fetch-and-ingest.yaml

name: Fetch and ingest

on:
  schedule:
    # Note times are in UTC, which is 1 or 2 hours behind CET depending on daylight savings.
    #
    # Note the actual runs might be late.
    # Numerous people were confused, about that, including me:
    #  - https://github.community/t/scheduled-action-running-consistently-late/138025/11
    #  - https://github.com/github/docs/issues/3059
    #
    # Note, '*' is a special character in YAML, so you have to quote this string.
    #
    # Docs:
    #  - https://docs.github.com/en/actions/learn-github-actions/events-that-trigger-workflows#schedule
    #
    # Tool that deciphers this particular format of crontab string:
    #  - https://crontab.guru/
    #
    # Runs at 4pm UTC (12pm EDT) since curation by NCBI happens on the East Coast.
    - cron:  '0 16 * * 2'

  repository_dispatch:
    types:
      - fetch-and-ingest

  # Manually triggered using GitHub's UI
  workflow_dispatch:
    inputs:
      trial_name:
        description: 'Short name for a trial run. WARNING: without this we will overwrite files in s3://nextstrain-data/files/workflows/rsv'

jobs:
  set_config_overrides:
    runs-on: ubuntu-latest
    steps:
      - id: s3_dst
        run: |
          S3_DST=s3://nextstrain-data/files/workflows/rsv

          if [[ -n "$TRIAL_NAME" ]]; then
            S3_DST+=/trial/"$TRIAL_NAME"
          fi

          echo "s3_dst=$S3_DST" >> "$GITHUB_OUTPUT"
        env:
          TRIAL_NAME: ${{ inputs.trial_name }}
      - id: trigger_rebuild
        run: |
          TRIGGER_REBUILD=true

          if [[ -n "$TRIAL_NAME" ]]; then
            TRIGGER_REBUILD=false
          fi

          echo "trigger_rebuild=$TRIGGER_REBUILD" >> "$GITHUB_OUTPUT"
    outputs:
      s3_dst: ${{ steps.s3_dst.outputs.s3_dst }}
      trigger_rebuild: ${{ steps.trigger_rebuild.outputs.trigger_rebuild }}

  fetch-and-ingest:
    needs: [set_config_overrides]
    permissions:
      id-token: write
    uses: nextstrain/.github/.github/workflows/pathogen-repo-build.yaml@master
    secrets: inherit
    with:
      runtime: aws-batch
      run: |
        nextstrain build \
          --aws-batch \
          --detach \
          --no-download \
          --cpus 8 \
          --memory 32gib \
          --env PAT_GITHUB_DISPATCH="$GH_TOKEN_NEXTSTRAIN_BOT_WORKFLOW_DISPATCH" \
          --env S3_DST \
          --env TRIGGER_REBUILD \
          ingest \
            --configfiles config/config.yaml config/optional.yaml \
            --config s3_dst="$S3_DST" trigger_rebuild="$TRIGGER_REBUILD" \
            --printshellcmds
      env: |
        S3_DST: ${{ needs.set_config_overrides.outputs.s3_dst }}
        TRIGGER_REBUILD: ${{ needs.set_config_overrides.outputs.trigger_rebuild }}