Skip to content

Trading Bot with focus on Evolutionary Algorithms and Machine Learning

License

Notifications You must be signed in to change notification settings

alirezaseifi/mosquito

 
 

Repository files navigation

Flexible Trading Bot with main focus on Machine Learning and Genetic Algorithms, inspired by zenbot.

Build Status Codacy Badge Gitter chat codebeat badge

About

Mosquito is a crypto currency trading bot written in Python, with main focus on modularity, so it is straight forward to plug-in new exchange.

The idea to build a new bot came because of I was missing following easy-access features in all of available bots available at the time mid 2017:

  • Multi-currency bot - Be able to monitor and exchange several currencies in one strategy (no need to run 5 same strategies for 5 different currencies).
  • Easy AI plug - Possibility to easily plug any of the existing AI/ML libraries (for ex. scikit or keras)

Please be AWARE that Mosquito is still in beta and under heavy development. Please use it in Live trading VERY carefully.

Supported Exchanges

Mosquito currently supports following exchanges:

  • Poloniex - supporting fillOrKill and immediateOrCancel trading types. postOnly type is not supported. You can read more about trading types here.
  • Bittrex - supporting Trade Limit Buy/Sell Orders

Requirements

Quick Start

Install

  1. clone repo
git clone https://github.com/miti0/mosquito.git
  1. install mongodb & required python packages

  2. set-up mosquito.ini (if you want to use sample config, just rename mosquito.sample.ini to mosquito.ini)

  3. Run desired command (full list of commands below)

All parameters in the program can be overridden with input arguments. You can get list of all available arguments with:

 python3 mosquito.py --help

Backfill

Backfill gets history data from exchange and stores them to mongodb. Data can be after that used for testing your simulation strategies.

usage: backfill.py [-h] [--pairs PAIRS] [--all] --days DAYS

optional arguments:
  -h, --help   show this help message and exit
  --pairs PAIRS PairS to backfill. For ex. [BTC_ETH, BTC_* (to get all BTC_*
               prefixed pairs]
  --all        Backfill data for ALL currencies
  --days DAYS  Number of days to backfill

Example 1) Load historical data for BTC_ETH pair for the last 5 days:

python3 backfill --days 5 --pairs BTC_USD

Example 2) Load historical data for ALL pairs for the last 2 days

python3 backfill --days 3 --all

Example 3) Load historical data for all pairs starting with BTC_ for the last day

python3 backfill --days 1 --pairs BTC_*

Trading

This is the main module that handles passed strategy and places buy/sell orders.

Architecture and logic of mosquito is made so, that it should be easy to set and tune all strategy parameters with program arguments. Below is a list of main arguments that can be either configured via the mosquito.ini config file or by passing the value/values as argument.

-h, --help            show this help message and exit
 --polo_api_key POLO_API_KEY
                       Poloniex API key (default: None)
 --polo_secret POLO_SECRET
                       Poloniex secret key (default: None)
 --polo_txn_fee POLO_TXN_FEE
                       Poloniex txn. fee (default: None)
 --polo_buy_order POLO_BUY_ORDER
                       Poloniex buy order type (default: None)
 --polo_sell_order POLO_SELL_ORDER
                       Poloniex sell order type (default: None)
 --bittrex_api_key BITTREX_API_KEY
                       Bittrex API key (default: None)
 --bittrex_secret BITTREX_SECRET
                       Bittrex secret key (default: None)
 --bittrex_txn_fee BITTREX_TXN_FEE
                       Bittrex txn. fee (default: None)
 --exchange EXCHANGE   Exchange (default: None)
 --db_url DB_URL       Mongo db url (default: None)
 --db_port DB_PORT     Mongo db port (default: None)
 --db DB               Mongo db (default: None)
 --pairs PAIRS         Pairs (default: None)
 --use_real_wallet     Use/not use fictive wallet (only for paper simulation)
                       (default: False)
 --backtest_from BACKTEST_FROM
                       Backtest epoch start datetime (default: None)
 --backtest_to BACKTEST_TO
                       Backtest epoch end datetime (default: None)
 --backtest_days BACKTEST_DAYS
                       Number of history days the simulation should start
                       from (default: None)
 --wallet_currency WALLET_CURRENCY
                       Wallet currency (separated by comma) (default: None)
 --wallet_amount WALLET_AMOUNT
                       Wallet amount (separated by comma) (default: None)
 --backtest            Simulate your strategy on history ticker data
                       (default: False)
 --paper               Simulate your strategy on real ticker (default: False)
 --live                REAL trading mode (default: False)
 --plot                Generate a candle stick plot at simulation end
                       (default: False)
 --ticker_size TICKER_SIZE   Simulation ticker_size (default: 5)
 --root_report_currency ROOT_REPORT_CURRENCY
                       Root currency used in final plot (default: None)
 --buffer_size BUFFER_SIZE
                       Buffer size in days (default: 30)
 --prefetch            Prefetch data from history DB (default: False)
 --plot_pair PLOT_PAIR
                       Plot pair (default: None)
 --all ALL             Include all currencies/tickers (default: None)
 --days DAYS           Days to pre-fill (default: None)
 -c CONFIG, --config CONFIG
                       config file path (default: mosquito.ini)
 -v, --verbosity       Verbosity (default: False)
 --strategy STRATEGY   Strategy (default: None)
 --fixed_trade_amount FIXED_TRADE_AMOUNT
                       Fixed trade amount (default: None)

Currently Trading supports following modes:

  • Backtest - fast simulation mode using past data and placing fictive buy/sell orders.
  • Paper - mode simulating live ticker with placing fictive buy/sell orders.
  • Live - live trading with placing REAL buy/sell orders.

Backtest and Paper trading are using immediate buy/sell orders by using the last ticker closing price. This results to NOT 100% accurate strategy results, what you should be aware of.

Backtest

Fast simulation mode using past data and placing fictive buy/sell orders. Simulation configuration is done via config.ini file (some of the parameters can be overridden with command line arguments).

Below is an example of running a backtest together with final buy/sell plot generated at the end of the simulation.

python3 mosquito.py --backtest --plot

! Please be aware that Backtest should 99% work, but it is currently under final verification test.

Paper

Trading mode that simulates live ticker with placing fictive buy/sell orders. Simulation configuration is done via config.ini file (some of the parameters can be overridden with command line arguments).

Below is an example of running a backtest together with final buy/sell plot generated at the end of the simulation.

python3 mosquito.py --paper

! Please be aware that Paper should 99% work, but it is currently under final verification test.

Live

Live trading with placing REAL buy/sell orders. Configuration is done via config.ini file (some of the parameters can be overridden with command line arguments). Below is an example of running a backtest together with final buy/sell plot generated at the end of the simulation.

python3 mosquito.py --live

! Please be aware that Live should 99% work, but it is currently under final verification test.

Plot and Statistics

Mosquito has a simple plot utility for visualizing current pair combined with trading history. Visualization uses external library plotly. Below You can see an example visualizing ticker price plot, together with simulated buy/sell orders.

Below is an example of Final Simulation Report summary:

****************************************************
              Final simulation report:               
****************************************************
Wallet at Start: | 50.0DGB |
Wallet at End: | 51.3464723121DGB |
Strategy result: -5.68%
Buy & Hold: -8.16%
Strategy vs Buy & Hold: 2.47%
Total txn: 10
Simulated (data time): 0 days, 4 hours and 55 minutes
Transactions per hour: 2.03
Simulation run time: 0 hours 1 minutes and 13 seconds

AI

Blueprint

Blueprint is a part of AI package. Main function of the module is to generate datasets which can be used for training AI. Logic of Blueprint module is following:

  1. Create a blueprint file/module which contains features, indicators and output parameters. As an example you can take a look at ai/blueprints/minimal.py or ai/blueprints/junior.py

  2. Decide how many days you would like to run the Blueprint. Backfill data for that period.

  3. Choose which pair/pairs you would like to include. Following combinations should work [BTC_ETH] - single pair, [BTC_ETH, BTC_LTC] - list of pairs, [BTC_*] - all pairs with prefix BTC

  4. Start blueprint with following parameters (example below)

python3 blueprint.py --features junior --days 200

As a result you should see *.csv file in your Mosquito's root folder, which should contain the dataset.

Utilities

Wallet Lense

Simple module which sends up to 24h winners/losers market pairs summary by email in user specified intervals (sample below).

Usage

# You need to have configured email parameters in ini file, or pass them as input arguments.
python3 lense.py 

Donate

If you would like to support the project in other way than code-contributing, you can donate Mosquito development on following Bitcoin address:

16v94untkyuTXyE4euWzDA5r4vKd996CP8


License: GNU GENERAL PUBLIC LICENSE

  • Copyright (C) 2017 Miroslav Karpis (miti0)

    The GNU General Public License is a free, copyleft license for software and other kinds of works.

    The licenses for most software and other practical works are designed to take away your freedom to share and change the works. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change all versions of a program--to make sure it remains free software for all its users. We, the Free Software Foundation, use the GNU General Public License for most of our software; it applies also to any other work released this way by its authors. You can apply it to your programs, too.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

About

Trading Bot with focus on Evolutionary Algorithms and Machine Learning

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Jupyter Notebook 72.0%
  • Python 28.0%