Add Malayalam characters to char classes so periods tokenize correctly (#12898)

[vineethsaivs]

Description

Fixes #12898.

Malayalam sentence-terminal periods were not split off the final word:

import spacy
nlp = spacy.blank("ml")
[t.text for t in nlp("ഭാഷയാണ് മലയാളം.")]
# before: ['ഭാഷയാണ്', 'മലയാളം.']
# after:  ['ഭാഷയാണ്', 'മലയാളം', '.']

Malayalam letters were not part of the _uncased character group in spacy/lang/char_classes.py, so the tokenizer did not treat them as alphabetic characters and the suffix rules never separated the trailing period.

Following @svlandeg's suggestion on the issue, this defines the Malayalam Unicode block (U+0D00-U+0D7F) and adds it to _uncased, mirroring the existing Indic scripts (Tamil, Telugu, Kannada). A tokenizer regression test is added under spacy/tests/lang/ml/. Tamil and the existing Malayalam tests are unaffected.

Types of change

Bug fix (Malayalam tokenization).

Checklist

• I ran the tests, and all new and existing tests passed (spacy/tests/lang/ml/).
• Contributor agreement: I have not added .github/contributors/ because the spaCy Contributor Agreement is a copyright-assignment that must be signed by me, the human contributor; I will add .github/contributors/vineethsaivs.md before merge.

---

Disclosure: prepared with AI assistance; reviewed by me and I can explain every line.