TvTrader fullstack

Description

This is a full stack Sanic+Vue application for easily collect and process TradingView alerts and possibly execute custom trading strategies. Includes a modular alert distribution layer in the agents directory.

Also a good candidate to learn more about test automation. Pytest-BDD is used for test automation, see the tests folder for examples.

Do not use in production as it is created for learning and personal use only. No authentication is implemented yet.

Overview

[C4_diagram_here]

@startuml C4_Elements
!include  <C4/C4_Container>

LAYOUT_WITH_LEGEND()

AddElementTag("dev", $fontColor="yellow")
AddRelTag("future", $textColor="lightgrey", $lineColor="grey", $lineStyle = DashedLine())
Person(user, "Users")

System_Ext(tradingview, "TradingView", "Data source")
Container(tvtrader_frontend, "Frontend", "VueJs", "Alert dashboard")

System_Boundary(data_backend, "Data Backend") {
  Container(fetcher, "Fetcher", "REST, NestJs", "Allows users to dinamycally fetch price data from TradingView")
  SystemDb(influx, "Store", "InfluxDB, FluxScript")
  Container(tvtrader_backend, "Backend", "WS, REST", "API backend for the dashboard")
  Container(tvtrader_distributor, "Distributor", "Python", "Receiver and distributor for the alerts")
}
SystemQueue(alertq, "Alert Queue")

System_Boundary(dist_targets, "Distribution Targets") {
  Person(sysdev, "Target developer", $tags="dev")
  Container(carbon, "CarbonDB", "Optional target for the alerts")
  Container(grafana, "Grafana", "Optional target for the alerts")
  Container(logging, "Logging", "Optional loogging for the alerts")
  Container(opt_target, "...", "Other types of distribution targets")
  Container_Ext(future_frontend, "Future Frontend", "Vue3+Quasar")
  Rel_D(sysdev, opt_target, "Implements distribution target")
  Rel_U(sysdev, carbon, "Implements distribution target")
  Rel_L(sysdev, grafana, "Implements distribution target")
  Rel_R(sysdev, logging, "Implements distribution target")
  Rel_U(sysdev, future_frontend, "Implements features", $tags="future")
}

System_Boundary(strategies, "Strategy evaluators") {
  Container(strategy_1, "Jupyter", "NumPy, Pandas", "Evaluator for the strategy")
  Container(opt_strategy, "Others", "Various technologies", "Additional evaluators")
  Person(quant, "Quantitative Developer", $tags="dev")
  System_Boundary(browser, "Browser") {
    Container_Ext(tvui, "TradingView UI", "User interface for TradingView")
    Container(alertmonitor, "Monitoring script", "Monitor for the alert panel")
    Container(monkey, "TamperMonkey plugin", "Executor")
    Rel_D(monkey, alertmonitor, "Injects and executes the script")
    Rel(tvui, alertmonitor, "Provides alerts")
    Rel(tradingview, tvui, "Sends prices, executes PineScript strategies")
  }

}

Rel(tvtrader_distributor, alertq, "Distributes the alerts", "ThreadPool")
Rel(alertq, dist_targets, "Receives and processes the alerts")

Rel(tradingview, fetcher, "Fetches prices", "HTTPS GET")
Rel(fetcher, influx, "Gets specific prices periodaically", "REST GET")
Rel_R(influx, strategies, "Reads the prices and calculates the strategy action", "API")
Rel_U(strategies, tvtrader_backend, "Sends strategy alerts", "REST POST")
Rel_R(tvtrader_backend, tvtrader_frontend, "Receives alerts for visualization", "WebSocket")
Rel(tvtrader_backend, tvtrader_distributor, "Receives alerts for distribution", "WebSocket")
Rel_U(user, tvtrader_frontend, "Monitors the alerts")
Rel_U(user, future_frontend, "Monitors the alerts", $tags="future")
Rel_U(quant, opt_strategy, "Implements strategies")
Rel_U(quant, strategy_1, "Implements strategies")
@enduml

Features

Fetcher

• Receiving price information from TradingView
• Publishing the prices upon HTTP GET in JSON format

Store

• InfluxDB FluxScript to store the price information from the fetcher - (WIP)

Backend

• Receiving alerts via HTTP PUST (for example by custom scripts injected to TradingView using TamperMonkey)
• Forwarding received alerts to the connected websocket clients
• REST API (start the backend and go to [http://localhost:8089/swagger] for details)

Distributor

• Receive alerts from various sources (currently the backend's websocket)
• Forwarding alerts to the connected distribution targets
• Loading of distribution targets upon startup from a directory
  • Example console target implementation is included with a CLI

Frontend

• Receiving incoming alerts from the connected backend
• Displaying latest alerts on the Web UI
• Configurable timeout for the alerts
• Color coding the status of the alerts

Tech stack

• Python
  • Sanic framework for the backend
• Javascript
  • Vue with Vuex and Vuetify for the frontend
  • Nest.js for the fetcher
  • For the TamperMonkey scripts
• Docker for containerization (coming soon)

Installation

Clone the repository and change the current directory to it.

git clone https://github.com/dzooli/tvtrader.git
cd tvtrader

Testing

Use the requirements_dev.txt for virtualenv creation:

python -m venv .
source venv/bin/activate
pip install -r backend/requirements.txt
pip install -r backend/requirements_dev.txt
pytest backend
pytest agents/tvt_agents

Backend

Using a Linux environment:

python -m venv .
source bin/activate
pip install -r requirements.txt

Start backend

cd backend
./start.sh

# or after manual venv activation
python -m src.server

Under Windows the virtualenv creates a different directory structure and activation will also be different. Please read the virtualenv related documentation first.

Frontend

cd frontend
npm -g i yarn
yarn install

Start frontend

cd frontend
yarn serve # For development. Use 'yarn build' for production use and deploy it in your preferred way.

Usage

Alert catching setup

Install TamperMonkey Chrome extension and add the scripts from the scripts directory. Restrict them to run only on https://tradingview.com/charts/* on the settings page and do other security related settings as you wish. Do not forget, TamperMonkey is a powerful but double-edged tool.

Open TradingView, login and display the alerts panel. Add a properly formatted alert following the directions below.

TradingView alert setup

General rules

• Do not add blocking pop-up window
• Do not add disturbing actions
• Use the strategy.comment as "BUY" or "SELL" only
• Keep the Alerts panel open (the page script is based on this panel)

Alert message content

{
    "stratId":1,
    "stratName": "STRATEGY_NAME1",
    "symbol": "{{exchange}}:{{ticker}}",
    "interval": {{interval}},
    "direction": "{{strategy.order.comment}}",
    "timestamp": "{{timenow}}"
}

Example settings

Enjoy

Start the backend first and the frontend. Open the frontend with your browser on [http://localhost:8080/] and wait for the alerts.

Extras

In the doc folder you can find various useful information as well as the developer documentation (WIP).