gavindiaz/DinQuant
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: Multi-user system with PostgreSQL - WIP temporary save

Browse Source
This commit is contained in:
TIANHE
2026-01-14 05:29:55 +08:00
parent 996e3b38fe
commit 61a5e5e6aa
68 changed files with 91057 additions and 1920 deletions
Download Patch File Download Diff File Expand all files Collapse all files
backend_api_python/README.md
+137 -56
View File
@@ -1,17 +1,16 @@
# QuantDinger Python API (backend)
Flask-based local-first backend for QuantDinger: market data, indicators, AI analysis, backtesting, and a strategy runtime (with an optional pending-order worker).
This repository is intentionally simple: **no external database is required by default**. Data is stored in a local SQLite file (`quantdinger.db`) created/updated automatically on startup.
Flask-based backend for QuantDinger: market data, indicators, AI analysis, backtesting, and a strategy runtime with multi-user support.
## What you get
- **Multi-market data layer**: factory-based providers (crypto / US stocks / CN&HK stocks / futures, etc.)
- **Indicators + backtesting**: persisted runs/history in SQLite
- **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)
- **Local auth (single-user)**: `/login` with env-configured admin credentials (JWT)
- **Multi-user authentication**: role-based access control (admin/manager/user/viewer)
- **User management**: admin can create/edit/delete users and reset passwords
## Project layout
@@ -22,8 +21,10 @@ backend_api_python/
│ ├─ config/ # Settings (env-driven)
│ ├─ data_sources/ # Data sources + factory
│ ├─ routes/ # REST endpoints
│ ├─ services/ # Analysis, agents, strategies, search, ...
│ └─ utils/ # SQLite helpers, config loader, logging, HTTP utils
│ ├─ 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)
@@ -31,46 +32,99 @@ backend_api_python/
└─ README.md
```
## Quick start (local development)
## Quick start (Docker - Recommended)
### 1) Configure environment
Create `.env` file in project root:
```bash
# 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
```
### 2) Start services
```bash
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`
### 3) Access the system
- Frontend: `http://localhost:8888`
- Backend API: `http://localhost:5000`
- Login with your configured admin credentials
## Quick start (Local Development)
### Prerequisites
- Python 3.10+ recommended
- PostgreSQL 14+ installed and running
### 1) Install dependencies
### 1) Setup PostgreSQL
```bash
# 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
```
### 2) Install dependencies
```bash
cd backend_api_python
pip install -r requirements.txt
```
### 2) Create your local `.env`
### 3) Create your local `.env`
Windows (CMD):
```bash
copy env.example .env
```
Windows (PowerShell):
```bash
Copy-Item env.example .env
```
Then edit `.env` and set at least:
Then edit `.env` and set:
- `SECRET_KEY`
- `ADMIN_USER`
- `ADMIN_PASSWORD`
```bash
# 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 common:
# Optional but recommended
OPENROUTER_API_KEY=your_api_key
```
- `OPENROUTER_API_KEY` (for AI analysis)
- `FINNHUB_API_KEY` / `SEARCH_GOOGLE_*` / `SEARCH_BING_API_KEY` (for richer data/search)
- `PROXY_PORT` or `PROXY_URL` (if your network blocks some providers)
### 3) Start the API server
### 4) Start the API server
```bash
python run.py
@@ -78,47 +132,74 @@ python run.py
Default address: `http://localhost:5000`
## Database (SQLite)
## Database (PostgreSQL)
- Default file: `backend_api_python/data/quantdinger.db` (override via `SQLITE_DATABASE_FILE`)
- Tables are created/updated automatically on startup (see `app/utils/db.py`)
- `qd_addon_config` exists for backward compatibility, but **this backend reads secrets from `.env` / OS env**, not from the database (see `app/utils/config_loader.py`)
- 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
## AI memory augmentation (local-only)
## User Roles & Permissions
This backend includes a lightweight, privacy-first **memory-augmented multi-agent** system:
| Role | Permissions |
|------|-------------|
| admin | Full access + user management |
| manager | Strategy, backtest, portfolio, settings |
| user | Strategy, backtest, portfolio (own data) |
| viewer | Dashboard view only |
- Memory DBs (per role): `backend_api_python/data/memory/*_memory.db`
- Reflection DB (optional auto-verify loop): `backend_api_python/data/memory/reflection_records.db`
- API hooks:
- `POST /api/analysis/multi` (main entry)
- `POST /api/analysis/reflect` (manual learn from post-trade outcomes)
- Controls: see `.env` / `env.example`:
- `ENABLE_AGENT_MEMORY`, `AGENT_MEMORY_*`
- `ENABLE_REFLECTION_WORKER`, `REFLECTION_WORKER_INTERVAL_SEC`
## Frontend integration (Vue dev server)
The Vue dev server proxies `/api/*` to this backend by default:
- Frontend: `http://localhost:8000`
- Backend: `http://localhost:5000`
Proxy config: `quantdinger_vue/vue.config.js`
## Useful endpoints
## API Endpoints
### Authentication
```text
GET /health
POST /login
GET /info
POST /api/user/login - User login
POST /api/user/logout - User logout
GET /api/user/info - Get current user info
```
### User Management (Admin only)
```text
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
```
### Self-Service
```text
GET /api/users/profile - Get own profile
PUT /api/users/profile/update - Update own profile
POST /api/users/change-password - Change own password
```
### Other Endpoints
```text
GET /api/health
GET /api/indicator/kline
POST /api/analysis/multi
```
## Production (optional)
## AI memory augmentation
Gunicorn example:
This backend includes a lightweight, privacy-first **memory-augmented multi-agent** system:
- Memory DBs stored in PostgreSQL
- API hooks:
- `POST /api/analysis/multi` (main entry)
- `POST /api/analysis/reflect` (manual learn from post-trade outcomes)
- Controls in `.env`:
- `ENABLE_AGENT_MEMORY`, `AGENT_MEMORY_*`
- `ENABLE_REFLECTION_WORKER`, `REFLECTION_WORKER_INTERVAL_SEC`
## Frontend integration
For Vue dev server:
- Frontend: `http://localhost:8000`
- Backend: `http://localhost:5000`
- Proxy config: `quantdinger_vue/vue.config.js`
## Production (Gunicorn)
```bash
gunicorn -c gunicorn_config.py "run:app"
@@ -126,11 +207,11 @@ gunicorn -c gunicorn_config.py "run:app"
## Troubleshooting
- If outbound data/search requests fail, configure `PROXY_PORT` (or `PROXY_URL`) in `.env`.
- If you don’t want strategies to auto-restore on startup, set `DISABLE_RESTORE_RUNNING_STRATEGIES=true`.
- If you don’t want the pending-order worker, set `ENABLE_PENDING_ORDER_WORKER=false`.
- **Database connection failed**: Check `DATABASE_URL` format and PostgreSQL service status
- **Outbound requests fail**: Configure `PROXY_PORT` or `PROXY_URL` in `.env`
- **Disable auto-restore**: Set `DISABLE_RESTORE_RUNNING_STRATEGIES=true`
- **Disable pending-order worker**: Set `ENABLE_PENDING_ORDER_WORKER=false`
## License
Apache License 2.0. See repository root `LICENSE`.