tex2word 0.8.1__tar.gz

tex2word-0.8.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Yifan Yang
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

tex2word-0.8.1/PKG-INFO

@@ -0,0 +1,209 @@
+Metadata-Version: 2.4
+Name: tex2word
+Version: 0.8.1
+Summary: Open-source, production-grade LaTeX -> Microsoft Word (.docx) converter with native OMML math and live fields
+Keywords: latex,word,docx,converter,ooxml,omml,tex,office,document
+Author: Yifan Yang
+Author-email: Yifan Yang <yfyang.86@hotmail.com>
+License-Expression: MIT
+License-File: LICENSE
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Science/Research
+Classifier: Intended Audience :: End Users/Desktop
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Text Processing :: Markup :: LaTeX
+Classifier: Topic :: Text Processing :: Markup :: XML
+Classifier: Topic :: Office/Business :: Office Suites
+Classifier: Typing :: Typed
+Requires-Dist: lxml>=5.0
+Requires-Dist: pylatexenc>=2.10
+Requires-Dist: citeproc-py>=0.6 ; extra == 'csl'
+Requires-Dist: matplotlib>=3.7 ; extra == 'mathimg'
+Requires-Dist: latex2mathml>=3.77 ; extra == 'mathml'
+Requires-Dist: pypdfium2>=4 ; extra == 'pdf'
+Requires-Dist: pillow>=10 ; extra == 'pdf'
+Maintainer: Yifan Yang
+Maintainer-email: Yifan Yang <yfyang.86@hotmail.com>
+Requires-Python: >=3.12
+Project-URL: Changelog, https://github.com/yfyang86/tex2word/blob/main/CHANGELOG.md
+Project-URL: Homepage, https://github.com/yfyang86/tex2word
+Project-URL: Issues, https://github.com/yfyang86/tex2word/issues
+Project-URL: Repository, https://github.com/yfyang86/tex2word
+Provides-Extra: csl
+Provides-Extra: mathimg
+Provides-Extra: mathml
+Provides-Extra: pdf
+Description-Content-Type: text/markdown
+
+# latex2word
+
+An open-source, cross-platform **LaTeX → Microsoft Word (`.docx`)** converter
+that produces *genuinely editable* Word: native paragraph styles, **native OMML
+equations** (editable in Word's equation editor, not images), and **live,
+auto-renumbering fields** for equation/figure/table numbers and
+cross-references.
+
+> **Status: production-grade.** Foundation, math core (direct LaTeX→OMML), the
+> live cross-reference/field plumbing (the differentiator), image embedding, the
+> BibTeX bibliography, and the robustness layer (math cascade, coverage report,
+> OOXML validator, round-trip manifest) are all in. See
+> [`CHANGELOG.md`](CHANGELOG.md) for the release history.
+
+## Why
+
+Pandoc/`texmath` is the open-source reference but **drops equation numbers**,
+can dump raw LaTeX for labelled equations, and emits *static* cross-references.
+No open tool produces editable styles **and** native OMML **and** live
+field-based numbering. That gap is the product.
+
+## Install & use
+
+Requires Python 3.12+.
+
+From PyPI:
+
+```bash
+pip install tex2word                 # core (PNG/JPEG figures)
+pip install "tex2word[pdf]"          # + PDF figure rasterisation (pypdfium2, Apache-2.0)
+pip install "tex2word[mathml]"       # + LaTeX->MathML->OMML for hard math (latex2mathml)
+pip install "tex2word[csl]"          # + real CSL citation styles (citeproc-py)
+pip install "tex2word[pdf,mathml,csl,mathimg]"   # everything
+
+latex2word convert paper.tex -o paper.docx
+latex2word convert paper.tex -o paper.docx --report report.json
+latex2word convert paper.tex -o paper.docx --reference-doc journal.docx
+```
+
+Or, for a development checkout with [uv](https://docs.astral.sh/uv/):
+
+```bash
+uv sync --all-extras
+uv run latex2word convert paper.tex -o paper.docx
+```
+
+Or from Python:
+
+```python
+from latex2word import convert_source, convert_file
+
+out_path, result = convert_file("paper.tex")
+print(result.report.summary())   # math coverage + warnings
+```
+
+## What works today
+
+- **Reference Word templates** ★: `--reference-doc TEMPLATE.docx` adopts a
+  journal/corporate template's styles, theme and page geometry (size + margins),
+  so the output matches the required look — while keeping the live fields below.
+  Our custom styles are merged in so nothing renders unstyled.
+- **Structure & styles**: `\title`/`\author`/`\date`/`abstract`, `\section`…
+  `\subparagraph` → Word Title/Heading 1–4 (visible in the Navigation pane),
+  paragraphs, `\textbf`/`\emph`/`\texttt`/`\underline`/`\textsc`, quotes, code.
+  Sections are **auto-numbered** (multilevel `1` / `1.1` / `1.1.1`) like LaTeX,
+  with `\section*` unnumbered; `\ref` to a section shows its live number. In
+  **book/report** documents `\chapter` is the top level (sections nest under it)
+  and `\appendix` switches to lettered headings (`A`, `A.1`).
+- **Math (direct LaTeX→OMML)**: inline `$…$`, display `\[…\]`,
+  `equation`/`align`/`gather`; fractions, sub/superscripts, roots, `\sum`/`\int`
+  with limits, accents, `\left…\right` delimiters, matrices/`cases`, Greek and
+  hundreds of symbols, `\mathbb`/`\mathcal`/`\mathbf`, functions (`\sin`, `\lim`).
+  `align*`/`aligned` line up at the `&` (a column-justified matrix); numbered
+  `align` keeps a live number per line.
+- **Live fields** ★: numbered equations get `SEQ Equation` fields inside
+  bookmarks; `\ref`/`\eqref`/`\pageref` become `REF`/`PAGEREF` fields; figure
+  and table captions get `SEQ Figure`/`SEQ Table`. Numbers auto-renumber in
+  Word on field refresh. `--number-by-section` switches to `N.M` per-section
+  numbering (`STYLEREF` + `SEQ \s`), book/report style.
+- **Table of contents** ★: `\tableofcontents` → a live Word `TOC` field (rebuilds
+  from heading styles on refresh); `\listoffigures`/`\listoftables` → caption-
+  sequence lists. Schema-valid and round-tripping.
+- **Lists, tables, figures**: `itemize`/`enumerate`, `tabular`/`longtable` with
+  `booktabs`, `\multicolumn`→column span, `\multirow`→vertical merge, and
+  repeating header rows; captioned `figure`/`table`, `\includegraphics`
+  (PNG/JPEG embedded directly; **PDF figures rasterised** to PNG when the
+  optional `tex2word[pdf]` extra — pypdfium2 — is installed). An
+  `\includegraphics` in running text (an icon/logo) is embedded **inline**.
+- **Custom macros**: `\newcommand`/`\renewcommand`/`\def` are expanded before
+  parsing. Common `mathtools`/`physics` math (`\abs`, `\norm`, `\dv`, `\ket`, …)
+  and `siunitx` (`\SI{9.81}{\meter\per\second\squared}` → `9.81 m/s²`, `\num`,
+  `\ang`) work as built-ins when not user-defined. **Acronyms** (`glossaries`):
+  `\newacronym` + `\gls`/`\acrshort`/`\acrlong`/`\acrfull` expand with the
+  first-use "long (short)" rule.
+- **Footnotes**: `\footnote` → native Word footnotes (`footnotes.xml`), not
+  inlined text; footnote bodies keep their formatting and math.
+- **Inline verbatim & smart refs**: `\verb|...|` → literal monospace;
+  `\cref`/`\Cref`/`\autoref` add cleveref-style type prefixes ("fig. N" /
+  "Figure N").
+- **Theorem environments**: `theorem`/`lemma`/`proof`/`definition`/… render
+  with a bold numbered lead (live `SEQ` per kind), optional `[title]`, and a
+  QED mark for proofs; `\ref` to a theorem shows its number.
+- **Algorithms**: `algorithm` + `algorithmic`/`algpseudocode`/`algorithm2e` →
+  numbered, indented pseudocode with bold keywords, inline OMML math, and a live
+  `SEQ Algorithm` caption.
+- **Graceful degradation**: unknown constructs never abort; they pass through
+  best-effort and are logged to the conversion report (math coverage telemetry
+  included). The math **decision-cascade** (direct OMML → LaTeX→MathML→OMML
+  secondary path → image fallback `--math-image-fallback` → raw) records which
+  path each equation took.
+- **Round-trip**: the IR is embedded as a JSON manifest custom part, so the
+  exact IR can be recovered from the `.docx` (`latex2word.roundtrip.recover_ir`)
+  and converted **back to LaTeX** (`latex2word to-latex out.docx`); the corpus
+  `latex→docx→latex` keeps the same block structure. Reconcile (on by default)
+  merges Word edits against the manifest, and **Word Track Changes are accepted**
+  on read (insertions kept, deletions dropped).
+- **Reports & validation**: `--report report.json|report.html` writes a coverage
+  report; `latex2word.validate.validate_docx` structurally validates output;
+  `latex2word benchmark <dir>` reports a quantitative baseline (math-OMML %,
+  validity, warnings, 0-abort) across a paper set (CI-gated on the corpus + UATs:
+  currently 100% native-OMML math, 100% valid, 0 aborts).
+- **Reproducible**: set `SOURCE_DATE_EPOCH` and the same input yields
+  byte-identical output (the `.docx` ZIP is built deterministically).
+- **Live citations** (opt-in `--citations zotero`): emit
+  `ADDIN ZOTERO_ITEM CSL_CITATION` / `CSL_BIBLIOGRAPHY` fields so citations are
+  editable by Zotero/Mendeley in Word (default is static formatted text).
+- **Real CSL styles** (opt-in `--csl style.csl`, needs `tex2word[csl]`): a
+  genuine `citeproc-py` engine formats in-text citations and the reference list
+  against any `.csl` style, with proper sorting; the built-in heuristic is the
+  fallback. `\nocite{key}`/`\nocite{*}` are honoured.
+
+- **Front-end choice**: the default **`pure`** front-end (pylatexenc-based) is
+  the validated engine — it converts the corpus and three real-paper UATs at
+  100% native-OMML math, 100% valid output, 0 aborts. `--frontend latexml` is
+  **experimental**: it shells out to a real `latexml` install for genuine TeX
+  expansion, but is not yet proven end-to-end (it silently falls back to `pure`
+  on any failure; see the advisory `real-tool` CI lane).
+
+## Architecture
+
+```
+LaTeX ─▶ front-end (preprocess, macro-expand, pylatexenc walk) ─▶ IR
+      ─▶ transforms (cross-reference resolution) ─▶ IR
+      ─▶ back-end (raw OOXML via lxml: document/styles/numbering) ─▶ .docx
+```
+
+The **IR** ([`src/latex2word/ir.py`](src/latex2word/ir.py)) is the format-neutral seam, so a LaTeXML front-end can replace the static parser post-V1 without touching the back-end.
+
+## Development
+
+```bash
+uv run pytest          # tests
+uv run ruff check src tests
+uv run mypy src
+uv run pre-commit install   # optional: run the lint/type gate on every commit
+```
+
+Releases: pushing a `vX.Y.Z` tag builds the wheel/sdist and publishes to PyPI
+(via the `Release` workflow, using PyPI Trusted Publishing). Notable changes are
+recorded in [`CHANGELOG.md`](CHANGELOG.md).
+
+## License
+
+MIT — see [`LICENSE`](LICENSE).
+
+## Author
+
+Yifan Yang <yfyang.86@hotmail.com>

tex2word-0.8.1/README.md

@@ -0,0 +1,168 @@
+# latex2word
+
+An open-source, cross-platform **LaTeX → Microsoft Word (`.docx`)** converter
+that produces *genuinely editable* Word: native paragraph styles, **native OMML
+equations** (editable in Word's equation editor, not images), and **live,
+auto-renumbering fields** for equation/figure/table numbers and
+cross-references.
+
+> **Status: production-grade.** Foundation, math core (direct LaTeX→OMML), the
+> live cross-reference/field plumbing (the differentiator), image embedding, the
+> BibTeX bibliography, and the robustness layer (math cascade, coverage report,
+> OOXML validator, round-trip manifest) are all in. See
+> [`CHANGELOG.md`](CHANGELOG.md) for the release history.
+
+## Why
+
+Pandoc/`texmath` is the open-source reference but **drops equation numbers**,
+can dump raw LaTeX for labelled equations, and emits *static* cross-references.
+No open tool produces editable styles **and** native OMML **and** live
+field-based numbering. That gap is the product.
+
+## Install & use
+
+Requires Python 3.12+.
+
+From PyPI:
+
+```bash
+pip install tex2word                 # core (PNG/JPEG figures)
+pip install "tex2word[pdf]"          # + PDF figure rasterisation (pypdfium2, Apache-2.0)
+pip install "tex2word[mathml]"       # + LaTeX->MathML->OMML for hard math (latex2mathml)
+pip install "tex2word[csl]"          # + real CSL citation styles (citeproc-py)
+pip install "tex2word[pdf,mathml,csl,mathimg]"   # everything
+
+latex2word convert paper.tex -o paper.docx
+latex2word convert paper.tex -o paper.docx --report report.json
+latex2word convert paper.tex -o paper.docx --reference-doc journal.docx
+```
+
+Or, for a development checkout with [uv](https://docs.astral.sh/uv/):
+
+```bash
+uv sync --all-extras
+uv run latex2word convert paper.tex -o paper.docx
+```
+
+Or from Python:
+
+```python
+from latex2word import convert_source, convert_file
+
+out_path, result = convert_file("paper.tex")
+print(result.report.summary())   # math coverage + warnings
+```
+
+## What works today
+
+- **Reference Word templates** ★: `--reference-doc TEMPLATE.docx` adopts a
+  journal/corporate template's styles, theme and page geometry (size + margins),
+  so the output matches the required look — while keeping the live fields below.
+  Our custom styles are merged in so nothing renders unstyled.
+- **Structure & styles**: `\title`/`\author`/`\date`/`abstract`, `\section`…
+  `\subparagraph` → Word Title/Heading 1–4 (visible in the Navigation pane),
+  paragraphs, `\textbf`/`\emph`/`\texttt`/`\underline`/`\textsc`, quotes, code.
+  Sections are **auto-numbered** (multilevel `1` / `1.1` / `1.1.1`) like LaTeX,
+  with `\section*` unnumbered; `\ref` to a section shows its live number. In
+  **book/report** documents `\chapter` is the top level (sections nest under it)
+  and `\appendix` switches to lettered headings (`A`, `A.1`).
+- **Math (direct LaTeX→OMML)**: inline `$…$`, display `\[…\]`,
+  `equation`/`align`/`gather`; fractions, sub/superscripts, roots, `\sum`/`\int`
+  with limits, accents, `\left…\right` delimiters, matrices/`cases`, Greek and
+  hundreds of symbols, `\mathbb`/`\mathcal`/`\mathbf`, functions (`\sin`, `\lim`).
+  `align*`/`aligned` line up at the `&` (a column-justified matrix); numbered
+  `align` keeps a live number per line.
+- **Live fields** ★: numbered equations get `SEQ Equation` fields inside
+  bookmarks; `\ref`/`\eqref`/`\pageref` become `REF`/`PAGEREF` fields; figure
+  and table captions get `SEQ Figure`/`SEQ Table`. Numbers auto-renumber in
+  Word on field refresh. `--number-by-section` switches to `N.M` per-section
+  numbering (`STYLEREF` + `SEQ \s`), book/report style.
+- **Table of contents** ★: `\tableofcontents` → a live Word `TOC` field (rebuilds
+  from heading styles on refresh); `\listoffigures`/`\listoftables` → caption-
+  sequence lists. Schema-valid and round-tripping.
+- **Lists, tables, figures**: `itemize`/`enumerate`, `tabular`/`longtable` with
+  `booktabs`, `\multicolumn`→column span, `\multirow`→vertical merge, and
+  repeating header rows; captioned `figure`/`table`, `\includegraphics`
+  (PNG/JPEG embedded directly; **PDF figures rasterised** to PNG when the
+  optional `tex2word[pdf]` extra — pypdfium2 — is installed). An
+  `\includegraphics` in running text (an icon/logo) is embedded **inline**.
+- **Custom macros**: `\newcommand`/`\renewcommand`/`\def` are expanded before
+  parsing. Common `mathtools`/`physics` math (`\abs`, `\norm`, `\dv`, `\ket`, …)
+  and `siunitx` (`\SI{9.81}{\meter\per\second\squared}` → `9.81 m/s²`, `\num`,
+  `\ang`) work as built-ins when not user-defined. **Acronyms** (`glossaries`):
+  `\newacronym` + `\gls`/`\acrshort`/`\acrlong`/`\acrfull` expand with the
+  first-use "long (short)" rule.
+- **Footnotes**: `\footnote` → native Word footnotes (`footnotes.xml`), not
+  inlined text; footnote bodies keep their formatting and math.
+- **Inline verbatim & smart refs**: `\verb|...|` → literal monospace;
+  `\cref`/`\Cref`/`\autoref` add cleveref-style type prefixes ("fig. N" /
+  "Figure N").
+- **Theorem environments**: `theorem`/`lemma`/`proof`/`definition`/… render
+  with a bold numbered lead (live `SEQ` per kind), optional `[title]`, and a
+  QED mark for proofs; `\ref` to a theorem shows its number.
+- **Algorithms**: `algorithm` + `algorithmic`/`algpseudocode`/`algorithm2e` →
+  numbered, indented pseudocode with bold keywords, inline OMML math, and a live
+  `SEQ Algorithm` caption.
+- **Graceful degradation**: unknown constructs never abort; they pass through
+  best-effort and are logged to the conversion report (math coverage telemetry
+  included). The math **decision-cascade** (direct OMML → LaTeX→MathML→OMML
+  secondary path → image fallback `--math-image-fallback` → raw) records which
+  path each equation took.
+- **Round-trip**: the IR is embedded as a JSON manifest custom part, so the
+  exact IR can be recovered from the `.docx` (`latex2word.roundtrip.recover_ir`)
+  and converted **back to LaTeX** (`latex2word to-latex out.docx`); the corpus
+  `latex→docx→latex` keeps the same block structure. Reconcile (on by default)
+  merges Word edits against the manifest, and **Word Track Changes are accepted**
+  on read (insertions kept, deletions dropped).
+- **Reports & validation**: `--report report.json|report.html` writes a coverage
+  report; `latex2word.validate.validate_docx` structurally validates output;
+  `latex2word benchmark <dir>` reports a quantitative baseline (math-OMML %,
+  validity, warnings, 0-abort) across a paper set (CI-gated on the corpus + UATs:
+  currently 100% native-OMML math, 100% valid, 0 aborts).
+- **Reproducible**: set `SOURCE_DATE_EPOCH` and the same input yields
+  byte-identical output (the `.docx` ZIP is built deterministically).
+- **Live citations** (opt-in `--citations zotero`): emit
+  `ADDIN ZOTERO_ITEM CSL_CITATION` / `CSL_BIBLIOGRAPHY` fields so citations are
+  editable by Zotero/Mendeley in Word (default is static formatted text).
+- **Real CSL styles** (opt-in `--csl style.csl`, needs `tex2word[csl]`): a
+  genuine `citeproc-py` engine formats in-text citations and the reference list
+  against any `.csl` style, with proper sorting; the built-in heuristic is the
+  fallback. `\nocite{key}`/`\nocite{*}` are honoured.
+
+- **Front-end choice**: the default **`pure`** front-end (pylatexenc-based) is
+  the validated engine — it converts the corpus and three real-paper UATs at
+  100% native-OMML math, 100% valid output, 0 aborts. `--frontend latexml` is
+  **experimental**: it shells out to a real `latexml` install for genuine TeX
+  expansion, but is not yet proven end-to-end (it silently falls back to `pure`
+  on any failure; see the advisory `real-tool` CI lane).
+
+## Architecture
+
+```
+LaTeX ─▶ front-end (preprocess, macro-expand, pylatexenc walk) ─▶ IR
+      ─▶ transforms (cross-reference resolution) ─▶ IR
+      ─▶ back-end (raw OOXML via lxml: document/styles/numbering) ─▶ .docx
+```
+
+The **IR** ([`src/latex2word/ir.py`](src/latex2word/ir.py)) is the format-neutral seam, so a LaTeXML front-end can replace the static parser post-V1 without touching the back-end.
+
+## Development
+
+```bash
+uv run pytest          # tests
+uv run ruff check src tests
+uv run mypy src
+uv run pre-commit install   # optional: run the lint/type gate on every commit
+```
+
+Releases: pushing a `vX.Y.Z` tag builds the wheel/sdist and publishes to PyPI
+(via the `Release` workflow, using PyPI Trusted Publishing). Notable changes are
+recorded in [`CHANGELOG.md`](CHANGELOG.md).
+
+## License
+
+MIT — see [`LICENSE`](LICENSE).
+
+## Author
+
+Yifan Yang <yfyang.86@hotmail.com>

tex2word-0.8.1/pyproject.toml

@@ -0,0 +1,103 @@
+[project]
+name = "tex2word"
+version = "0.8.1"
+description = "Open-source, production-grade LaTeX -> Microsoft Word (.docx) converter with native OMML math and live fields"
+readme = "README.md"
+requires-python = ">=3.12"
+authors = [
+    { name = "Yifan Yang", email = "yfyang.86@hotmail.com" }
+]
+maintainers = [
+    { name = "Yifan Yang", email = "yfyang.86@hotmail.com" }
+]
+license = "MIT"
+license-files = ["LICENSE"]
+keywords = [
+    "latex",
+    "word",
+    "docx",
+    "converter",
+    "ooxml",
+    "omml",
+    "tex",
+    "office",
+    "document",
+]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Environment :: Console",
+    "Intended Audience :: Science/Research",
+    "Intended Audience :: End Users/Desktop",
+    "Operating System :: OS Independent",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Topic :: Text Processing :: Markup :: LaTeX",
+    "Topic :: Text Processing :: Markup :: XML",
+    "Topic :: Office/Business :: Office Suites",
+    "Typing :: Typed",
+]
+dependencies = [
+    "lxml>=5.0",
+    "pylatexenc>=2.10",
+]
+
+[project.urls]
+Homepage = "https://github.com/yfyang86/tex2word"
+Repository = "https://github.com/yfyang86/tex2word"
+Issues = "https://github.com/yfyang86/tex2word/issues"
+Changelog = "https://github.com/yfyang86/tex2word/blob/main/CHANGELOG.md"
+
+[project.optional-dependencies]
+# Rasterise PDF figures to embeddable PNG (Word can't embed PDF directly).
+# pypdfium2 (PDFium) + Pillow are both permissive (Apache-2.0/BSD/HPND) -- no GPL.
+pdf = ["pypdfium2>=4", "pillow>=10"]
+# Secondary math path: LaTeX -> presentation MathML -> OMML for the hard 20%.
+mathml = ["latex2mathml>=3.77"]
+# Image fallback for math that can't become OMML (pure-Python; no TeX needed).
+mathimg = ["matplotlib>=3.7"]
+# Real CSL citation processor: format citations/bibliography against a .csl style.
+csl = ["citeproc-py>=0.6"]
+
+[project.scripts]
+latex2word = "latex2word.cli:main"
+
+[build-system]
+requires = ["uv_build>=0.8.17,<0.9.0"]
+build-backend = "uv_build"
+
+# The PyPI distribution is published as "tex2word", but the import package
+# remains "latex2word", so point the build backend at the actual module dir.
+[tool.uv.build-backend]
+module-name = "latex2word"
+
+[dependency-groups]
+dev = [
+    "pytest>=8.0",
+    "ruff>=0.6",
+    "mypy>=1.11",
+    "lxml-stubs>=0.5.1",
+    "pypdfium2>=4",
+    "pillow>=10",
+    "latex2mathml>=3.77",
+    "matplotlib>=3.7",
+]
+
+[tool.ruff]
+line-length = 100
+target-version = "py312"
+
+[tool.ruff.lint]
+select = ["E", "F", "I", "UP", "B"]
+
+[tool.ruff.lint.per-file-ignores]
+# These modules embed long XML/HTML string literals (templates / reports).
+"src/latex2word/backend/package.py" = ["E501"]
+"src/latex2word/report.py" = ["E501"]
+
+[tool.mypy]
+python_version = "3.12"
+ignore_missing_imports = true
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]

tex2word-0.8.1/src/latex2word/__init__.py

@@ -0,0 +1,13 @@
+"""latex2word: an open-source LaTeX -> Microsoft Word (.docx) converter.
+
+See ``README.md`` for the architecture overview. The public entry points are
+:func:`~latex2word.pipeline.convert_source` and
+:func:`~latex2word.pipeline.convert_file`.
+"""
+
+from __future__ import annotations
+
+from .pipeline import ConversionResult, convert_file, convert_source
+
+__all__ = ["ConversionResult", "convert_file", "convert_source"]
+__version__ = "0.8.1"

tex2word-0.8.1/src/latex2word/backend/__init__.py

@@ -0,0 +1 @@
+"""OOXML back-end: turns the IR into a `.docx` (OPC) package."""