words2num2 0.3.0.dev1__tar.gz → 0.3.1__tar.gz

{words2num2-0.3.0.dev1 → words2num2-0.3.1}/PKG-INFO

@@ -1,17 +1,6 @@
 Metadata-Version: 2.4
 Name: words2num2
-Version: 0.3.0.dev1
-Summary: Inverse of num2words2: convert spoken-form numbers back to numeric values across 100+ languages.
-Home-page: https://github.com/jqueguiner/words2num2
-Author: Jean-Louis Queguiner
-Author-email: Jean-Louis Queguiner <jlqueguiner@gladia.io>
-Maintainer: Jean-Louis Queguiner
-Maintainer-email: Jean-Louis Queguiner <jlqueguiner@gladia.io>
-License: LGPL-2.1
-Project-URL: Homepage, https://github.com/jqueguiner/words2num2
-Project-URL: Repository, https://github.com/jqueguiner/words2num2.git
-Project-URL: Bug Tracker, https://github.com/jqueguiner/words2num2/issues
-Keywords: number,word,numbers,words,parse,parser,convert,conversion,i18n,localisation,localization,internationalisation,internationalization,asr,speech,llm
+Version: 0.3.1
 Classifier: Development Status :: 4 - Beta
 Classifier: Intended Audience :: Developers
 Classifier: Programming Language :: Python :: 3
@@ -20,21 +9,23 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Programming Language :: Python :: 3.14
-Classifier: Programming Language :: Python :: 3.15
+Classifier: Programming Language :: Rust
 Classifier: Topic :: Software Development :: Internationalization
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Classifier: Topic :: Software Development :: Localization
 Classifier: Topic :: Text Processing :: Linguistic
-Requires-Python: >=3.10
-Description-Content-Type: text/x-rst
 License-File: COPYING
-Requires-Dist: docopt>=0.6.2
-Requires-Dist: num2words2>=0.1.0.dev0
-Dynamic: author
-Dynamic: home-page
-Dynamic: license
-Dynamic: license-file
-Dynamic: maintainer
+Summary: Rust-powered inverse of num2words2: parse spoken-form numbers back to numeric values across 100+ languages, optimized for ASR/LLM/speech.
+Keywords: number,word,numbers,words,parse,parser,convert,conversion,i18n,localisation,localization,internationalisation,internationalization,asr,speech,llm,rust
+Author-email: Jean-Louis Queguiner <jean-louis.queguiner@gmail.com>
+Maintainer-email: Jean-Louis Queguiner <jean-louis.queguiner@gmail.com>
+License-Expression: LGPL-2.1-only
+Requires-Python: >=3.10
+Description-Content-Type: text/x-rst; charset=UTF-8
+Project-URL: Bug Tracker, https://github.com/jqueguiner/words2num2/issues
+Project-URL: Documentation, https://github.com/jqueguiner/words2num2/wiki
+Project-URL: Homepage, https://github.com/jqueguiner/words2num2
+Project-URL: Repository, https://github.com/jqueguiner/words2num2.git
 
 words2num2 — words to numbers
 =============================
@@ -104,9 +95,21 @@ entries**) and adds a free-text *auto-parse* mode that handles
 currencies, units, configurable thousands/decimal separators, and
 ASR/LLM-style mixed text.
 
-The project is hosted on GitHub_. Contributions are welcome.
+It is **Rust-powered with a thin Python binder**. The whole parsing engine —
+the English grammar, the generic reverse-lookup backend, the sentence walker,
+number-format parsing and auto-parse — runs in a compiled Rust core
+(PyO3/``abi3``); Python only shapes arguments and results. The core embeds the
+num2words2 conversion engine natively, so ``words2num2`` is **self-contained
+and has no runtime dependencies**, while running several times faster than the
+former pure-Python implementation. Output is unchanged — validated against a
+frozen corpus of ~11,000 round-trip cases.
+
+The project is hosted on GitHub_, and the full documentation is available
+in the Wiki_. Contributions are welcome.
 
 .. _GitHub: https://github.com/jqueguiner/words2num2
+.. _Wiki: https://github.com/jqueguiner/words2num2/wiki
+.. _GitHub Releases: https://github.com/jqueguiner/words2num2/releases
 
 Why this library
 ----------------
@@ -130,13 +133,59 @@ the forward direction. ``words2num2``:
 * Pluralizes long-form units in expand mode (``5 dollars`` /
   ``1 dollar``, ``5 feet`` / ``1 foot``, ``5 yen`` / ``1 yen``).
 
+Performance
+-----------
+
+The parsing engine is compiled Rust (PyO3/``abi3``), so there is no
+Python-level tokenising or table walking on the hot path. Typical native
+throughput (Apple M-series, nanoseconds per call, after warmup):
+
+.. list-table::
+   :header-rows: 1
+   :widths: 60 20
+
+   * - Operation
+     - ns/call
+   * - English cardinal (``"eight thousand seven hundred sixty-five"``)
+     - ~5,900
+   * - English ordinal (``"forty-second"``)
+     - ~4,100
+   * - French cardinal (``"trois cent quatre"``)
+     - ~1,900
+   * - Russian cardinal (``"сорок два"``)
+     - ~1,700
+   * - ``parse_number_string("1.234.567,89", lang="de")``
+     - ~240
+   * - ``auto_parse("$12,345.00")``
+     - ~990
+
+For the 100+ generic locales, the first call in a given language builds the
+``{words → number}`` reverse table once (~35 ms) and caches it in the core;
+every call after that is a native lookup. Output is byte-for-byte identical to
+the former pure-Python implementation, validated against a frozen corpus of
+~11,000 round-trip cases and the full test suite.
+
 Installation
 ------------
 
-**pip**::
+**pip** (Python 3.10+)::
 
     pip install words2num2
 
+That's it — the wheel bundles the Python binder and the compiled Rust
+extension, and ``words2num2`` has **no runtime dependencies** (the num2words2
+conversion engine is compiled in, so you do *not* need the num2words2 package).
+
+Prebuilt wheels are published for:
+
+* **Linux** — x86_64 and aarch64 (manylinux2014)
+* **macOS** — Apple Silicon (arm64); Intel (x86_64) from a later release
+* **Windows** — x86_64
+
+On any other platform (or an older release without a matching wheel), pip
+falls back to the source distribution and builds the extension locally — this
+needs a **stable Rust toolchain** (`rustup <https://rustup.rs>`_); nothing else.
+
 **Arch Linux / Manjaro (AUR):**
 
 .. code-block:: bash
@@ -150,14 +199,15 @@ Installation
     cd python-words2num2
     makepkg -si
 
-**From source**::
+**From source** (needs a stable Rust toolchain and `maturin
+<https://www.maturin.rs>`_)::
 
     git clone https://github.com/jqueguiner/words2num2
     cd words2num2
-    pip install -e .
+    pip install -e .          # builds the Rust extension via maturin
+    # or, to produce a wheel:  maturin build --release
 
-``num2words2`` is a runtime dependency for the generic multi-language
-backend and is installed automatically.
+Wheels for every version are also attached to the `GitHub Releases`_ page.
 
 Quickstart
 ----------
@@ -303,6 +353,13 @@ entries** including:
 
 Aliases: ``jp`` → ``ja``, ``cn`` → ``zh_CN``.
 
+Wiki
+----
+For the full documentation, including installation, API details, CLI
+usage, supported locales, sentence conversion, auto-parse behavior, and
+migration guidance, please check the Wiki_. Feel free to propose wiki
+enhancements.
+
 Conversion types
 ~~~~~~~~~~~~~~~~
 
@@ -387,3 +444,4 @@ License
 -------
 
 LGPL-2.1, mirroring ``num2words2``. See `COPYING`.
+

words2num2-0.3.0.dev1/words2num2.egg-info/PKG-INFO → words2num2-0.3.1/README.rst

@@ -1,41 +1,3 @@
-Metadata-Version: 2.4
-Name: words2num2
-Version: 0.3.0.dev1
-Summary: Inverse of num2words2: convert spoken-form numbers back to numeric values across 100+ languages.
-Home-page: https://github.com/jqueguiner/words2num2
-Author: Jean-Louis Queguiner
-Author-email: Jean-Louis Queguiner <jlqueguiner@gladia.io>
-Maintainer: Jean-Louis Queguiner
-Maintainer-email: Jean-Louis Queguiner <jlqueguiner@gladia.io>
-License: LGPL-2.1
-Project-URL: Homepage, https://github.com/jqueguiner/words2num2
-Project-URL: Repository, https://github.com/jqueguiner/words2num2.git
-Project-URL: Bug Tracker, https://github.com/jqueguiner/words2num2/issues
-Keywords: number,word,numbers,words,parse,parser,convert,conversion,i18n,localisation,localization,internationalisation,internationalization,asr,speech,llm
-Classifier: Development Status :: 4 - Beta
-Classifier: Intended Audience :: Developers
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Classifier: Programming Language :: Python :: 3.12
-Classifier: Programming Language :: Python :: 3.13
-Classifier: Programming Language :: Python :: 3.14
-Classifier: Programming Language :: Python :: 3.15
-Classifier: Topic :: Software Development :: Internationalization
-Classifier: Topic :: Software Development :: Libraries :: Python Modules
-Classifier: Topic :: Software Development :: Localization
-Classifier: Topic :: Text Processing :: Linguistic
-Requires-Python: >=3.10
-Description-Content-Type: text/x-rst
-License-File: COPYING
-Requires-Dist: docopt>=0.6.2
-Requires-Dist: num2words2>=0.1.0.dev0
-Dynamic: author
-Dynamic: home-page
-Dynamic: license
-Dynamic: license-file
-Dynamic: maintainer
-
 words2num2 — words to numbers
 =============================
 
@@ -104,9 +66,21 @@ entries**) and adds a free-text *auto-parse* mode that handles
 currencies, units, configurable thousands/decimal separators, and
 ASR/LLM-style mixed text.
 
-The project is hosted on GitHub_. Contributions are welcome.
+It is **Rust-powered with a thin Python binder**. The whole parsing engine —
+the English grammar, the generic reverse-lookup backend, the sentence walker,
+number-format parsing and auto-parse — runs in a compiled Rust core
+(PyO3/``abi3``); Python only shapes arguments and results. The core embeds the
+num2words2 conversion engine natively, so ``words2num2`` is **self-contained
+and has no runtime dependencies**, while running several times faster than the
+former pure-Python implementation. Output is unchanged — validated against a
+frozen corpus of ~11,000 round-trip cases.
+
+The project is hosted on GitHub_, and the full documentation is available
+in the Wiki_. Contributions are welcome.
 
 .. _GitHub: https://github.com/jqueguiner/words2num2
+.. _Wiki: https://github.com/jqueguiner/words2num2/wiki
+.. _GitHub Releases: https://github.com/jqueguiner/words2num2/releases
 
 Why this library
 ----------------
@@ -130,13 +104,59 @@ the forward direction. ``words2num2``:
 * Pluralizes long-form units in expand mode (``5 dollars`` /
   ``1 dollar``, ``5 feet`` / ``1 foot``, ``5 yen`` / ``1 yen``).
 
+Performance
+-----------
+
+The parsing engine is compiled Rust (PyO3/``abi3``), so there is no
+Python-level tokenising or table walking on the hot path. Typical native
+throughput (Apple M-series, nanoseconds per call, after warmup):
+
+.. list-table::
+   :header-rows: 1
+   :widths: 60 20
+
+   * - Operation
+     - ns/call
+   * - English cardinal (``"eight thousand seven hundred sixty-five"``)
+     - ~5,900
+   * - English ordinal (``"forty-second"``)
+     - ~4,100
+   * - French cardinal (``"trois cent quatre"``)
+     - ~1,900
+   * - Russian cardinal (``"сорок два"``)
+     - ~1,700
+   * - ``parse_number_string("1.234.567,89", lang="de")``
+     - ~240
+   * - ``auto_parse("$12,345.00")``
+     - ~990
+
+For the 100+ generic locales, the first call in a given language builds the
+``{words → number}`` reverse table once (~35 ms) and caches it in the core;
+every call after that is a native lookup. Output is byte-for-byte identical to
+the former pure-Python implementation, validated against a frozen corpus of
+~11,000 round-trip cases and the full test suite.
+
 Installation
 ------------
 
-**pip**::
+**pip** (Python 3.10+)::
 
     pip install words2num2
 
+That's it — the wheel bundles the Python binder and the compiled Rust
+extension, and ``words2num2`` has **no runtime dependencies** (the num2words2
+conversion engine is compiled in, so you do *not* need the num2words2 package).
+
+Prebuilt wheels are published for:
+
+* **Linux** — x86_64 and aarch64 (manylinux2014)
+* **macOS** — Apple Silicon (arm64); Intel (x86_64) from a later release
+* **Windows** — x86_64
+
+On any other platform (or an older release without a matching wheel), pip
+falls back to the source distribution and builds the extension locally — this
+needs a **stable Rust toolchain** (`rustup <https://rustup.rs>`_); nothing else.
+
 **Arch Linux / Manjaro (AUR):**
 
 .. code-block:: bash
@@ -150,14 +170,15 @@ Installation
     cd python-words2num2
     makepkg -si
 
-**From source**::
+**From source** (needs a stable Rust toolchain and `maturin
+<https://www.maturin.rs>`_)::
 
     git clone https://github.com/jqueguiner/words2num2
     cd words2num2
-    pip install -e .
+    pip install -e .          # builds the Rust extension via maturin
+    # or, to produce a wheel:  maturin build --release
 
-``num2words2`` is a runtime dependency for the generic multi-language
-backend and is installed automatically.
+Wheels for every version are also attached to the `GitHub Releases`_ page.
 
 Quickstart
 ----------
@@ -303,6 +324,13 @@ entries** including:
 
 Aliases: ``jp`` → ``ja``, ``cn`` → ``zh_CN``.
 
+Wiki
+----
+For the full documentation, including installation, API details, CLI
+usage, supported locales, sentence conversion, auto-parse behavior, and
+migration guidance, please check the Wiki_. Feel free to propose wiki
+enhancements.
+
 Conversion types
 ~~~~~~~~~~~~~~~~
 

words2num2-0.3.1/pyproject.toml

@@ -0,0 +1,94 @@
+[build-system]
+requires = ["maturin>=1.5,<2.0"]
+build-backend = "maturin"
+
+[project]
+name = "words2num2"
+# Static base version. The release CI overrides it from the pushed git tag
+# (see .github/workflows/release.yml); bump here for local source builds.
+version = "0.3.1"
+license = "LGPL-2.1-only"
+license-files = ["COPYING"]
+description = "Rust-powered inverse of num2words2: parse spoken-form numbers back to numeric values across 100+ languages, optimized for ASR/LLM/speech."
+readme = "README.rst"
+requires-python = ">=3.10"
+# Self-contained: the num2words2 conversion engine is embedded in the compiled
+# core (num2words2-core), so there is no runtime dependency on the num2words2
+# package and no docopt (the CLI uses only the standard library).
+dependencies = []
+authors = [
+    {name = "Jean-Louis Queguiner", email = "jean-louis.queguiner@gmail.com"}
+]
+maintainers = [
+    {name = "Jean-Louis Queguiner", email = "jean-louis.queguiner@gmail.com"}
+]
+keywords = ["number", "word", "numbers", "words", "parse", "parser", "convert",
+            "conversion", "i18n", "localisation", "localization",
+            "internationalisation", "internationalization", "asr", "speech",
+            "llm", "rust"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Developers",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
+    "Programming Language :: Rust",
+    "Topic :: Software Development :: Internationalization",
+    "Topic :: Software Development :: Libraries :: Python Modules",
+    "Topic :: Software Development :: Localization",
+    "Topic :: Text Processing :: Linguistic",
+]
+
+[project.urls]
+Homepage = "https://github.com/jqueguiner/words2num2"
+Documentation = "https://github.com/jqueguiner/words2num2/wiki"
+Repository = "https://github.com/jqueguiner/words2num2.git"
+"Bug Tracker" = "https://github.com/jqueguiner/words2num2/issues"
+
+[project.scripts]
+words2num2 = "words2num2.__main__:main"
+
+# ---- maturin: mixed Rust/Python project -------------------------------------
+# The compiled engine lives in ./rust (vendored crates `words2num2-py`, lib
+# `_rust`, plus its `num2words2-core` dependency); the thin Python binder is
+# ./words2num2. maturin bundles both into a single per-platform wheel. The
+# wheel embeds num2words2-core natively — no lingua-rs models, so it stays
+# small (~4MB), and there is only one variant.
+[tool.maturin]
+manifest-path = "rust/words2num2-py/Cargo.toml"
+python-source = "."
+module-name = "words2num2._rust"
+features = ["pyo3/extension-module"]
+strip = true
+
+[tool.isort]
+profile = "black"
+line_length = 88
+skip = ["words2num2/_version.py"]
+multi_line_output = 3
+include_trailing_comma = true
+force_grid_wrap = 0
+use_parentheses = true
+ensure_newline_before_comments = true
+
+[tool.black]
+line-length = 88
+target-version = ['py310']
+include = '\.pyi?$'
+exclude = '''
+/(
+    \.git
+  | \.hg
+  | \.mypy_cache
+  | \.tox
+  | \.venv
+  | _build
+  | buck-out
+  | build
+  | dist
+  | words2num2/_version.py
+)/
+'''