Usage

databooks is a CLI tool, but there are many different ways in which one can make use of the tool - such as CI or hooks.

CLI tool

The most straightforward way is to use it in the terminal, whenever desired. That can be error-prone and have "dirty" notebooks in your git repo. Check CLI documentation for more information.

A safer alternative is to automate this step, by setting up CI in your repo or a pre-commit hook.

GitHub Actions

GitHub Actions are a GitHub-hosted solution for CI/CD. All you need to get started is a file in project_root/.github/workflows/nb-meta.yml.

An example workflow to clean any notebook metadata and commit changes at every push:

name: 'nb-meta'
on: [push]

jobs:
  nb-meta:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2
      - uses: actions/setup-python@v2
        with:
          python-version: 3.8
      - name: Configure git user
        run: |
          git config --global user.name 'github-actions[bot]'
          git config --global user.email 'github-actions[bot]@users.noreply.github.com'
      - name: Install dependencies and clean metadata
        run: |
          pip install databooks
          databooks meta . --overwrite
      - name: Commit changes and push
        run: |
          git commit -am "Automated commit - clean notebook metadata"
          git push

Alternatively, one can choose to avoid having CI systems making code changes. In that case, we can only check whether notebooks have any undesired metadata.

name: 'nb-meta'
on: [push]

jobs:
  nb-meta:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2
      - uses: actions/setup-python@v2
        with:
          python-version: 3.8
      - name: Install dependencies and check metadata
        run: |
          pip install databooks
          databooks meta . --check