Add Oracle AI Database + LangChain.js RAG Assistant

[Ela689]

Summary

Adds a runnable TypeScript/JavaScript RAG assistant that demonstrates the LangChain.js integration with Oracle AI Database.

The sample shows how JavaScript and TypeScript developers can use @oracle/langchain-oracledb and OracleVS to store document chunks, embeddings, and JSON metadata in Oracle AI Database, then run retrieval and grounded answer generation from a browser-based developer assistant.

What is included

• Adds a full Node.js/TypeScript application, not a notebook.
• Includes a React + Vite frontend in apps/web.
• Includes a Hono API service in services/api.
• Includes shared LangChain.js RAG logic in packages/core.
• Includes Oracle connection and OracleVS helpers in packages/db.
• Seeds a markdown knowledge base from samples/knowledge-base.
• Documents local Oracle Database Free / Oracle AI Database and Autonomous AI Database connection options.
• Demonstrates:
  • OracleVS table creation and ingest
  • OracleVS.fromDocuments
  • similarity search with scores
  • metadata-filtered retrieval
  • MMR retrieval
  • optional HNSW vector index creation
  • grounded LLM answers with retrieved evidence
  • Oracle table, column, row count, vector index, and query vector inspection
  • representative vector-search SQL shown in the UI

Documentation

• README.md explains setup, seeding, local development, and Docker usage.
• docs/article.md provides the technical walkthrough with diagrams and screenshots.
• docs/local-development.md provides local troubleshooting notes.

Notes

This contribution is intentionally a runnable TypeScript/JavaScript application rather than a Python notebook. The goal is to show the end-to-end LangChain.js integration path: React UI, Node.js API, LangChain.js retrieval workflow, OracleVS vector store, Oracle AI Database storage/search, and grounded answers.