env-example.txt

# ========== Pinecone Vector Database ==========
PINECONE_API_KEY=<your-key-goes-here>
PINECONE_ENVIRONMENT=gcp-starter
PINECONE_INDEX_NAME=edith-notes
VECTOR_DIMENSION=384

# ========== LLaMA Model Settings ==========
# Path to your local LLaMA model file (.gguf format recommended)
# Using LLaMA 3.1 8B Instruct Q4_K_M quantization
LLAMA_MODEL_PATH=llama3.1:8b-instruct-q4_K_M
LLAMA_MODEL_TYPE=gguf
USE_GPU=true
MAX_TOKENS=4096
TEMPERATURE=0.7

# ========== Embedding Model ==========
# Default uses a small, efficient sentence transformer
EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2
EMBEDDING_DEVICE=cuda  # or 'cpu' if no GPU

# ========== Document Processing ==========
NOTES_DIRECTORY=./src/data/notes
USE_OCR=true
CHUNK_SIZE=1000
CHUNK_OVERLAP=200

# ========== RAG Settings ==========
TOP_K_RESULTS=5
SIMILARITY_THRESHOLD=0.3

# ========== Summary Settings ==========
SUMMARY_MAX_LENGTH=500
SUMMARY_MIN_LENGTH=100
SUMMARY_STYLE=comprehensive  # comprehensive, bullet, or brief

# ========== Application Settings ==========
LOG_LEVEL=INFO
BATCH_SIZE=10
MAX_NOTES_TO_PROCESS=100