olmocr/pyproject.toml

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

[project]
# See https://setuptools.pypa.io/en/latest/userguide/quickstart.html for more project configuration options.
name = "olmocr"
description = "Fast, efficient, and high quality OCR powered by open visual language models"
dynamic = ["version"]
readme = "README.md"
classifiers = [
    "Intended Audience :: Science/Research",
    "Development Status :: 3 - Alpha",
    "License :: OSI Approved :: Apache Software License",
    "Programming Language :: Python :: 3",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
]
authors = [
    {name = "Allen Institute for Artificial Intelligence", email = "jakep@allenai.org"}
]
requires-python = ">=3.11"
dependencies = [
  "cached-path",
  "smart_open",
  "pypdf>=5.2.0",
  "pypdfium2",
  "cryptography",
  "lingua-language-detector",
  "Pillow",
  "ftfy",
  "bleach",
  "markdown2",
  "filelock",
  "orjson",
  "requests",
  "zstandard",
  "boto3",
  "httpx",
  "torch>=2.7.0",
  "transformers>=4.51.1",
  "img2pdf",
  "beaker-py",
]
license = {file = "LICENSE"}

[project.urls]
Homepage = "https://github.com/allenai/olmocr"
Repository = "https://github.com/allenai/olmocr"
Changelog = "https://github.com/allenai/olmocr/blob/main/CHANGELOG.md"
# Documentation = "https://olmocr.readthedocs.io/"

[project.optional-dependencies]
gpu = [
    "vllm==0.9.1"
]

dev = [
    "ruff",
    "mypy",
    "black",
    "isort",
    "pytest",
    "pytest-sphinx",
    "pytest-cov",
    "twine>=1.11.0",
    "build",
    "setuptools",
    "wheel",
    "Sphinx>=4.3.0,<7.1.0",
    "furo==2023.7.26",
    "myst-parser>=1.0,<2.1",
    "sphinx-copybutton==0.5.2",
    "sphinx-autobuild==2021.3.14",
    "sphinx-autodoc-typehints==1.23.3",
    "packaging",
    "necessary",
    "peft",
    "datasets",
    "omegaconf",
    "spacy",
]

bench = [
    "tinyhost",
    "fuzzysearch",
    "rapidfuzz",
    "sequence_align",
    "syntok",
    "openai",
    "google-genai",
    "playwright",
    "mistralai",
    "lxml",
    "flask",
]

train = [
    "torch",
    "torchvision",
    "accelerate",
    "datasets",
    "peft",
    "wandb",
    "omegaconf",
    "s3fs",
    "necessary",
    "einops",
    "transformers>=4.45.1"
]

elo = [
    "numpy",
    "scipy",
    "pandas",
    "matplotlib"
]

[tool.setuptools.packages.find]
exclude = [
    "*.tests",
    "*.tests.*",
    "tests.*",
    "tests",
    "docs*",
    "scripts*"
]

[tool.setuptools]
include-package-data = true

[tool.setuptools.package-data]
olmocr = [
    "py.typed",
    "viewer/*.html",
    "eval/*.html",
]

[tool.setuptools.dynamic]
version = {attr = "olmocr.version.VERSION"}

[tool.black]
line-length = 160
include = '\.pyi?$'
exclude = '''
(
      __pycache__
    | \.git
    | \.mypy_cache
    | \.pytest_cache
    | \.vscode
    | \.venv
    | \bdist\b
    | \bdoc\b
)
'''

[tool.isort]
profile = "black"
multi_line_output = 3

# You can override these pyright settings by adding a personal pyrightconfig.json file.
[tool.pyright]
reportPrivateImportUsage = false

[tool.ruff]
line-length = 160
target-version = "py311"
exclude = ["olmocr/train/molmo", "tests/*"]
ignore = ["E722"] #igore bare except

[tool.ruff.per-file-ignores]
"__init__.py" = ["F401"]

[tool.mypy]
ignore_missing_imports = true
no_site_packages = true
check_untyped_defs = true
exclude = ["olmocr/train/molmo/", "tests/*"]

[[tool.mypy.overrides]]
module = "tests.*"
strict_optional = false

[tool.pytest.ini_options]
testpaths = "tests/"
python_classes = [
  "Test*",
  "*Test"
]
log_format = "%(asctime)s - %(levelname)s - %(name)s - %(message)s"
log_level = "DEBUG"
markers = [
    "nonci: mark test as not intended for CI runs"
]