Create a minimal folder structure for a data science project.
The instructions in this file have only been verified on Linux and MacOS systems. At the moment, it is not known if these instructions will work on a Windows system.
-
Install the Python packages
cookiecutter
tox
using
pip3 install cookiecutter tox
-
Install
Make
(link) -
Install
git
from here.
- Clone this repo
cd Downloads git clone https://github.com/elsdes3/cookiecutter-portfolio.git
- Set your prefered values for all variables in
cookiecutter-portfolio/cookiecutter-project/config.yaml
- Change into the project directory
cd cookiecutter-portfolio
- Create the templated project, run code formatting checks in the resulting project and run the resulting starter notebook (
01_get_data.ipynb
and02_process_data.ipynb
) programmatically usingmake build
- To test that the expected template is produced
make test clean-tests
- Every variable present in
cookiecutter.json
must also be present inconfig.yaml
. Values will only be taken fromcookiecutter-portfolio/cookiecutter-project/config.yaml
. Values incookiecutter-portfolio/cookiecutter-project/cookiecutter.json
will be ignored. - This template a customized version of the
cookiecutter-datascience
template (v2). - The Python library
tox
is used for managing Python virtual environments. See these links (1, 2) for details about howtox
can be used to do this for a machine learning project.
If you encounter any problems, please file an issue along with a detailed description.
Contributions are welcome, and they are greatly appreciated! Credit will always be given.
Distributed under the terms of the MIT license, cookiecutter-portfolio
is free and open source software.