23AI Query Assistant
0
Converts natural language questions into SQL or semantic search queries over data stored in Oracle Autonomous Database.
About
This proof-of-concept system transforms natural language questions into executable SQL or semantic vector search queries for data residing in Oracle Autonomous Database (ADB). It leverages Oracle 23ai's native vector search capabilities alongside a locally hosted LLaMA 3.2 LLM via Ollama to provide intelligent, unified access to both structured and unstructured data, eliminating the need for an external vector database.
Key Features
- Locally hosted LLaMA 3.2 LLM (via Ollama) interprets natural language and generates appropriate queries.
- FastAPI server acts as an intermediary, managing LLM prompts, query execution, and result formatting.
- Stores all vector data and performs queries directly within Oracle Autonomous Database.
- Intelligently decides between traditional SQL and semantic vector queries based on the user's question.
- Utilizes Oracle 23ai's built-in vector search and embedding models for document and query embeddings.
- 0 GitHub stars
Use Cases
- Developing natural language interfaces for Oracle Autonomous Database.
- Enabling semantic search capabilities over unstructured text within Oracle databases.
- Building AI-powered data retrieval systems for structured and unstructured enterprise data.
Sponsored