- **Dynamic RAG System:** Transforms user questions into SQL queries using a retrieval-augmented generation (RAG) approach.
- **PostgreSQL Backend:** Connects to a test PostgreSQL database with 5 tables (each with 5 columns and over 1000 rows), optimized for retrieval efficiency with indexes and tailored queries.
- **Offline LLM Integration:** Utilizes an offline Ollama LLM (with quantization/optimization) for generating precise SQL queries and refining user input.
- **Intelligent Agent Workflow:
- **User Typo Correction:** Automatically corrects typos in user messages to ensure query clarity.
- **Query Routing:** Classifies input as either a RAG query or a general user query.
- **RAG Node:** Generates and executes the SQL query against the database, then formats the result into a user-friendly response.
- **User-Friendly Node:** Provides conversational responses when the query does not require database retrieval.
- **Graph-Based Orchestration:** Leverages a modular state graph to seamlessly manage and transition between different processing nodes for efficient query handling.
| اسم المستقل | Omar S. |
| عدد الإعجابات | 0 |
| عدد المشاهدات | 2 |
| تاريخ الإضافة |
