# TheDataGuy Chat Configuration
# Copy this file to .env and fill in your values

# OpenAI API Key - Required for LLM and embeddings
OPENAI_API_KEY=your_openai_api_key_here

# Vector Store Configuration
VECTOR_STORAGE_PATH=./db/vector_store_tdg
QDRANT_COLLECTION=thedataguy_documents

# Model Configuration
EMBEDDING_MODEL=Snowflake/snowflake-arctic-embed-l
LLM_MODEL=gpt-4o-mini
LLM_TEMPERATURE=0

# For evaluation and synthetic data generation (optional)
SDG_LLM_MODEL=gpt-4.1
EVAL_LLM_MODEL=gpt-4.1

# Blog Configuration
DATA_DIR=data/
BLOG_BASE_URL=https://thedataguy.pro/blog/

# Search Configuration
MAX_SEARCH_RESULTS=5

# Document Chunking Configuration
CHUNK_SIZE=1000
CHUNK_OVERLAP=200

# Vector Database Creation Configuration
FORCE_RECREATE=False
OUTPUT_DIR=./stats
USE_CHUNKING=True
SHOULD_SAVE_STATS=True