PyPI Data Pipeline #305
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: PyPI Data Pipeline | |
| permissions: | |
| id-token: write | |
| on: | |
| workflow_dispatch: | |
| inputs: | |
| database_name: | |
| description: 'Database name' | |
| required: false | |
| default: 'duckdb_stats' | |
| type: string | |
| job_type: | |
| description: 'Job to run' | |
| required: true | |
| default: 'all' | |
| type: choice | |
| options: | |
| - ingest | |
| - transform | |
| - all | |
| - update | |
| start_date: | |
| description: 'Start Date' | |
| required: false | |
| default: '' | |
| end_date: | |
| description: 'End Date' | |
| required: false | |
| default: '' | |
| schedule: | |
| - cron: '0 0 * * *' | |
| jobs: | |
| setup: | |
| runs-on: ubuntu-latest | |
| outputs: | |
| start_date: ${{ steps.dates.outputs.START_DATE }} | |
| end_date: ${{ steps.dates.outputs.END_DATE }} | |
| run_ingest: ${{ steps.set-flags.outputs.run_ingest }} | |
| run_transform: ${{ steps.set-flags.outputs.run_transform }} | |
| run_update: ${{ steps.set-flags.outputs.run_update }} | |
| steps: | |
| - name: Calculate Dates | |
| id: dates | |
| run: | | |
| if [ -n "${{ github.event.inputs.start_date }}" ] && [ -n "${{ github.event.inputs.end_date }}" ]; then | |
| echo "START_DATE=${{ github.event.inputs.start_date }}" >> $GITHUB_OUTPUT | |
| echo "END_DATE=${{ github.event.inputs.end_date }}" >> $GITHUB_OUTPUT | |
| else | |
| START_DATE=$(date -d '2 days ago' +%Y-%m-%d) | |
| END_DATE=$(date -d '1 day ago' +%Y-%m-%d) | |
| echo "START_DATE=${START_DATE}" >> $GITHUB_OUTPUT | |
| echo "END_DATE=${END_DATE}" >> $GITHUB_OUTPUT | |
| fi | |
| - name: Set job flags | |
| id: set-flags | |
| run: | | |
| if [[ "${{ github.event_name }}" == "schedule" ]] || [[ "${{ github.event.inputs.job_type }}" == "all" ]]; then | |
| echo "run_ingest=true" >> $GITHUB_OUTPUT | |
| echo "run_transform=true" >> $GITHUB_OUTPUT | |
| echo "run_update=true" >> $GITHUB_OUTPUT | |
| elif [[ "${{ github.event.inputs.job_type }}" == "ingest" ]]; then | |
| echo "run_ingest=true" >> $GITHUB_OUTPUT | |
| echo "run_transform=false" >> $GITHUB_OUTPUT | |
| echo "run_update=false" >> $GITHUB_OUTPUT | |
| elif [[ "${{ github.event.inputs.job_type }}" == "transform" ]]; then | |
| echo "run_ingest=false" >> $GITHUB_OUTPUT | |
| echo "run_transform=true" >> $GITHUB_OUTPUT | |
| echo "run_update=false" >> $GITHUB_OUTPUT | |
| elif [[ "${{ github.event.inputs.job_type }}" == "update" ]]; then | |
| echo "run_ingest=false" >> $GITHUB_OUTPUT | |
| echo "run_transform=false" >> $GITHUB_OUTPUT | |
| echo "run_update=true" >> $GITHUB_OUTPUT | |
| fi | |
| ingest-pypi-data: | |
| needs: setup | |
| runs-on: ubuntu-latest | |
| if: needs.setup.outputs.run_ingest == 'true' | |
| steps: | |
| - name: Checkout repository | |
| uses: actions/checkout@v4 | |
| with: | |
| sparse-checkout: | | |
| Makefile | |
| sparse-checkout-cone-mode: false | |
| - id: 'auth' | |
| name: 'Authenticate to GCP' | |
| uses: 'google-github-actions/auth@v2' | |
| with: | |
| credentials_json: '${{ secrets.GCP_SECRETS }}' | |
| - name: Pull latest Docker image | |
| run: docker pull ghcr.io/${{ github.repository }}:latest | |
| - name: Ingest pypi data | |
| env: | |
| START_DATE: ${{ needs.setup.outputs.start_date }} | |
| END_DATE: ${{ needs.setup.outputs.end_date }} | |
| PYPI_PROJECT: duckdb | |
| DATABASE_NAME: ${{ github.event.inputs.database_name || 'duckdb_stats' }} | |
| GCP_PROJECT: ${{ secrets.GCP_PROJECT_ID }} | |
| TIMESTAMP_COLUMN: timestamp | |
| DESTINATION: md | |
| DOCKER_IMAGE: ghcr.io/${{ github.repository }}:latest | |
| AWS_PROFILE: None | |
| S3_PATH: None | |
| motherduck_token: ${{ secrets.MOTHERDUCK_TOKEN }} | |
| GOOGLE_APPLICATION_CREDENTIALS: ${{ steps.auth.outputs.credentials_file_path }} | |
| run: | | |
| make pypi-ingest DOCKER=true | |
| transform-pypi-data: | |
| needs: [setup, ingest-pypi-data] | |
| runs-on: ubuntu-latest | |
| if: | | |
| always() && | |
| needs.setup.outputs.run_transform == 'true' && | |
| (needs.ingest-pypi-data.result == 'success' || needs.ingest-pypi-data.result == 'skipped') | |
| steps: | |
| - name: Checkout repository | |
| uses: actions/checkout@v4 | |
| with: | |
| sparse-checkout: | | |
| Makefile | |
| sparse-checkout-cone-mode: false | |
| - id: 'auth' | |
| name: 'Authenticate to GCP' | |
| uses: 'google-github-actions/auth@v2' | |
| with: | |
| credentials_json: '${{ secrets.GCP_SECRETS }}' | |
| - name: Pull latest Docker image | |
| run: docker pull ghcr.io/${{ github.repository }}:latest | |
| - name: Transform pypi data using dbt | |
| env: | |
| START_DATE: ${{ needs.setup.outputs.start_date }} | |
| END_DATE: ${{ needs.setup.outputs.end_date }} | |
| DATABASE_NAME: ${{ github.event.inputs.database_name || 'duckdb_stats' }} | |
| DBT_TARGET: prod | |
| DOCKER_IMAGE: ghcr.io/${{ github.repository }}:latest | |
| motherduck_token: ${{ secrets.MOTHERDUCK_TOKEN }} | |
| GOOGLE_APPLICATION_CREDENTIALS: ${{ steps.auth.outputs.credentials_file_path }} | |
| run: | | |
| make pypi-transform DOCKER=true DBT_TARGET=${DBT_TARGET} | |
| update-motherduck-share: | |
| needs: transform-pypi-data | |
| runs-on: ubuntu-latest | |
| if: needs.transform-pypi-data.result == 'success' || needs.transform-pypi-data.result == 'skipped' | |
| steps: | |
| - name: Setup DuckDB | |
| uses: opt-nc/[email protected] | |
| with: | |
| version: v1.1.0 | |
| - name: Update MotherDuck share | |
| env: | |
| motherduck_token: ${{ secrets.MOTHERDUCK_TOKEN }} | |
| DATABASE_NAME: ${{ github.event.inputs.database_name || 'duckdb_stats' }} | |
| run: | | |
| echo "ATTACH 'md:';" >> ~/.duckdbrc && duckdb -c "UPDATE SHARE $DATABASE_NAME;" | |