A collection of Kaggle competitions I have participated in.
script/create_venv.sh
cd kaggles
mkdir -p <competition_name>/kernels
script/download_data.sh
script/pull_kernels.sh <competition_name>
kaggle k push -p <competition_name>/kernels/<kernel_dir>
Kaggle supports utility scripts to help avoid code duplication but they can be a little tricky to set up with a repo. This is how it's done here:
- Create a new kernel and add the utility code to it. E.g.
home-data-for-ml-course/src/
- Push the kernel to Kaggle.
kaggle k push -p home-data-for-ml-course/src/
- Go to Kaggle, open the utility kernel with the editor and click
File -> Set as utility script
. - For kernels that depend on the utility script, add the utility kernel to the
kernel_sources
list in thekernel-metadata.json
file. E.g.Alternatively, click"kernel_sources": [ "danielmcampos/house-price-utils" ],
File -> Add utility script
and select the utility script.