Skip to content

README.md

Latest commit

 

History

History
102 lines (80 loc) · 2.36 KB

README.md

File metadata and controls

102 lines (80 loc) · 2.36 KB

Prism - AI Portfolio Analysis (MVP)

Prism is an AI-driven fund portfolio analysis system for the Chinese market. It is designed for research and decision support, not for real-time trading.

What Works (MVP)

  1. Wizard-based portfolio setup:
    • Portfolio name
    • Fund codes and allocations
    • Risk preference and investment horizon
  2. Async analysis workflow:
    • POST /api/portfolios/{id}/analyze returns a task_id
    • Background job runs PrismKernel
    • Progress is pushed via WebSocket
  3. Dashboard consumption:
    • View portfolio details
    • Track task status (queued/running/completed/failed)
    • Read latest analysis result

Architecture

  1. Frontend (TypeScript + React + Vite):
    • Wizard
    • Dashboard
    • Task progress via WebSocket
  2. Backend (Python + FastAPI):
    • REST APIs for portfolios, analysis, task status
    • WebSocket endpoint for progress events
  3. Core Engine (Python):
    • PrismKernel with LangGraph orchestration
    • Fetchers for fund and market data
    • Analysis synthesis to structured recommendations
  4. Storage:
    • SQLite for MVP
    • PostgreSQL via async driver for production

Key Endpoints

  1. POST /api/portfolios
  2. GET /api/portfolios
  3. GET /api/portfolios/{id}
  4. POST /api/portfolios/{id}/analyze
  5. GET /api/tasks/{task_id}
  6. GET /api/portfolios/{id}/analysis
  7. WS /ws/tasks/{task_id}

Local Setup

# One-command start (backend + frontend)
./start.sh

# Fast start without reinstalling dependencies
./start.sh --skip-install
 
# Override ports/host
PRISM_BACKEND_PORT=8001 PRISM_FRONTEND_PORT=5174 ./start.sh

# Stop backend/frontend services
./stop.sh

# Smoke test the backend flow:
# health -> create portfolio -> analyze -> poll task -> get analysis
./smoke.sh

# Or run services separately:
# Install backend dependencies
uv sync

# Initialize env
cp .env.example .env

# Run backend
uv run uvicorn src.api.main:app --host 0.0.0.0 --port 8000 --reload
# Run frontend
cd frontend
npm install
npm run dev

Quality Checks

# Backend
ruff check src
python -m compileall src

# Frontend
cd frontend
npm run lint
npm run build

Notes

  1. This project does not execute trades and does not connect to brokerage APIs.
  2. Outputs are for research/decision support only and not investment advice.
  3. Black swan monitoring and periodic snapshot service are planned for next phase.