pyproject.toml

[project]
name = "ed-schema"
version = "0.1.0"
description = "AI-Enabled Schema Mapping for LIF - comparing embedding approaches vs LLMs"
readme = "README.md"
requires-python = ">=3.12"
dependencies = [
    # Tier 1: Embedding/NLP
    "rapidfuzz>=3.0", # Fuzzy string matching
    "sentence-transformers>=3.0", # Sentence embeddings (all-MiniLM, bge-large)
    "model2vec>=0.3", # Static embeddings (potion-base-32M)
    # LLM API clients
    "anthropic>=0.76", # Claude models (latest: 0.76.0)
    "google-genai>=1.57", # Gemini models (google-generativeai deprecated Aug 2025)
    "openai>=2.15", # OpenAI API (latest: 2.15.0)
    "groq>=0.33", # Groq inference for OSS models (latest: 0.33.0)
    # Data handling
    "pandas>=2.0",
    "numpy>=1.26",
    # ML utilities
    "scikit-learn>=1.5", # Metrics, cosine similarity
    # Experiment tracking & observability
    "wandb>=0.18",
    "langfuse>=2.0", # LLM observability & tracing
    # Visualization
    "matplotlib>=3.8",
    "seaborn>=0.13",
    # Validation & utilities
    "pydantic>=2.0", # Structured output validation
    "python-dotenv>=1.0", # Environment variables
    "tqdm>=4.66", # Progress bars
    "questionary>=2.1.1",
    "jsonata-python>=0.6.1",
]

[project.optional-dependencies]
dev = [
    "pytest>=8.0",
    "pytest-timeout>=2.0",
    "ipykernel>=6.0",           # Jupyter kernel
    "jupyter>=1.0",
]

[tool.pytest.ini_options]
testpaths = ["tests"]
markers = [
    "integration: tests that make real API calls (require API keys)",
]
filterwarnings = [
    "ignore::DeprecationWarning",
]

[dependency-groups]
dev = [
    "python-docx>=1.2.0",
]

[project.scripts]
test = "pytest:main"