pyproject.toml

[project]
name = "athena-research-assistant"
version = "0.1.0"
description = "Multi-agent research copilot with deterministic citation verification"
readme = "README.md"
requires-python = ">=3.10"
license = { text = "MIT" }
authors = [{ name = "Junye Zhao", email = "junyezhao9@gmail.com" }]

# Core runtime: enough to run retrieval, the pipeline, citation validation,
# the local PDF RAG (hashing backend), and the Streamlit demo.
dependencies = [
    "pydantic>=2.0",
    "pydantic-settings>=2.0",
    "python-dotenv>=1.0",
    "langgraph>=0.2.0",
    "langgraph-checkpoint-sqlite>=2.0.0",
    "langchain-core>=0.3.0",
    "openai>=1.0",
    "arxiv>=2.1",
    "requests>=2.31",
    "rapidfuzz>=3.0",
    "diskcache>=5.6",
    "numpy>=1.24",
    "pymupdf>=1.24",
    "streamlit>=1.28",
]

[project.optional-dependencies]
# Higher-quality semantic embeddings + FAISS acceleration for the PDF RAG module.
# The default hashing backend works without these (offline, deterministic).
rag = [
    "faiss-cpu>=1.8",
    "sentence-transformers>=3.0",
]
# RQ1/RQ2/RQ3 evaluation experiments (figures + statistics).
eval = [
    "matplotlib>=3.8",
    "scipy>=1.11",
]
dev = [
    "pytest>=8.0",
    "ruff>=0.6",
]

[tool.pytest.ini_options]
testpaths = ["tests"]
pythonpath = ["."]
filterwarnings = [
    "ignore::UserWarning",
]

[tool.ruff]
line-length = 100
target-version = "py310"
extend-exclude = [".vendor", ".venv", ".venv-eval"]

[tool.ruff.lint]
select = ["E", "F", "I"]
# Line length is enforced by the formatter; E501 would otherwise flag long
# LLM prompt string literals that must not be reflowed (it changes the prompt).
ignore = ["E501"]

[tool.ruff.lint.per-file-ignores]
# CLI entry points inject the repo root into sys.path before importing athena.*
"scripts/*" = ["E402"]