feat(anonymize): add NULL, NAME, PHONE_NUMBER, DATE_OF_BIRTH, TEXT, WORD types

[Andarius]

Summary

anonymize previously supported only EMAIL. This extends FieldType so config files can cover more PII columns without falling back to custom SQL post-steps:

• NULL — blank the column. Useful for hashes / tokens / anything that shouldn't carry information in a shared dump.
• FIRST_NAME, LAST_NAME, NAME — faker.first_name() / last_name() / name().
• PHONE_NUMBER — faker.phone_number().
• DATE_OF_BIRTH — faker.date_of_birth(), accepts minimum_age / maximum_age extra args.
• TEXT, WORD — faker.text() / faker.word().

All types forward their extra_args to the underlying Faker call, so the existing EMAIL domain: pattern works for the new ones too.

Heads-up

YAML's bare NULL parses as a real null, so users need to quote the type:

fields:
  - column: password_hash
    type: "NULL"          # quoted, otherwise parsed as None

Documented in the README example.

Tests

• 7 new parametrized cases over _get_fake_value (one per supported type) verifying the output matches the type's shape.
• Explicit pytest.raises(ValueError, match="unimplemented field type") to keep the fallback honest.
• All 8 unit tests pass locally (DB-dependent tests untouched).

Test plan

• CI green
• Manual smoke against a dev DB with a config exercising each new type