import os
import subprocess

# Set environment variables
os.environ['OPENAI_API_KEY'] = 'sk-55f6e57f1d834b0e93ceaf98cc2cb715'
os.environ['DEEPSEEK_API_KEY'] = 'sk-55f6e57f1d834b0e93ceaf98cc2cb715'
os.environ['PYTHONIOENCODING'] = 'utf-8'
os.environ['OLLAMA_EMBEDDING_MODEL'] = 'snowflake-arctic-embed:latest'
os.environ['OLLAMA_RERANKER_MODEL'] = 'jina-reranker-v2:latest'  # Updated to v2 model
os.environ['OPENAI_API_MODEL'] = 'deepseek-chat'
os.environ['OPENAI_API_BASE'] = 'https://api.deepseek.com/v1'
os.environ['LLM_BINDING_HOST'] = 'https://api.deepseek.com/v1'
# Ollama rerank configuration - using local Ollama server
os.environ['RERANK_BINDING_HOST'] = 'http://localhost:11434'  # Local Ollama server
os.environ['RERANK_BINDING_API_KEY'] = ''  # No API key needed for local Ollama

# Set database environment variables
os.environ['REDIS_URI'] = 'redis://localhost:6379'
os.environ['NEO4J_URI'] = 'bolt://localhost:7687'
os.environ['NEO4J_USERNAME'] = 'neo4j'
os.environ['NEO4J_PASSWORD'] = 'jleu1212'
os.environ['QDRANT_URI'] = 'http://localhost:6333/'
os.environ['POSTGRES_URI'] = 'postgresql://jleu3482:jleu1212@localhost:5432/rag_anything'

# Command to start the LightRAG server with entity extraction disabled
cmd = [
    'python', '-m', 'lightrag.api.lightrag_server',
    '--port', '3015',
    '--working-dir', 'rag_storage',
    '--input-dir', 'inputs',
    '--key', 'jleu1212',
    '--auto-scan-at-startup',
    '--llm-binding', 'openai',
    '--embedding-binding', 'ollama',
    '--rerank-binding', 'ollama',  # Changed from 'jina' to 'ollama' for local Ollama rerank
    '--summary-max-tokens', '0',  # Disable entity extraction by setting summary tokens to 0
    '--timeout', '600'            # Increase server timeout to 600 seconds to avoid nginx 504
]

# Run the server
subprocess.run(cmd)