jqueguiner / words2num2

59%

main: 59%

LAST BUILD ON BRANCH refactor/thin-python-binder

branch: refactor/thin-python-binder

CHANGE BRANCH

x

Reset

Sync Branches

• refactor/thin-python-binder
• feat/digits-mode-v2
• feat/digits-sequence-mode
• fix-scale-cardinals
• fix-scale-plurals-connectors
• fix-spoken-year-forms
• fix/release-macos-cross
• main
• refactor/pure-rust-core
• rust-port
• v0.1.1
• v0.2.0
• v0.2.1
• v0.2.2
• v0.2.3
• v0.3.0
• v0.3.1

Build # 29606390633

Build Type

Pull #3

github

Committed by jqueguiner

Commit Message

refactor: thin Python binder — delete the 120-file Python conversion layer

words2num2/ is now a thin binder over the Rust core, matching num2words2/. All
locale resolution, en-grammar parsing, reverse-table lookup and mode dispatch
live in words2num2-core; the Python package only binds _rust and surfaces the
exception/result types.

- words2num2-core: add top-level `words2num(text, lang, to)` (resolve_lang +
  CONVERTER_TYPES check + en-vs-reverse-table dispatch) so no dispatch remains
  in Python. Bump to 0.1.1.
- binder: expose `#[pyfunction] words2num`.
- words2num2/__init__.py: thin pass-throughs to _rust (words2num,
  words2num_sentence, auto_parse, auto_parse_sentence, parse_number_string,
  normalize, supported_langs) + Words2NumError / Quantity surface.
- Delete 122 files: all 118 lang_*.py, formats.py, utils.py, compat.py,
  converters/sentence.py. base.py keeps only Words2NumError; converters/auto.py
  keeps only Quantity + pluralize (the names _rust imports).
- tests: drop the 2 dispatch tests referencing the removed CONVERTER_CLASSES.

Verified: cargo test green, clippy clean, words2num2-core 0.1.1 published;
pytest 114 passed / 1 skipped (was 116/1 — the 2 removed are the deleted-layer
dispatch tests).

Pull Request Pull Request #3: refactor: thin Python binder (delete the 120-file Python layer)

Coverage Stats

14 of 16 new or added lines in 2 files covered. (87.5%)

44 of 74 relevant lines covered (59.46%)

3.54 hits per line