as an object, some significant performance speedups could be realized.
More to the point, though slightly off-topic to this PR: it is unclear, and not easy to infer what happens when a lab color is created. Passing in l,a & b creates the color, but the function has to look up values for angle and luminescence. The headers indicate a return of ndarray, when array is used instead.