Getting around with stock data.
Live preview:
- http://qwde.no/ - index page
- http://qwde.no/bb/twtr/20150102?toDate=20170301 - bollinger bands
- http://qwde.no/sma/twtr/20150102?toDate=20170301 - simple moving averages (you can change dates and tickers in URL from around 20140101 to around 20170330)
The general idea is to analyze stock data. One day in some future we hope to make this into a trading robot. For now, to get code done, its mostly important to have fun and build something interesting.
The project reads closing-prices and info from the project pystock-data, which is inserted into a database (sqlite). From there, we do simple analysis and present it using a websocket implementation combined with the all-powerful plotly and tablesaw.
analyticsfolder contains code for ML, analysis, etc.dataproviderfolder contains code for getting stock-data and making it available for other projectswebfolder contains code for presenting views and running web server
git clone --recursive https://github.com/kwrl/qwde
if you forgot recursive:
git submodule update --init
Install openjdk 11 or higher to run the code.
You can build the project with ./gradlew build
We ship with a java dependency that takes care of all database and network. If you want to look at/debug the database, install sqlite version 3 or higher. After that, you can run something like sqlite3 $XDG_CACHE_HOME/qwde/database.db.
./gradlew :web:run
or
./gradlew build to produce a runnable webserver at ./web/build/libs/shadow-<VERSION>-all.jar. You can run this with java -jar. To make sure you find the data, make sure to either run (and copy) the jar file to the project root, or point $XDG_DATA_HOME to the folder containing pystock-data. Our configuration directory contains the setup we use at Hetzner-server.
By default a sqlite database copy of pystock-data is written to
- Unix:
$XDG_CACHE_HOME(default ~/.cache/qwde) - Windows:
%APPDATA%/qwde/cache
Upon running the application, it is checked whether or not the database exists and has data. If not, it is generated. To re-generate the data, you can point $XDG_CACHE_HOME elsewhere, or delete the database.db file.
Code is continuously deployed to http://qwde.no. Its a droplet with 2vCPU and 8 GB RAM. If you want access, send me an email or PM.
Pull- and feature-requests welcome! 😄
The main requirements for any PR is that the tests still run. In general there will be no thorough code review unless you explicitly ask for it. This is to make it easier to "just fix things" and avoid the corporate procedures we normally go through 9 to 5. At least until the project is of a certain size.
Also, to speed up development, the policy is that all code can be moved around, refactored or deleted. As long as the tests run. Though, I wouldn't touch other people's code too much unless I saw a good reason to do so.
If you contribute with (scientific) code, a small explanation is much obliged. We also put some small footnotes on the wiki.