segni49/OkooAI-Tourism_Assistant-

OkooAI is a Retrieval-Augmented Generation (RAG) system built with LangChain, FastAPI, and ChromaDB. It answers tourism-related questions about Ethiopia using real documents, strict fallback logic, and adaptive routing. # 🧠 OkooAI — Tourism Assistant Powered by RAG OkooAI is a Retrieval-Augmented Generation (RAG) system built with LangChain, FastAPI, and ChromaDB. It answers tourism-related questions about Ethiopia using real documents, strict fallback logic, and adaptive routing. ## 🚀 Features - ✅ Naive + Adaptive RAG pipeline - ✅ LangGraph-based intent routing - ✅ Strict fallback: no hallucinations - ✅ Modular nodes for planning, comparison, exploration - ✅ PDF ingestion and chunking - ✅ Ollama-powered local LLM (Qwen 0.5b) - ✅ FastAPI backend with clean endpoints ## 🧱 Architecture ```bash User Query │ ▼ Intent Classifier ──► LangGraph Router │ │ ▼ ▼ Retriever ┌─────────────┐ │ │ Nodes: │ ▼ │ - ask_fact │ LLM + Prompt │ - plan_trip │ │ │ - compare │ ▼ │ - explore │ Reflection Node ◄───┘ │ ▼ Final Answer + Source Chunks ```bash ## 📂 Project Structure ```bash advanced_rag_ai/ ├── api/ │ ├── main.py # FastAPI entry point │ ├── tourism_graph.py # LangGraph workflow │ ├── planner_node.py # Trip planner logic │ ├── hotel_comparison_node.py │ ├── explore_place_node.py │ ├── intent_classifier.py │ ├── adaptive_retriever.py │ └── self_reflective_rag.py ├── data/ # Indexed tourism PDFs ├── requirements.txt ├── Dockerfile └── README.md ``` ## 🧪 Demo Instructions ### 1. Start Ollama ```bash ollama run qwen:0.5b ``` ### 2. Run the API ```bash uvicorn api.main:app --reload ``` ### 3. Test the Chat Endpoint ```bash curl -X POST localhost \ -H "Content-Type: application/json" \ -d '{"session_id": "demo", "question": "Plan a trip to Gondar", "model": "qwen:0.5b"}' ``` ### 4. Upload a PDF ``` bash curl -X POST localhost \ -F "file=@data/03_Gondar_Bahir_Dar_Lake_Tana_Blue_Nile.pdf" ``` ## 📦 Deployment Use Docker for production: ```bash docker build -t oko …