pyproject.toml

[build-system]
build-backend = "maturin"
requires = ["maturin>=1.13.1,<2"]

[dependency-groups]
dev = [
    "covdefaults>=2.3.0",
    "httpx>=0.28.1",
    "maturin>=1.13.1",
    "pyrefly>=1.1.1",
    "pytest>=9.0.3",
    "pytest-asyncio>=1.3.0",
    "pytest-cov>=7.1.0",
    "pytest-mock>=3.15.1",
    "pytest-rerunfailures>=16.2",
    "pytest-timeout>=2.4.0",
    "types-pillow>=10.2",
    "types-psutil>=7.2.2.20260508",
    "validate-pyproject>=0.24.1",
]
bench-docling = ["docling>=2.64.1"]
bench-markitdown = ["markitdown[all]>=0.1.4"]
# `six` is an undeclared runtime import in mineru's pytorch_paddle OCR module
# (mineru.model.utils.pytorchocr) — without it every extraction dies with
# `ModuleNotFoundError: No module named 'six'`. Pin it in until mineru fixes
# its own metadata.
bench-mineru = ["mineru[pipeline]>=2.6.7", "onnxruntime", "six"]
bench-pymupdf4llm = ["pillow>=10.0.0", "pymupdf-layout>=0.0.1", "pymupdf4llm>=0.0.17"]
# Format-specific extras for the `fast` (no-OCR) partition strategy the
# benchmark uses (pdfminer/python-docx/python-pptx/openpyxl). Avoids the
# `[all-docs]` `unstructured-inference`/`pypdfium2>=5` stack that pulled the
# Python-3.11-incompatible llvmlite 0.36 and conflicted with mineru's pypdfium2.
bench-unstructured = ["unstructured[docx,pdf,pptx,xlsx,md,odt,rtf,epub,org,rst]==0.18.32"]
doc = ["mkdocstrings>=1.0.4", "mkdocstrings-python>=2.0.3", "zensical"]

[tool.uv]
# Restrict the resolution universe to the Python versions our CI and dev
# environments actually use. Without this, requires-python = ">=3.10" forces
# uv to keep both onnxruntime 1.24.x (last build with 3.10 wheels) and 1.26.x
# (3.11+) in the lock; the install step then ambiguously matches both rows
# on a single bench group and errors with "Found multiple packages matching
# `onnxruntime`" when syncing bench-mineru or bench-unstructured. Restricting
# uv's environment universe to 3.11+ collapses each transitive dep to a
# single version while leaving the published abi3-py310 wheel untouched (that
# distribution is driven by maturin, not uv).
environments = ["python_version >= '3.11'"]
workspace.members = ["packages/python"]
conflicts = [
    [
        { package = "xberg", group = "bench-docling" },
        { package = "xberg", group = "bench-markitdown" },
        { package = "xberg", group = "bench-mineru" },
        { package = "xberg", group = "bench-pymupdf4llm" },
        { package = "xberg", group = "bench-unstructured" },
    ],
]
cache-keys = [
    { file = "pyproject.toml" },
    { file = "Cargo.toml" },
    { file = "Cargo.lock" },
    { file = "crates/**/*.rs" },
    { file = "crates/**/Cargo.toml" },
]

[tool.pytest]
ini_options.timeout = 300
ini_options.testpaths = ["packages/python/tests"]
ini_options.norecursedirs = ["dist", "build", "*.egg"]
ini_options.markers = [
    "slow: marks tests as slow (deselect with '-m \"not slow\"')",
    "integration: marks tests as integration tests (requires running services)",
]
ini_options.asyncio_mode = "auto"
ini_options.asyncio_default_fixture_loop_scope = "function"
ini_options.filterwarnings = [
    "ignore:Exception ignored in:pytest.PytestUnraisableExceptionWarning",
    "ignore:pkg_resources is deprecated as an API:DeprecationWarning",
    "ignore:ast.Num is deprecated and will be removed in Python 3.14:DeprecationWarning",
    "ignore:datetime.datetime.utcfromtimestamp() is deprecated:DeprecationWarning",
    "ignore:Deprecated call to `pkg_resources.declare_namespace('google')`:DeprecationWarning",
]

[tool.coverage]
run.branch = true
run.omit = ["benchmarks/*", "packages/python/tests/*", "scripts/*"]
run.plugins = ["covdefaults"]
run.source = ["packages/python/xberg"]
report.exclude_lines = [
    "@(abc\\.)?abstractmethod",
    "class .*\\bProtocol\\):",
    "def __repr__",
    "except ImportError:",
    "if __name__ == .__main__.:",
    "if sys.version_info",
    "if TYPE_CHECKING:",
    "pragma: no cover",
    "raise NotImplementedError",
]
report.fail_under = 90