marcogorelli/kaggle

Select Python Interpreter

ctrl + shift + p -> select interpreter -> enter interpreter path -> find -> /opt/conda/bin/python

Extensions

Install and enable PyLance.

Kaggle API key

Follow instructions here to get an API key. Temporarily save your kaggle.json file to this repository.

Set up development environment

Open this repository in vs code - see here for instructions for developing inside a container.

Move Kaggle API key

From within the container in vscode run

mkdir /root/.kaggle
mv kaggle.json /root/.kaggle/

Verify it works by running:

kaggle competitions list

New competition

Make a new repository for it on GitHub. Then, for each new competition, do e.g.

git submodule add [email protected]:MarcoGorelli/tweet-sentiment-extraction.git

Any input for the competition goes in the input folder here.

Push notebook to Kaggle

If you want to enable push a notebook to Kaggle for a particular competition, you can do

echo kaggle kernels push > .git/modules/<COMPETITION SLUG GOES HERE>/hooks/pre-push
chmod +x .git/modules/<COMPETITION SLUG GOES HERE>/hooks/pre-push

Then, any time you push a commit to GitHub, you will be deploying your latest changes to Kaggle. If you want to push a commit without deploying a new notebook to Kaggle, you can add the --no-verify flag when you push.