How aki-008/evaluation-ai-hybrid-chat works
File structure
• Top-level Python scripts:
• config.py – central configuration (credentials, endpoints)
• load_to_neo4j.py – builds the graph database from the JSON dataset
• pinecone_upload.py – builds the vector index in Pinecone
• hybrid_chat.py – the core chat interface combining graph + vector retrieval
• visualize_graph.py & neo4j_viz.html – graph visualization tools
• requirements.txt – dependencies (e.g., neo4j, pinecone-client, OpenAI)
• vietnam_travel_dataset.json – source data for ingestion
• AI_hybrid_chat_evaluation.pdf – likely design or evaluation report
• No frontend vs. backend, but distinct components: data ingestion, storage layers, chat logic, visualization.
README for additional insights
• README only says “first commit,” so the architecture must be inferred from file names.