[build-system]
build-backend = "maturin"
requires = [ "maturin>=1,<2" ]

[project]
name = "kreuzberg"
version = "5.0.0rc3"
description = "High-performance document intelligence library"
keywords = [ "document", "extraction", "ocr", "pdf", "text" ]
license = "Elastic-2.0"
license-files = [ "LICENSE" ]
authors = [ { name = "Na'aman Hirschfeld <naaman@kreuzberg.dev>" } ]
requires-python = ">=3.10"
classifiers = [
  "Programming Language :: Python :: 3 :: Only",
  "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",
]
urls.repository = "https://github.com/kreuzberg-dev/kreuzberg"
homepage = "https://kreuzberg.dev"

[dependency-groups]
dev = [ "mypy>=1.19", "ruff>=0.14.8" ]

[tool.maturin]
module-name = "kreuzberg._kreuzberg"
manifest-path = "../../crates/kreuzberg-py/Cargo.toml"
# abi3-py310 produces a single wheel per platform that loads on Python 3.10+,
# avoiding a per-Python-version build matrix.
features = [ "pyo3/extension-module", "pyo3/abi3-py310" ]
python-packages = [ "kreuzberg" ]
# Bundle the core Rust crate so `pip install` can build from sdist on
# platforms without a precompiled wheel (e.g. Alpine/musl). Without this
# the workspace [patch.crates-io] (when present) points at a path that is
# missing from the tarball and the source build fails.
include = [
  { path = "../../crates/kreuzberg/**/*", format = "sdist" },
]

[tool.ruff]
target-version = "py310"
line-length = 120
format.docstring-code-line-length = 120
format.docstring-code-format = true
lint.select = [ "ALL" ]
lint.ignore = [
  "ANN401",
  "ASYNC109",
  "ASYNC110",
  "BLE001",
  "COM812",
  "D100",
  "D104",
  "D107",
  "D205",
  "E501",
  "EM",
  "FBT",
  "FIX",
  "ISC001",
  "PD011",
  "PGH003",
  "PLR2004",
  "PLW0603",
  "S104",
  "S110",
  "S603",
  "TD",
  "TRY",
]
lint.per-file-ignores."kreuzberg/__init__.py" = [ "I001" ]
# The alef Python codegen still emits cosmetic warnings on the wrapper
# modules: api.py keeps the legacy `from typing import AsyncIterator` and a
# single-line import block, options.py carries # noqa: TC001 / F401 markers
# that turn out unused on every regen, __init__.py star-imports re-sort with
# a different convention. Silence these specific rules on the wrappers until
# the codegen is updated to emit ruff-clean output.
lint.per-file-ignores."kreuzberg/api.py" = [ "F401", "I001", "UP035" ]
lint.per-file-ignores."kreuzberg/options.py" = [ "F401", "RUF100" ]
lint.per-file-ignores."tests/**" = [ "ANN", "D103", "PLR2004", "S101" ]
lint.mccabe.max-complexity = 15
lint.pydocstyle.convention = "google"
lint.pylint.max-args = 10
lint.pylint.max-branches = 15
lint.pylint.max-returns = 10

[tool.mypy]
python_version = "3.10"
strict = true
show_error_codes = true
implicit_reexport = false
namespace_packages = true
overrides = [
  # The alef-emitted `api.py` wrapper has a structural mismatch between its
  # `options.*` dataclass signatures and the `_internal_bindings.*` pyclass
  # types pyo3 accepts/returns at runtime. pyo3 reconciles them dynamically via
  # FromPyObject — the Python e2e suite exercises the runtime path — but mypy
  # sees only the static-type discrepancy. Disable the four error codes the
  # discrepancy raises until the codegen emits matching `_to_rust_*` calls and
  # casts the return values.
  { module = "kreuzberg.api", disable_error_code = [ "call-arg", "arg-type", "return-value", "attr-defined" ] },
]