stage4(sagemaker-deployment): Chat Persistence, History, Eviction, Delete Endpoint, Tests & Demo Description

[shreyashx16]

Description - Fixes #6

Summary
This PR completes the Stage 4 features by adding persistent chat session management, including session history, TTL-based eviction, and a safe delete endpoint. It also adds the necessary tests, a small demo script, and updates to the README to ensure the app can be easily run and demoed locally. This work is hackathon-ready and fully tested.

Key Changes

• Persistence

  • Added chat session management functions: init_chat_db, save_session, load_session, delete_session, and list_sessions.
  • Chat session persistence is SQLite-based, and the database path can be dynamically set via the CHAT_DB environment variable.

• Chat API & Session Management

  • POST /chat: Creates or continues a session. Saves sessions after each message.
  • GET /chat/{session_id}/history: Retrieves chat history for a session (in-memory or persisted).
  • DELETE /chat/{session_id}: Deletes a session from both memory and persistence.
  • Eviction: Periodic background eviction of expired sessions, triggered based on CHAT_SESSION_TTL_SECONDS and CHAT_EVICT_INTERVAL_SECONDS.

• Tests

  • Added tests for chat history retrieval, session persistence, delete functionality, and TTL-based eviction.

• Docs & Demo

  • Updated README with instructions on setting up environment variables and running the demo.
  • Added demo_demo.py to simulate upload, analysis, chat, and history operations.

How to Run Locally

1. Install Dependencies

   python -m venv .venv
   pip install -r requirements.txt

2. Run Tests

   $env:PYTHONPATH='.'; pytest -q

3. Start the Server

   uvicorn app.app:app --reload --host 0.0.0.0 --port 8000

4. Run the Demo

   python demo_demo.py

5. Access Interactive API Docs
   http://127.0.0.1:8000/docs

Environment Variables

• CHAT_DB — Path to SQLite DB for chat sessions (default: sessions.db).
• CHAT_SESSION_TTL_SECONDS — Session TTL in seconds (default: 3600). Set <= 0 to disable expiry.
• CHAT_CONTEXT_TURNS — Number of turns to include in the LLM prompt (default: 10).
• CHAT_EVICT_INTERVAL_SECONDS — Background eviction interval (default: 60).

Next Steps

• Add AWS client wrappers for SageMaker integration.
• Teach KnowledgeStore to use SageMaker embeddings with fallback options.
• Add a CI job to mock AWS/SageMaker interactions.
• Optional: Implement Redis-backed session store (for multi-instance environments).

PR Checklist:

• Added chat session persistence with SQLite.
• Implemented /chat/{session_id}/history and /chat/{session_id}/delete endpoints.
• Added background session eviction (TTL-based).
• Added tests for chat functionality (history, persistence, TTL eviction, delete).
• Updated README with setup instructions and environment variables.
• Added demo script for showcasing functionality.

[github-actions]

CodeGuardian scan — no results found

[github-actions]

CodeGuardian scan — no results found

[github-actions]

CodeGuardian scan — no results found

[github-actions]

CodeGuardian scan — no results found