The purpose of this repository is to document and backup for easy reproduction, our server configurations for analysis reproducability.
Only dependency needed to set up this system is docker which prepares all the other dependencies in a reproducable way.
# Create working directory
mkdir -p fdudatamining
cd fdudatamining
# Clone relevant repositories
git clone https://github.com/fdudatamining/system.git
git clone https://github.com/fdudatamining/framework.git
# Start system
cd system
MYSQL_ROOT_PASSWORD=your_secret_password docker-compose up
# Wait for the system to be completely built and come up
# Populate the database with our cleaned open healthcare data (optional)
# NOTE: Plenty of RAM (~24G) and about 16GB of free storage is recommended to process all the data
git clone https://github.com/fdudatamining/data.git
cd data
docker-compose build
MYSQL_ROOT_PASSWORD=your_secret_password docker-compose run data
# Note you can also specify a single table of interest at the end of this command
# Clear cache to free up space (optional)
rm -r .cache
# Point your favorite web browser to http://localhost/
Work should be saved in ./work, the database is saved in a docker volume named data. PhpMyAdmin sessions are saved in a docker volume named sessions.
Note that volumes are persistent accross instances--if something failed and you want a clean slate, execute:
docker-compose rm -v