Skip to content

Latest commit

 

History

History
177 lines (136 loc) · 4.17 KB

README.md

File metadata and controls

177 lines (136 loc) · 4.17 KB

QuantsLab

QuantsLab is a Python project designed for quantitative research with Hummingbot. It provides functionalities for fetching historical data, calculating metrics, backtesting, and generating trading configurations.

Installation

Prerequisites

Steps

  1. Clone the repository:

    git clone https://github.com/hummingbot/quants-lab.git
cd quants-lab

  2. Create and activate the Conda environment:

    make install

    This command will create a new Conda environment and install all the necessary dependencies.

  3. Activate the environment:

    conda activate quants-lab

You're now ready to use QuantsLab!

Usage

1. Research Notebooks

  • Under the research notebooks folder you will find a folder for each strategy. The idea is that you can use them as an inspiration to do research on your own strategies.
  • The main steps are:
    • Exploratory Data Analysis
    • Design the controller
    • Backtest a simple controller
    • Optimize and find the best parameters

2. Task Orchestration

Prerequisites

  1. Ensure Docker and Docker Compose are installed. If not, you can install them with the following commands:

    curl -fsSL https://get.docker.com -o get-docker.sh
sh get-docker.sh

  2. Verify the Docker installation by running:

    docker --version
docker compose version

Configuration

  1. Modify Task Settings:

    • Navigate to the config folder and locate the tasks.yml file.
    • Update the file to include the specific tasks you want to execute. By default, it is configured to run the pool-fetching task.

  2. Customize Database Credentials (Optional):

    • The default credentials for MongoDB and PostgreSQL are specified in the docker-compose-db.yml file.
    • Update these credentials if necessary, especially for production environments.

Steps to Run Tasks

  1. Build the Docker Image: Build the local Quants-Lab Docker image by running:

    make build

  2. Start Databases: Start the necessary databases (MongoDB and PostgreSQL) using:

    make run-db

  3. Run the Task Runner: Execute tasks specified in tasks.yml with the following command:

    make run-task config=tasks.yml

  4. Monitor Database Activity: Use Mongo Compass UI to inspect the database data:

    • Open your web browser and visit: 
      http://localhost:28081/
    • Default credentials:
      • Username: admin
      • Password: changeme
    • Update these credentials in docker-compose-db.yml if needed.

    Replace localhost with your machine's IP address if accessing remotely.

Stopping Services

  1. Stop the Task Runner:

    make stop-task

  2. Stop the Databases:

    make stop-db

Notes:

  • Ensure all required ports are open and accessible.
  • Regularly check the logs for errors using docker logs <container_name>.
  • Make sure to re-build the local Docker image using make build after any changes are made to tasks.yml.

Data Source

  • CLOB (Central Limit Order Book)

    • Last Traded Price
    • Current Order Book
    • Historical Candles
    • Historical Trades
    • Trading Rules
    • Funding Info

  • AMM (Automated Market Maker)

    • Last Traded Price
    • Current Liquidity
    • Pool Stats
      • Fees Collected
      • Volume (24h)
    • Historical Trades

  • GeckoTerminal

    • Networks
    • Dexes by Network
    • Top Pools by Network
    • Top Pools by Network Dex
    • Top Pools by Network Token
    • New Pools by Network
    • New Pools (All Networks)
    • OHLCV

  • CoinGecko

    • Top Tokens Stats
    • Top Exchange Stats
    • Market Stats by Token
    • Market Stats by Exchange

  • Spice (DuneAnalytics)

    • Queries

Modules

  • Labeling

    • Triple Barrier Method

  • Backtesting

  • Optimization

  • Visualization

    • OHLC
    • Order Book
    • Backtesting Report

  • Features

    • Signals