estebanthi/binance-trading-bot

Basic Overview

Using crypto historical data, backtest your strategies, optimize the parameters and analyze the results. Once you found a promising strategy, automatically trade with it in live using Binance API. Be notified on Telegram about what's happening within the bot.

Visualize the results

Key features

Develop trading components easily :
- Strategies : to decide whether to enter / exit a trade or not.
- Indicators : to give information about data to your strategies.
- Analyzers : to analyze the results of your backtests or trading sessions.
- Sizers : to define how much quantity of assets the engine should buy depending on parameters.
- Observers : to define what you plot on the results chart.
- Timers : to define actions to do at a precise time, repeatedly or not.
Backtest and optimize your strategies to find the best parameters.
Visualize the results on a chart you can configure.
Automatic live trading with fake or real money, using Binance API.
Telegram notifications when the bot does specific actions (entering a trade, loading data, ...)
MongoDB support if you don't want to store data locally.

How to install

# clone the repo
git clone https://github.com/estebanthi/BinanceTradingBotV4

# go to folder
cd BinanceTradingBotV4/bot

# install requirements
pip3 install -r requirements.txt
pip3 install git+git://github.com/Dave-Vallance/bt-ccxt-store

Go into backtrader package, in folder plot, open file locator.py, and remove warnings in line 35, because it is deprecated :

from matplotlib.dates import (HOURS_PER_DAY, MIN_PER_HOUR, SEC_PER_MIN,
                              MONTHS_PER_YEAR, DAYS_PER_WEEK,
                              SEC_PER_HOUR, SEC_PER_DAY,
                              num2date, rrulewrapper, YearLocator,
                              MicrosecondLocator, warnings) # Remove warnings here

How to use

To use the bot, just go in the folder bot and create your first run script, or edit the one already existing. But first, you have to create the config.yml file.

Create and configure the config file

First, create a config.yml in the bot file. Fill it like the following :

api_key: "<YOUR BINANCE API KEY>"
api_secret: "<YOUR BINANCE API SECRET>"

telegram_token: "<YOUR TELEGRAM BOT TOKEN>" # If you want to use Telegram
user: "<YOUR TELEGRAM USER ID>"

mongo_url: "<YOUR MONGO DATABASE URL>" # If you want to use MongoDB for storing data

If you don't know how to use Binance API, you can check it on https://www.binance.com/en-NG/support/faq/360002502072.

If you don't know how to get a Telegram token, you can check it on https://sendpulse.com/knowledge-base/chatbot/create-telegram-chatbot.

Setting up your trading components

First, you need to decide what you want to use with the engine. So, for example, you can set up your components like that :

# (you can use multiple strategies at the same time for backtesting or live trading)
strategies = [TripleEMA(logging=True), MACD_Crossings(period_me1=12)]
sizer = DefaultSizer() # you can use only one sizer (it's logic !)
analyzers = [TradeAnalyzer(), Returns()]
observers = [ValueObserver()]
timers = [StopSession(when=dt.time(0), weekdays=[7])]

Use additional features

You can use some more features, like a writer to write details about your session into a file, or a TelegramBot to be notified on Telegram.

telegram_bot = TelegramBot()
write_to = "results.txt"

Set up your engine configuration

config = EngineConfiguration(
	mode="BACKTEST", # Chose between BACKTEST, OPTIMIZE, PAPER or LIVE
    symbol="BTC/EUR", # What you want to trade
    start_date="2021/01/01 00:00:00", # Session start date (format YYYY/MM/DD HH:MM:SS)
    end_date=dt.datetime.now(), # Session end
    timedelta=dt.timedelta(days=100), # Can replace start date because it will be calculated using this end_date - timedelta
    timeframe=bt.TimeFrame.Days, # Timeframe
    compression=1, # Timeframe * compression = resolution
    strategies=strategies,
    analyzers=analyzers,
    sizer=sizer,
    observers=observers,
    timers=timers,
    cash=100_000, # Virtual cash for the broker
    commission=0.2 # 0.2% broker commission 
    currency="EUR" # What you use to live trade
    telegram_bot=telegram_bot, # If you want to use Telegram
    write_to=write_to, # If you want to save results file
    debug=True, # If you want to debug the datafeed
    stdstats=True, # For default chart plotting
    exchange=ccxt.binance() # Exchange you want to use for data extraction (default is bitfinex)
)

Run the engine

When you have your config, you can just run your engine.

engine = Engine() # Instantiate the engine
engine.set_configuration(config) # Set the config
result = engine.run() # Get the results
engine.plot() # Display the chart

# If you use a telegram bot, you can send the results to your Telegram
telegram_bot.send_file(open(f"data/backtesting_results/{write_to}", "r"))

Develop !

If you want to use the bot the best way, you will have to develop strategies, indicators, analyzers, ... First, you have to understand the backtrader python library, because everything operates around this. Find the documentation here : https://www.backtrader.com/docu/.

I will add the documentation later for more details...

Notes

Raspberry PI

For backtesting, it's better to have a strong processor, but for live trading, a nice implementation for this trading bot is on Raspberry PI, so that it can runs 24/7.

Multiple timeframes

Be careful when using multiple strategies with multiple timeframes, because it can cause some issues. Check the code for more details.

estebanthi / binance-trading-bot