add chat feature and healthomics backend

[nbulsara]

Problem Statement: The NGS360 API Server needed AI chatbot capabilities to allow users to query genomics data through natural language, plus support for AWS HealthOmics workflow execution as an alternative to AWS Batch.

What Was Done:

• Added a complete AI chat feature using AWS Bedrock (Claude 3.5 Sonnet) with Strands Agent framework and MCP (Model Context Protocol) server exposing ~25 NGS360 REST API endpoints as tools
• Implemented streaming and non-streaming chat endpoints with JWT-based authentication, conversation persistence, and error handling
• Added AWS HealthOmics backend support for demultiplexing workflows alongside existing AWS Batch integration
• Updated job tracking to use UUIDs instead of AWS job IDs as primary keys, added viewed flag, and created separate endpoints for finding jobs by AWS ID
• Reduced Gunicorn workers from 4 to 2 and added graceful timeout configuration
• Added comprehensive test coverage for chat, MCP server, and job management features

Next Steps: The chat feature is production-ready with dependencies added (bedrock-agentcore, mcp, strands-agents, httpx) and can be deployed once environment variables are configured (BEDROCK_MODEL_ID, BEDROCK_REGION, AGENTCORE_MEMORY_ID, CHAT_API_BASE_URL, CHAT_TIMEOUT_SECONDS).