pyproject.toml

[build-system]
requires = ["setuptools>=68", "wheel"]
build-backend = "setuptools.build_meta"

[project]
name = "exactdoc"
version = "1.0.1"
description = "High-fidelity PDF to DOCX that survives Google Docs, with a render-back diff that proves it"
readme = "README.md"
requires-python = ">=3.9"
license = { text = "Apache-2.0" }
authors = [{ name = "Ebin Babu Thomas" }]
keywords = [
    "pdf", "docx", "pdf-to-docx", "pdf-to-word", "google-docs",
    "document-conversion", "ooxml", "layout-analysis", "pdfium",
    "whitepaper", "research-paper",
]
classifiers = [
    "Development Status :: 5 - Production/Stable",
    "Intended Audience :: Developers",
    "License :: OSI Approved :: Apache Software License",
    "Programming Language :: Python :: 3",
    "Topic :: Text Processing :: Markup",
    "Topic :: Office/Business :: Office Suites",
]
dependencies = [
    "pypdfium2>=4.25",
    "python-docx>=1.1",
    "numpy>=1.24",
    "pillow>=10.0",
    "lxml>=4.9",
]

[project.optional-dependencies]
# Test harness: adversarial corpus generation + measurement.
test = ["reportlab>=4.0", "fpdf2>=2.7"]
# The PyMuPDF parser, and the base-14 text metrics the quality ladder shapes
# with. **PyMuPDF is AGPL-3.0-or-later; installing this extra changes your
# obligations for anything you distribute.** Nothing in a default install
# imports it: `backend="pymupdf"` and `--ladder` are the only doors, and both
# say so by name when the extra is absent. See docs/license-audit.md.
mupdf = ["pymupdf>=1.23"]
# Google Docs oracle -- measure fidelity in the renderer actually targeted.
gdocs = [
    "google-api-python-client>=2.100",
    "google-auth-httplib2>=0.1",
    "google-auth-oauthlib>=1.1",
]

[project.urls]
Homepage = "https://github.com/ebt55/exactdoc"
Issues = "https://github.com/ebt55/exactdoc/issues"

[project.scripts]
exactdoc = "exactdoc.cli:main"
exactdoc-gdocs = "exactdoc.gdocs:main"

[tool.setuptools]
packages = ["exactdoc"]