pyproject.toml

[build-system]
requires = ["setuptools>=61.0", "wheel"]
build-backend = "setuptools.build_meta"

[project]
name = "longparser"
version = "0.1.4"
description = "Privacy-first document intelligence engine — converts PDFs, DOCX, PPTX, XLSX, and CSV into AI-ready Markdown + structured JSON for RAG pipelines."
readme = {file = "README.md", content-type = "text/markdown"}
requires-python = ">=3.10"
license = "MIT"
authors = [
    {name = "ENDEVSOLS Team", email = "technology@endevsols.com"},
]
keywords = [
    "pdf", "document", "parsing", "ocr", "rag", "ai",
    "docling", "chunking", "extraction", "retrieval-augmented-generation",
    "longparser",
]
classifiers = [
    "Development Status :: 4 - Beta",
    "Intended Audience :: Developers",
    "Intended Audience :: Science/Research",
    "Operating System :: OS Independent",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
    "Programming Language :: Python :: 3.13",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
    "Topic :: Text Processing :: General",
    "Typing :: Typed",
]
dependencies = [
    "pydantic>=2.0,<3",
    "docling>=2.14",
    "docling-core>=2.13",
    "langgraph-checkpoint-mongodb>=0.3.1",
    "fast-langdetect>=0.3,<1.0",  # Apache-2.0 — document language detection
]

[project.optional-dependencies]
# PPTX support (indent-level detection via python-pptx)
pptx = [
    "python-pptx>=1.0",
]
# LangChain integration adapter
langchain = [
    "langchain-core>=0.2",
]
# LlamaIndex integration adapter
llamaindex = [
    "llama-index-core>=0.10",
]
# ----------- v0.1.4: Optional extraction backends -----------
# ⚠️ pymupdf4llm is AGPL-3.0 licensed. See LICENSE-THIRD-PARTY.md.
# Only loaded when user sets backend="pymupdf".
pymupdf = [
    "pymupdf4llm>=1.27",
]
# ⚠️ marker-pdf is GPL-3.0. GPU recommended. Future release.
# marker = [
#     "marker-pdf",
# ]
# ⚠️ surya-ocr is GPL-3.0. GPU recommended. Future release.
# surya = [
#     "surya-ocr>=0.17",
# ]
# FastAPI REST server + MongoDB + job queue + LangChain chat engine
server = [
    "fastapi>=0.115",
    "uvicorn[standard]>=0.34",
    "python-multipart>=0.0.9",
    "motor>=3.6",
    "arq>=0.26",
    "python-magic>=0.4.27",
    "python-dotenv>=1.0",
    "langchain>=0.3",
    "langchain-openai>=0.3",
    "langchain-google-genai>=2.0",
    "langchain-groq>=0.3",
    "langchain-mongodb>=0.3",
    "langchain-huggingface>=0.1",
    "langchain-chroma>=0.2",
    "langgraph>=0.2",
    "langgraph-checkpoint>=2.0",
    "tiktoken>=0.7",
    "redis>=5.0",
]

# ---------------------------------------------------------------------------
# Embedding engines
# Pre-install torch BEFORE these extras:
#   CPU:  pip install torch --index-url https://download.pytorch.org/whl/cpu
#   GPU:  pip install torch  (default PyPI = CUDA)
# ---------------------------------------------------------------------------
embeddings = [
    "sentence-transformers>=3.0",
]
# Alias kept for clarity — same package, torch variant controls CPU/GPU
embeddings-cpu = [
    "sentence-transformers>=3.0",
]
embeddings-gpu = [
    "sentence-transformers>=3.0",
]

# ---------------------------------------------------------------------------
# Vector databases — faiss has separate CPU and GPU packages on PyPI
# ---------------------------------------------------------------------------
chroma = [
    "chromadb>=0.5",
]
faiss-cpu = [
    "faiss-cpu>=1.8",
]
faiss-gpu = [
    "faiss-gpu>=1.7",
]
qdrant = [
    "qdrant-client>=1.12",
]

# ---------------------------------------------------------------------------
# LaTeX / equation OCR
# Pre-install torch (see above) before installing these extras.
# ---------------------------------------------------------------------------
# LaTeX OCR via pix2tex (PDF formulas)
latex-ocr = [
    "pix2tex>=0.1.4",
]
latex-ocr-cpu = [
    "pix2tex>=0.1.4",
]
latex-ocr-gpu = [
    "pix2tex>=0.1.4",
]

# LaTeX equation extraction from DOCX/PPTX (no torch needed)
docx-equations = [
    "docxlatex>=0.3.0",
    "defusedxml>=0.7.0",
]
# Page-level Math Formula Detection via pix2text YOLO (MIT licence)
mfd = [
    "pix2text>=1.1.1,<1.2",
]

# ---------------------------------------------------------------------------
# Convenience meta-extras
# Usage:
#   CPU-only (no CUDA):
#     pip install torch --index-url https://download.pytorch.org/whl/cpu
#     pip install "longparser[cpu]"
#
#   GPU / CUDA:
#     pip install torch          # default PyPI wheel = CUDA
#     pip install "longparser[gpu]"
# ---------------------------------------------------------------------------
cpu = [
    "longparser[server]",
    "longparser[embeddings-cpu]",
    "longparser[faiss-cpu]",
    "longparser[latex-ocr-cpu]",
    "longparser[chroma]",
    "longparser[pptx]",
    "longparser[langchain]",
    "longparser[llamaindex]",
]
gpu = [
    "longparser[server]",
    "longparser[embeddings-gpu]",
    "longparser[faiss-gpu]",
    "longparser[latex-ocr-gpu]",
    "longparser[chroma]",
    "longparser[pptx]",
    "longparser[langchain]",
    "longparser[llamaindex]",
]
# Legacy / catch-all (defaults to CPU-safe)
all = [
    "longparser[cpu]",
]
# Development / testing
dev = [
    "pytest>=8.0",
    "pytest-asyncio>=0.23",
    "pytest-cov>=5.0",
    "ruff>=0.4",
    "mypy>=1.10",
    "build>=1.0",
    "twine>=5.0",
    "httpx>=0.27",
    "anyio>=4.0",
]


[project.urls]
Homepage = "https://github.com/ENDEVSOLS/LongParser"
Repository = "https://github.com/ENDEVSOLS/LongParser"
Issues = "https://github.com/ENDEVSOLS/LongParser/issues"
Documentation = "https://endevsols.github.io/LongParser/"
Changelog = "https://github.com/ENDEVSOLS/LongParser/blob/main/CHANGELOG.md"

[tool.setuptools.packages.find]
where = ["src"]

[tool.ruff]
line-length = 100
target-version = "py310"

[tool.ruff.lint]
select = ["E", "F", "I", "UP", "B", "SIM"]
ignore = ["E501"]

[tool.ruff.lint.isort]
known-first-party = ["longparser"]

[tool.pytest.ini_options]
testpaths = ["tests"]
pythonpath = ["src"]
asyncio_mode = "auto"
addopts = "-v --tb=short"

[tool.mypy]
python_version = "3.10"
strict = false
ignore_missing_imports = true
warn_unused_ignores = true

[tool.coverage.run]
source = ["src/longparser"]
omit = [
    "tests/*",
    # These modules require heavy optional deps (server, langchain, docling GPU)
    # not installed in the lightweight CI environment
    "src/longparser/server/*",
    "src/longparser/extractors/docling_extractor.py",
    "src/longparser/extractors/latex_ocr.py",
    "src/longparser/integrations/*",
    "src/longparser/pipeline/*",
]


[tool.coverage.report]
show_missing = true
fail_under = 5

# CPU-only PyTorch index — saves ~2.3 GB vs CUDA build
[[tool.uv.index]]
name = "pytorch-cpu"
url = "https://download.pytorch.org/whl/cpu"
explicit = true

[tool.uv.sources]
torch = [{ index = "pytorch-cpu" }]
torchvision = [{ index = "pytorch-cpu" }]