pyproject.toml

[tool.poetry]
name = "semdoc"
version = "0.1.0"
description = "Exract semantic information from documents"
authors = ["Michael Volz <info@michael-volz.eu>"]
license = "AGPL-3.0-or-later"
readme = "README.md"

[build-system]
requires = ["poetry-core"]
build-backend = "poetry.core.masonry.api"

[tool.poetry.dependencies]
python = ">=3.11,<3.13"
rich = "^13.7.1"
layoutparser = {extras = ["layoutmodels"], version = "^0.3.4"}
typer = "^0.12.1"
torch = "^2.2.2"
#detectron2 = {path = "detectron2", develop = false}
pytesseract = "^0.3.10"
urwid = "^2.6.11"
seaborn = "^0.13.2"
surya-ocr = "^0.4.12"
diskcache = "^5.6.3"
pyxdg = "^0.28"
pytest-profiling = "^1.7.0"
pyside6 = "^6.7.2"
pikepdf = "^9.0.0"
pypdf = "^4.2.0"
pypdf2 = "^3.0.1"
pdfrw = "^0.4"
llpdf = { git = "https://github.com/mibifuzzi/llpdf" }
img2pdf = "^0.5.1"
python-levenshtein = "^0.25.1"
python-lsp-server = {extras = ["rope"], version = "^1.11.0"}
lxml = "^5.2.2"

pymupdf = "^1.24.9"
camelot-py = "^0.11.0"
[tool.poetry.scripts]
semdoc = "semdoc.cli:run"
semdoc_opencv = "semdoc.deprecated.opencv2:segment_image_into_blocks"
semdoc_tesseract = "semdoc.deprecated.tesseract:run"
semdoc_server = "semdoc.server:app"

[tool.poetry.group.dev.dependencies]
pytest = "^8.1.1"

[tool.ruff]
exclude = [
    "detectron2",
]

[tool.pytest.ini_options]
filterwarnings = [
  "ignore::DeprecationWarning:pydantic",
  "ignore::DeprecationWarning:pytesseract",
  "ignore::DeprecationWarning:llpdf",
  "ignore::FutureWarning:huggingface_hub",
  "ignore::DeprecationWarning:multiprocessing",
  "ignore::FutureWarning:transformers",
]