Regex for slicing categorical axes

[alexander-held]

StrCategory axes are very useful for grouping many datasets together in multi-dimensional histograms. The available indexing features in UHI seem mostly targeted at numerical axes. It is possible to slice categorical axes in the same way via integers after determining which index a given category corresponds to. Are there any methods in hist at the moment to extract them directly by matching strings? I imagine that something like a regex-based category matching could be useful.

[alexander-held]

This small example allows filtering for categories matching a regex:

import hist
import re

def cat_regex_slice(h, expression):
    return [hist.tag.loc(name) for name in h.axes["categories"] if re.search(expression, name)]

h = hist.Hist.new.StrCat(["aa", "ab", "bb"], name="categories").Double()

print(h.axes)
print(h[cat_regex_slice(h, r"^a")].axes)

output:

(StrCategory(['aa', 'ab', 'bb'], name='categories', label='categories'),)
(StrCategory(['aa', 'ab'], name='categories', label='categories'),)

The axis name needs to be specified in this implementation, but presumably it is possible somehow to infer which axis should be used in an automatic way?