Crypto Signals
Development state: Alpha (There are many bugs and documentation is often lagging)
Crypto Signals is a command line tool that automates your crypto currency Technical Analysis (TA) and trading.
Track over 500 coins across Bittrex, Bitfinex, GDAX, Gemini and more!
Technical Analysis Automated:
- Relative Strength Index (RSI)
- Ichimoku Cloud (Leading Span A, Leading Span B, Conversion Line, Base Line)
- Simple Moving Average
- Exponential Moving Average
- Breakouts / Pumps
- MACD
Alerts:
- SMS via Twilio
- Slack
- Telegram
Features:
- Modular code for easy trading strategy implementation
- Easy install with Docker
You can build on top of this tool and implement algorithm trading and some machine learning models to experiment with predictive analysis.
Coming Soon:
- Automated buying/selling
How to use (Docker)
- First make sure you have Docker installed
- Next, to create the docker image run
make buildin the root of the project directory. - Create a .env file which can be populated with settings in the format of OPTION=value which can be derived from the app/default-config.json file. For example if you want to change the how often it updates add SETTINGS_UPDATE_INTERVAL=600
- For lists of values separate them with commas. For instance if you want to use specific symbol pairs they are in the format of base_currency/quote_currency (i.e. SETTINGS_MARKET_PAIRS=BTC/ETH,BTC/USDT)
How to run
In the root directory run docker-compose run app or make build && make run --env-file=.env if you don't have docker-compose.
How to use (Local)
To install the dependencies for this project, perform the following...
- Ensure you are running python 3.6
- install TA-lib from https://www.ta-lib.org/ for your OS.
cd apppip install numpy==1.14.0pip install -r requirements.txt
You can add a secrets.json file to the app directory of your project to customize the configuration, the defaults are in app/default-config.json.
How to run
Navigate to the app directory in your terminal and run with "python app.py"
Behaviours
A behaviour is a functionality of the program that can be modified via the "selected_task" value in default-config.json. It current accepts three values: 'default', 'rsi_bot', 'reporter', and 'server'.
Default
"selected_task": "default"
This is the default behaviour of the bot. By default, it polls Bittrex (or any other exchange you configured) and reports the price analysis of each coin pair available. This currently consists of six indicators: Breakout, RSI, SMA, EMA, Ichimoku Cloud, and MACD. More will be added in future versions.
Simple Bot
"selected_task": "rsi_bot"
TODO
Reporter
"selected_task": "reporter"
TODO
Server (Backtesting)
"selected_task": "server"
Forked from the Cryptocurrency Trading Bot Tutorial on: https://youtube.com/cryptocurrencytrading
This behaviour runs a flask server hosting a website allowing you to test different backtesting strategies on various sets of historical data.
Installation
First, clone or download the repository to your computer.
Front End- Navigate to the app/behaviours/ui/www directory and run npm install. Make sure you have the latest version of node.js installed on your computer.
Back End- (Without Docker) The server should run with both python 2.7.x and 3.x. The only dependencies are numpy and flask, so just do a quick pip install flask, pip install numpy.
(With Docker) Everything should be installed already from the dependencies if you run make build in the root directory of crypto-signal.
Running the Application
First, you'll need to use webpack to bundle all of the React .jsx files on the front end. Navigate to the app/beahviours/ui/www directory and run npm run build.
(Without Docker) Next, navigate to the app/behaviours/ui directory and run python server.py.
(With Docker) Run docker-compose up in the root directory.
Now you're all set! Open up your favorite browser and navigate to http://localhost:5000/ and try it out.
How does it work?
First, you'll select the coin pair you want to trade with.
"Capital" is the amount of BTC you want to start out with trading.
"Time Unit" is the duration of each point on the time series of historical data (1m, 5m, 30m, etc.).
"Stop Loss" is the amount of BTC below each buy price that you will sell your position at. The smaller the stop loss, the less your risk.
"Start Time" is the start date that your backtesting data will be grabbed from.
Strategies
The middle panel on the website allows you to create customizable strategies to run over your historical data. You can choose from various indicators and select a comparision operator to compare the indicator on the left to another indicator or a number. The "value" field on the right can take either numbers (i.e. "30", "0.00034") or one of the suggested indicators from the dropdown (i.e. "RSI", "Current Price"). Any other inputs are invalid. This will be improved on in the future for better user experience.
There are two buttons below the buy and sell strategy fields. Clicking on the "+" button will add another condition to your strategy. Clicking on the "-" button will remove the most recently added condition. You may add as many of these as you want. As of right now, if multiply strategy conditions are present they will be evaluated conjunctively. In other words, if your "Buy When" conditions are
RSI < 40
Current Price < Moving Average (15 Period)
Then the bot will buy only when both conditions are true. The same applies for the sell conditions.
What's on the Graph?
Green dots represent buy points. Red dots represent sell points. The blue line is the plot of historical closing prices, and the dotted yellow lines are Bollinger Bands. Moving averages will appear red and green, but this will be customizable later.
Coming Soon
MACD plot, RSI plot
Liability
I am not your financial adviser, nor is this tool. Use this program as an educational tool, and nothing more. None of the contributors to this project are liable for any loses you may incur. Be wise and always do your own research.