[in_mha_banned] name framework migration

in_mha_banned extracts aliases and acronyms from combined name strings with hand-rolled regex and keyword sets instead of the name-cleaning framework (h.review_names / h.apply_reviewed_name_string).

Crawler: datasets/in/mha_banned/crawler.py

Ad-hoc logic to retire (~45 lines):

• REGEX_ACRONYM_PARENS pulling acronym-in-parens into alias — crawler.py:15
• COMPLEX_TERMS keyword set gating a lookup — crawler.py:18-30
• splitting names_string on ;@ for aliases + parse_names — crawler.py:45-58, 178-184

Migration (sanctions crawler — 3 steps):

1. Mirror the existing cleaned names into a suggested h.Names() and call h.review_names(..., default_accepted=True). Output unchanged while reviews accumulate; deploy step 3 ASAP afterwards.
2. Complete the accumulated name reviews for this dataset.
3. Replace the custom logic with h.apply_reviewed_names / h.apply_reviewed_name_string and delete it. Add any custom alias-marker phrases not covered by rigour.names.name_split_phrases_list() as reject_strings under names.schema_rules in in_mha_banned.yml.

Migration guide: zavod/docs/extract/names.md (section "Migrating to the name cleaning helpers").

Read zavod/docs/extract/names.md, especially the migration section, then inspect datasets/in/mha_banned/crawler.py and datasets/in/mha_banned/in_mha_banned.yml. Trace the existing name handling and review workflow before migrating the dataset and completing its accumulated reviews. Done means the custom extraction logic is removed, schema reject strings cover any missing phrases, and output remains unchanged.