[pengzhendong]

Summary

• Add native Marathi (mr) ITN and a new TN, completing the direction missing from NeMo. Both use raw-field taggers and semantic verbalizers, with independent language graphs and no NeMo dependency.
• Register both directions in the CLI, packaging, cache-manifest tests and README.
• Support Devanagari, ASCII and mixed digits, lakh/crore quantities, dates, decimals and Marathi time idioms. Preserve explicit punctuation spacing through the shared joiner.

Coverage and limitations

• All 131 original ITN fixture lines are retained verbatim.
• New TN is checked against reversed ITN fixtures and hand-written cases. Seven reverse cases remain excluded with explicit reasons: conflicting hundreds-count/colloquial time readings, singular minute wording and a trailing-space variant.
• Leading-zero or oversized digit strings are read digit by digit. No constructor options are added.

Test plan

• Original, reversed, raw-field and mapping tests
• CLI and cache registration regressions with NeMo imports blocked
• Full project suite on the merged master baseline (19824 passed, NeMo imports blocked)
• flake8 tn itn and git diff --check