deeprecsys is an open tool belt to speed up the development of modern data science projects at an enterprise level.
These words were chosen very carefully, and by them we mean:
- Open: we rely on OSS and distribute openly with a GNU GPLv3 license that won't change in the future. The official distribution channels are pypi (see deeprecsys at pypi) and GitHub (see deeprecsys at Github).
- Tool belt: this project contains code that may extract, process, analyse, aggregate, test, and present data.
- Modern: the code will be updated as much as possible to the newest versions, as long as they are stable and don't break pre-existing functionality.
- Data Science: This project will contain a mixture of data engineering, machine learning engineering, data analysis, and data visualization.
- Enterprise: The code deployed here will likely have been battle-tested by large organizations with millions of customers. Unless stated, it is production-ready. All code including dependencies is audited and secure.
If you're here from the research piece Optimized Recommender Systems with Deep Reinforcement Learning, please checkout the old branch origin/thesis for reproducibility. The README should contain instructions to get you started.
Installation depends on your framework, so you may need to adapt this. Here's an example using pip:
pip install deeprecsys
All source control is done in git, via GitHub. Make sure you have a modern version of git installed. For instance, you can checkout the project using SSH with:
git clone [email protected]:luksfarris/deeprecsys.git
All scripts are written using Taskfile. You can install it following Task's instructions. The file with all the tasks is Taskfile.yml.