# Core dependencies for Valori vector database
numpy>=1.19.0
scikit-learn>=0.24.0

# Document parsing dependencies
PyPDF2>=3.0.0
python-docx>=0.8.11
openpyxl>=3.1.0
python-pptx>=0.6.21
chardet>=5.0.0

# Text processing
nltk>=3.8
beautifulsoup4>=4.11.0

# Embedding models
sentence-transformers>=2.2.0

# Optional dependencies for better performance
# Uncomment if you need these features:
# faiss-cpu>=1.7.0  # For high-performance vector search
# hnswlib>=0.6.0    # For HNSW indexing
annoy>=1.17.0     # For approximate nearest neighbor search
# docling>=1.0.0    # Microsoft's Docling for advanced parsing
# llama-parse>=0.4.0  # LlamaIndex's LlamaParse for AI parsing

# Path utilities for older Python versions
pathlib2>=2.3.0; python_version<'3.4'
