fix late binding and proper reversal for funcs

[cvanelteren]

A secondary issue came up in #294 where a callable is not properly inverse.

This PR fixes a bug in the reversal of colormap segment data where callable segments were incorrectly captured due to Python’s late binding behavior in closures. Specifically, lambdas inside the dictionary comprehension referenced the loop variable data, which caused all callables to use the last data value. This is resolved by binding data as a default argument in the lambda. Explicit segment lists are unaffected and still reversed correctly.

[cvanelteren]

import ultraplot as uplt

fig, ax = uplt.subplots()

cmaps = ["jet", "jet_r", "rainbow", "rainbow_r"]
for i, loc in enumerate(["r", "l", "t", "b"]):
    ax.colorbar(uplt.Colormap(cmaps[i]), loc=loc, label=cmaps[i])
uplt.show(block=1)

[codecov]

Codecov Report

❌ Patch coverage is 85.71429% with 2 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
ultraplot/colors.py	75.00%	2 Missing ⚠️

📢 Thoughts on this report? Let us know!

[beckermr]

This method removes any of the characters in _r, not _r if the string ends with _r. I think you may want something like original_key.rsplit("_r", maxsplit=1)[0]

[cvanelteren]

You are correct, but then we should replace all strip functions by a split you suggest. My edit here merely replaces it with a more appropriate function. I think the original strip is wrong as it also deletes leading "_r".

[beckermr]

Right we should do that fix. We'll need another PR

[cvanelteren]

I can also include it in here, although I think this issue will not be a major things as the _r is added by ultraplot. If the colormap ends with _r it will be listed as {cmapnameendingwith_r}_r which should then split fine I think.

[cvanelteren]

We can go ahead and merge this I guess @beckermr?

[beckermr]

Didn't you want to fix the string _r thing here too?

[cvanelteren]

Ah ok that wasn't entirely clear. Will do that now

[cvanelteren]

done ;-)!