Create Chat API Endpoint for RAG Implementation

## Develop the main chat endpoint that will handle user queries using the RAG (Retrieval Augmented Generation) system.

Tasks:
- Create ChatMessage model to store chat history
- Implement `/api/chat/` endpoint for handling chat requests
- Develop vector similarity search using Pinecone
- Integrate Cohere?/Deepseek? for generating responses
- Implement prompt engineering for better responses
- Add context window management
- Implement error handling and rate limiting
- Add response streaming capability

Start by locating the existing Django/DRF API structure and any current model, embedding, or vector-search code; the payload names no files or tests. Define the ChatMessage shape and endpoint contract, then verify retrieval, response generation, context limits, errors, rate limiting, and streaming against a testable specification.

Written by the indexing model from the issue text.