Flask-based backend for QuantDinger: market data, indicators, AI analysis, backtesting, and a strategy runtime with multi-user support.

• Multi-market data layer: factory-based providers (crypto / US stocks / forex / futures, etc.)
• Indicators + backtesting: persisted runs/history in PostgreSQL
• AI multi-agent analysis: optional web search + OpenRouter LLM integration
• Strategy runtime: thread-based executor, with optional auto-restore on startup
• Pending orders worker (optional): polls queued orders and dispatches signals (webhook/notifications)
• Multi-user authentication: role-based access control (admin/manager/user/viewer)
• User management: admin can create/edit/delete users and reset passwords

backend_api_python/
|-- app/
|   |-- __init__.py                 # Flask app factory + startup hooks
|   |-- config/                     # Settings (env-driven)
|   |-- data_sources/               # Data sources + factory
|   |-- routes/                     # REST endpoints
|   |-- services/                   # Analysis, agents, strategies, search, user_service
|   |-- utils/                      # PostgreSQL helpers, config loader, logging, HTTP utils
|-- migrations/
|   |-- init.sql                    # PostgreSQL schema initialization
|-- env.example                     # Copy to .env for local config
|-- requirements.txt
|-- run.py                          # Entrypoint (loads .env, applies proxy env, starts Flask)
|-- gunicorn_config.py              # Optional production config
|-- README.md

• Backend module boundaries: ../docs/MODULE_BOUNDARIES.md
• Concurrency model: ../docs/CONCURRENCY_MODEL.md
• Canonical live-trading venue matrix: app/services/live_trading/capabilities.py
• Stable live order contracts: app/services/live_trading/contracts.py
• Structural regression guard:

python scripts/backend_quality_check.py

Exchange integrations should add offline fixtures and pass:

python scripts/exchange_smoke_test.py --offline-contracts

Create .env file in project root:

# Database
POSTGRES_USER=quantdinger
POSTGRES_PASSWORD=your_secure_password
POSTGRES_DB=quantdinger

# Admin account (created on first startup)
ADMIN_USER=admin
ADMIN_PASSWORD=your_admin_password

# Optional
OPENROUTER_API_KEY=your_api_key

AtlasCloud is supported as an OpenAI-compatible LLM provider. See the official AtlasCloud LLM API docs and API key guide, then set:

LLM_PROVIDER=atlascloud
ATLASCLOUD_API_KEY=your_api_key
ATLASCLOUD_MODEL=deepseek-v3
ATLASCLOUD_BASE_URL=https://api.atlascloud.ai/v1

Release builds inject the backend app version from the Git tag (v3.0.23 -> 3.0.23). Local source runs fall back to git describe and then the repo-root VERSION file; local Docker builds can override with APP_VERSION.

docker-compose up -d

This will:

• Start PostgreSQL database (port 5432)
• Initialize database schema automatically
• Start backend API (port 5000)
• Start frontend (port 8888)
• Create admin user from ADMIN_USER/ADMIN_PASSWORD

• Frontend: http://localhost:8888
• Backend API: http://localhost:5000
• Login with your configured admin credentials

• Python 3.10+ recommended
• PostgreSQL 14+ installed and running

# Create database and user
sudo -u postgres psql
CREATE DATABASE quantdinger;
CREATE USER quantdinger WITH ENCRYPTED PASSWORD 'your_password';
GRANT ALL PRIVILEGES ON DATABASE quantdinger TO quantdinger;
\q

# Initialize schema
psql -U quantdinger -d quantdinger -f migrations/init.sql

cd backend_api_python
pip install -r requirements.txt

Windows (CMD):

copy env.example .env

Windows (PowerShell):

Copy-Item env.example .env

Then edit .env and set:

# Required
DATABASE_URL=postgresql://quantdinger:your_password@localhost:5432/quantdinger
SECRET_KEY=your-secret-key-change-me
ADMIN_USER=admin
ADMIN_PASSWORD=your_admin_password

# Optional but recommended
OPENROUTER_API_KEY=your_api_key

For AtlasCloud instead, use:

LLM_PROVIDER=atlascloud
ATLASCLOUD_API_KEY=your_api_key
ATLASCLOUD_MODEL=deepseek-v3
ATLASCLOUD_BASE_URL=https://api.atlascloud.ai/v1

python run.py

Default address: http://localhost:5000

• Connection: configured via DATABASE_URL environment variable
• Schema: initialized via migrations/init.sql
• Tables are managed with foreign key constraints and indexes for performance
• User data isolation via user_id column in relevant tables

POST /api/user/login      - User login
POST /api/user/logout     - User logout
GET  /api/user/info       - Get current user info

GET    /api/users/list           - List all users
POST   /api/users/create         - Create user
PUT    /api/users/update?id=     - Update user
DELETE /api/users/delete?id=     - Delete user
POST   /api/users/reset-password - Reset password

GET  /api/users/profile         - Get own profile
PUT  /api/users/profile/update  - Update own profile
POST /api/users/change-password - Change own password

GET  /api/health
GET  /api/indicator/kline
GET  /api/global-market/adanos-sentiment?tickers=AAPL,TSLA
POST /api/fast-analysis/analyze    - Fast AI analysis (main entry)
GET  /api/fast-analysis/history    - Analysis history
GET  /api/fast-analysis/similar-patterns - RAG similar patterns
POST /api/fast-analysis/feedback   - User feedback on analysis

Set ADANOS_API_KEY to enable optional US stock sentiment enrichment from the Adanos Market Sentiment API. If the key is not configured, the endpoint returns enabled=false and the rest of QuantDinger continues to work normally.

ADANOS_API_KEY=your_adanos_key
ADANOS_SENTIMENT_SOURCE=reddit  # reddit, x, news, or polymarket

Example:

GET /api/global-market/adanos-sentiment?tickers=AAPL,TSLA&source=reddit&days=7

The response normalizes common compare fields across sources, including sentiment_score, buzz_score, bullish_pct, bearish_pct, mentions, trend, trend_history, and source-specific activity metrics such as subreddit_count, unique_tweets, source_count, trade_count, market_count, and total_liquidity.

The global-market economic calendar is free-first. By default it uses the no-key AkShare/WallstreetCN calendar fallback. If you configure TRADING_ECONOMICS_CLIENT and TRADING_ECONOMICS_KEY, QuantDinger will try Trading Economics as the official international calendar provider before falling back. Finnhub paid-only calendar and social-sentiment endpoints are skipped by default through FINNHUB_FREE_ONLY=true.

Only set FINNHUB_FREE_ONLY=false if your Finnhub plan explicitly includes those paid endpoints.

Uses FastAnalysisService (single LLM call, multi-factor):

• Memory: qd_analysis_memory in PostgreSQL
• API: POST /api/fast-analysis/analyze (main), /history, /similar-patterns, /feedback
• Calibration: AICalibrationService tunes BUY/SELL thresholds from validated outcomes

For local Vue dev (private frontend repo):

• Frontend dev server URL depends on your vue.config.js / Vite config
• Backend: http://localhost:5000
• Point devServer proxy at the backend URL above

gunicorn -c gunicorn_config.py "run:app"

• Database connection failed: Check DATABASE_URL format and PostgreSQL service status
• Outbound requests fail: Configure PROXY_URL in .env
• Disable auto-restore: Set DISABLE_RESTORE_RUNNING_STRATEGIES=true
• Disable pending-order worker: Set ENABLE_PENDING_ORDER_WORKER=false

Apache License 2.0. See repository root LICENSE.