[emrehangorgec]

Activating tr_TR translates words but leaves English numeric separators. intcomma(1_234_567.89) returns 1,234,567.89, and the Turkish string "1234567,89" is read as 123,456,789.

Changes proposed in this pull request:

• Register tr_TR with a dot for grouping and a comma for decimals, following Unicode CLDR.
• Extend the existing intcomma and intword locale tests, and add a README example.

Found through code inspection and reproduced locally. Three regression checks fail before the fix. No matching issue or PR was found.

Validation on Windows with Python 3.14:

• python -m pytest tests src/humanize --doctest-modules --benchmark-disable --cov=humanize: 873 passed, 99% coverage.
• python -m mypy .: passed for all 14 source files.
• prek run --all-files --show-diff-on-failure: passed.
• The README example passed all four doctest checks. git diff --check passed.

Translation catalogs were compiled before testing. Local hook setup required reinstalling validate-pyproject==0.26 to restore missing schema files. Benchmark timings and other platforms were not tested.

Does this handle Turkish numeric inputs as intended, including strings passed to intcomma?