sachink1729/legal-cases-search-using-self-query-qdrant-llama3-langchain
📦 Open Source Projectsachink1729
A RAG-based legal search engine using Qdrant, Llama 3, and LangChain for advanced metadata-filtered querying.
This project serves as a comprehensive guide and implementation for building a sophisticated legal document retrieval system. At its core, it utilizes the Self-Query Retriever pattern, which enables the system to translate natural language queries into structured metadata filters alongside vector similarity searches. By integrating Qdrant's efficient vector storage with Llama 3's reasoning capabilities, the system can accurately parse legal queries to filter by specific case attributes like jurisdiction, date, or case type. The repository includes Jupyter Notebooks that walk through the entire pipeline: from document ingestion and embedding generation to the configuration of the self-querying mechanism. It is an excellent resource for developers looking to bridge the gap between unstructured legal text and structured database querying using modern GenAI stacks.
💡Highlights
- ├─Self-querying metadata filtering
- ├─Qdrant vector database integration
- └─Llama 3 powered query parsing
🎯For
- ├─LegalTech Developers
- ├─AI Engineers
- └─Data Scientists