pyproject.toml

[project]
name = "mmore"
version = "0.1.0"
description = "mmore: Scalable multimodal document extraction pipeline for custom RAG integration."
authors = [
    {name = "Example Author", email = "author@example.com"}
]
readme = "README.md"
dependencies = [
    "torch>=2.5.1",
    "Pillow==10.2.0",
    "PyMuPDF==1.24.14",
    "Unidecode==1.3.8",
    "clean-text==0.6.0",
    "docx2pdf==0.1.8",
    "lxml_html_clean",
    "markdown==3.7",
    "markdownify==0.14.1",
    "marker-pdf==0.3.10",
    "moviepy==2.1.1",
    "numpy==1.26.3",
    "openpyxl==3.1.5",
    "pandas==2.2.3",
    "py7zr==0.22.0",
    "python-docx==1.1.2",
    "python-pptx==1.0.2",
    "rarfile==4.2",
    "requests==2.28.1",
    "selenium==4.27.1",
    "surya-ocr==0.6.13",
    "trafilatura==1.4.0",
    "validators==0.34.0",
    "xlrd==2.0.1",
    "transformers==4.47.0",
    "datasets==2.19.1",
    "langchain==0.2.17",
    "langchain-milvus==0.1.7",
    "langchain-huggingface==0.0.3",
    "langchain-openai==0.1.20",
    "langchain-anthropic==0.1.23",
    "langchain-mistralai==0.1.13",
    "langchain-cohere==0.2.4",
    "langserve==0.2.3",
    "pymilvus==2.5.0",
    "langserve[all]",
    "pymilvus[model]",
    "fastapi[standard]",
    "fastapi==0.115.5",
    "uvicorn==0.32.1",
    "ragas==0.2.6",
    "nltk==3.9.1",
    "python-dotenv==1.0.1",
    "dacite==1.8.1",
    "chonkie==0.2.1.post1",
    "click>=8.1.7",
    "dask-cuda>=24.10.0",
    "cuda-python>=12.6.2",
    "ucx-py-cu12>=0.40.0",
    "pytest>=8.3.4",
    "httpx==0.27.2",
    "emoji==1.6.3",
]
requires-python = ">= 3.8"

[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"

[tool.rye]
managed = true
dev-dependencies = []

[tool.hatch.metadata]
allow-direct-references = true

[tool.hatch.build.targets.wheel]
packages = ["src"]

[[tool.rye.sources]]
name = "pytorch"
url = "https://download.pytorch.org/whl/cu124"