Fix fractional(0) returning "0/1" instead of "0"

[Sreekant13]

fractional(0) returns "0/1" instead of "0", even though every other whole number renders bare:

>>> import humanize
>>> humanize.fractional(0)
'0/1'   # expected '0'
>>> humanize.fractional(3)
'3'

The whole-number return branch was gated on whole_number being truthy, so 0 (where whole_number == 0) skipped it and fell through to the fraction branch, producing "0/1". Since numerator == 0 and denominator == 1 already uniquely identifies a whole number (including 0), the whole_number and check is both redundant and wrong for 0. Removing it makes fractional(0) return "0", consistent with fractional(3), fractional(2.0), etc.

Fixes # (no linked issue - small self-contained bug fix)

Changes proposed in this pull request:

• Drop the redundant whole_number and guard in fractional() so 0 is treated as a whole number and returns "0" (previously "0/1").
• Add regression test cases (0, "0") and (0.0, "0") to test_fractional.

[Sreekant13]

Hi @hugovk, gentle nudge on this small fix whenever you have time. Thanks for maintaining humanize.

[Sreekant13]

Flagging an overlap, since I only noticed it now: @semx opened #354 a few days after this one, and it covers the same fractional(0) case plus a second one this PR does not touch, where the fractional part rounds up to a whole number and produces output like "2 1/1" for fractional(2.9999999).

Their fix folds any whole-valued fractional part into the integer part, which handles both cases in one place, so it is the more complete change of the two. I am happy to close this in favour of #354 if that is the direction you prefer, or to add their case here if you would rather keep this one. Whichever is less work for you.

[semx]

Thanks for the generous framing @Sreekant13 - for the record, your PR was here first, and priority counts for something too. Either direction works for me equally: if @hugovk prefers this one as first-come, I am happy to rebase #354 down to just the whole-rounding case on top of yours; if #354 goes in as the single change, credit for spotting fractional(0) first belongs here either way. Whatever is least work for the maintainer.

[Sreekant13]

Good outcome. Thanks @hugovk for picking one, and thanks @semx for the generous note. #354 was the more complete fix, and it covers the case this PR found, so landing it is the right result.