stage-build.yml

name: Stage Build

# NOTE! This is the *STAGE* workflow.
# Keep in mind that much of the configuration is repeated in `prod-build.yml`
# and `dev-build.yml`
#
# For a complete picture of all environments, see:
#
#  https://docs.google.com/spreadsheets/d/1VnnEl-iTtKYmlyN02FiEXygxZCgE4o_ZO8wSleebne4/edit?usp=sharing
#

env:
  DEFAULT_DEPLOYMENT_PREFIX: "main"
  DEFAULT_NOTES: ""
  DEFAULT_LOG_EACH_SUCCESSFUL_UPLOAD: "false"

on:
  schedule:
    # * is a special character in YAML so you have to quote this string
    - cron: "0 */24 * * *"

  workflow_dispatch:
    inputs:
      notes:
        description: "Notes"
        required: false
        default: ${DEFAULT_NOTES}

      # This is very useful when combined with the "Use workflow from"
      # feature that is built into the "Run workflow" button on
      # https://github.com/mdn/yari/actions?query=workflow%3A%22Production+Build%22
      # If you override the deployment prefix to something like the name
      # of the branch, you can deploy that entire branch to its own prefix
      # in S3 which means that it can be fully hosted as its own site.
      deployment_prefix:
        description: "Deployment prefix"
        required: false
        default: ${DEFAULT_DEPLOYMENT_PREFIX}

      log_each_successful_upload:
        description: "Deployer logs each success"
        required: false
        default: ${DEFAULT_LOG_EACH_SUCCESSFUL_UPLOAD}

      invalidate:
        description: "Invalidate CDN (use only in exceptional circumstances)"
        type: boolean
        required: false
        default: false

  workflow_call:
    secrets:
      GCP_PROJECT_NAME:
        required: true
      WIP_PROJECT_ID:
        required: true

permissions:
  contents: read
  id-token: write

jobs:
  build:
    environment: stage
    runs-on: ubuntu-latest

    # Only run the scheduled workflows on the main repo.
    if: github.repository == 'mdn/yari'

    steps:
      - uses: actions/checkout@v4

      - uses: actions/checkout@v4
        if: ${{ ! vars.SKIP_BUILD || ! vars.SKIP_FUNCTION }}
        with:
          repository: mdn/content
          path: mdn/content
          # Yes, this means fetch EVERY COMMIT EVER.
          # It's probably not sustainable in the far future (e.g. past 2021)
          # but for now it's good enough. We'll need all the history
          # so we can figure out each document's last-modified date.
          fetch-depth: 0

      - uses: actions/checkout@v4
        if: ${{ ! vars.SKIP_BUILD }}
        with:
          repository: mdn/mdn-studio
          path: mdn/mdn-studio
          lfs: true
          token: ${{ secrets.MDN_STUDIO_PAT }}

      - uses: actions/checkout@v4
        if: ${{ ! vars.SKIP_BUILD }}
        with:
          repository: mdn/curriculum
          path: mdn/curriculum

      # Our usecase is a bit complicated. When the cron schedule runs this workflow,
      # we rely on the env vars defined at the top of the file. But if it's a manual
      # trigger we rely on the inputs and only the inputs. That way, the user can
      # opt to type in 'false'.
      # It's not possible to express this with GitHub Workflow syntax, so we
      # have a dedicate set that conveniently sets these as env vars which we
      # can refer to later in `if: ....` lines or in bash with the `run: ...` blocks.
      - name: Merge dispatch inputs with default env vars
        run: |
          echo "DEPLOYER_BUCKET_PREFIX=${{ github.event.inputs.deployment_prefix || env.DEFAULT_DEPLOYMENT_PREFIX }}" >> $GITHUB_ENV
          echo "DEPLOYER_LOG_EACH_SUCCESSFUL_UPLOAD=${{ github.event.inputs.log_each_successful_upload || env.DEFAULT_LOG_EACH_SUCCESSFUL_UPLOAD }}" >> $GITHUB_ENV

      - uses: actions/checkout@v4
        if: ${{ ! vars.SKIP_BUILD || ! vars.SKIP_FUNCTION }}
        with:
          repository: mdn/translated-content
          path: mdn/translated-content
          # See matching warning for mdn/content checkout step
          fetch-depth: 0

      - uses: actions/checkout@v4
        if: ${{ ! vars.SKIP_BUILD }}
        with:
          repository: mdn/mdn-contributor-spotlight
          path: mdn/mdn-contributor-spotlight

      - name: Setup Node.js environment
        if: ${{ ! vars.SKIP_BUILD || ! vars.SKIP_FUNCTION }}
        uses: actions/setup-node@v4
        with:
          node-version-file: ".nvmrc"
          cache: yarn

      - name: Install all yarn packages
        if: ${{ ! vars.SKIP_BUILD }}
        run: yarn --frozen-lockfile
        env:
          # https://github.com/microsoft/vscode-ripgrep#github-api-limit-note
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

      - name: Install Python
        if: ${{ ! vars.SKIP_BUILD }}
        uses: actions/setup-python@v5
        with:
          python-version: "3.10"

      - name: Install Python poetry
        if: ${{ ! vars.SKIP_BUILD }}
        uses: snok/install-poetry@v1

      - name: Install deployer
        if: ${{ ! vars.SKIP_BUILD }}
        run: |
          cd deployer
          poetry install

      - name: Display Python & Poetry version
        if: ${{ ! vars.SKIP_BUILD }}
        run: |
          python --version
          poetry --version

      - name: Print information about build
        run: |
          echo "notes: ${{ github.event.inputs.notes || env.DEFAULT_NOTES }}"
          echo "log_each_successful_upload: ${{ github.event.inputs.log_each_successful_upload || env.DEFAULT_LOG_EACH_SUCCESSFUL_UPLOAD }}"
          echo "deployment_prefix: ${{ github.event.inputs.deployment_prefix || env.DEFAULT_DEPLOYMENT_PREFIX }}"

      - name: Print information about CPU
        run: cat /proc/cpuinfo

      - name: Build everything
        if: ${{ ! vars.SKIP_BUILD }}
        env:
          # Remember, the mdn/content repo got cloned into `pwd` into a
          # sub-folder called "mdn/content"
          CONTENT_ROOT: ${{ github.workspace }}/mdn/content/files
          CONTENT_TRANSLATED_ROOT: ${{ github.workspace }}/mdn/translated-content/files
          CONTRIBUTOR_SPOTLIGHT_ROOT: ${{ github.workspace }}/mdn/mdn-contributor-spotlight/contributors
          BLOG_ROOT: ${{ github.workspace }}/mdn/mdn-studio/content/posts
          CURRICULUM_ROOT: ${{ github.workspace }}/mdn/curriculum
          BASE_URL: "https://developer.allizom.org"

          # The default for this environment variable is geared for writers
          # (aka. local development). Usually defaults are supposed to be for
          # secure production but this is an exception and default
          # is not insecure.
          BUILD_LIVE_SAMPLES_BASE_URL: https://live.mdnyalp.dev
          BUILD_LEGACY_LIVE_SAMPLES_BASE_URL: https://live.mdnyalp.dev

          # Sign key for code samples
          BUILD_SAMPLE_SIGN_KEY: ${{ secrets.SAMPLE_SIGN_KEY }}

          # Use the stage version of interactive examples.
          BUILD_INTERACTIVE_EXAMPLES_BASE_URL: https://interactive-examples.mdn.allizom.net

          # Now is not the time to worry about flaws.
          BUILD_FLAW_LEVELS: "*:ignore"

          # This is the Google Analytics account ID for developer.mozilla.org
          # If it's used on other domains (e.g. stage or dev builds), it's OK
          # because ultimately Google Analytics will filter it out since the
          # origin domain isn't what that account expects.
          BUILD_GOOGLE_ANALYTICS_ACCOUNT: UA-36116321-5

          # This enables the Plus call-to-action banner and the Plus landing page
          REACT_APP_ENABLE_PLUS: true

          # This adds the ability to sign in (stage only for now)
          REACT_APP_DISABLE_AUTH: false

          # Use the stage version of interactive examples in react app
          REACT_APP_INTERACTIVE_EXAMPLES_BASE_URL: https://interactive-examples.mdn.allizom.net

          # Firefox Accounts and SubPlat settings
          REACT_APP_FXA_SIGNIN_URL: /users/fxa/login/authenticate/
          REACT_APP_FXA_SETTINGS_URL: https://accounts.stage.mozaws.net/settings/?brand=mozilla
          REACT_APP_MDN_PLUS_SUBSCRIBE_URL: https://accounts.stage.mozaws.net/subscriptions/products/prod_Jtbg9tyGyLRuB0?brand=mozilla
          REACT_APP_MDN_PLUS_5M_PLAN: price_1JFoTYKb9q6OnNsLalexa03p
          REACT_APP_MDN_PLUS_5Y_PLAN: price_1JpIPwKb9q6OnNsLJLsIqMp7
          REACT_APP_MDN_PLUS_10M_PLAN: price_1K6X7gKb9q6OnNsLi44HdLcC
          REACT_APP_MDN_PLUS_10Y_PLAN: price_1K6X8VKb9q6OnNsLFlUcEiu4

          # Surveys.
          REACT_APP_SURVEY_START_DISCOVERABILITY_AUG_2023: 0 # stage
          REACT_APP_SURVEY_END_DISCOVERABILITY_AUG_2023: 1691841600000 # new Date("2023-08-12 12:00:00Z").getTime()
          REACT_APP_SURVEY_RATE_FROM_DISCOVERABILITY_AUG_2023: 0.0
          REACT_APP_SURVEY_RATE_TILL_DISCOVERABILITY_AUG_2023: 0.10 # 10%

          # Telemetry.
          REACT_APP_GLEAN_CHANNEL: stage
          REACT_APP_GLEAN_ENABLED: true

          # Newsletter
          REACT_APP_NEWSLETTER_ENABLED: true

          # Placement
          REACT_APP_PLACEMENT_ENABLED: true

          # Playground
          REACT_APP_PLAYGROUND_BASE_HOST: mdnyalp.dev

          # Sentry.
          SENTRY_DSN_BUILD: ${{ secrets.SENTRY_DSN_BUILD }}
          SENTRY_ENVIRONMENT: stage
          SENTRY_RELEASE: ${{ github.sha }}

        run: |

          # Info about which CONTENT_* environment variables were set and to what.
          echo "CONTENT_ROOT=$CONTENT_ROOT"
          echo "CONTENT_TRANSLATED_ROOT=$CONTENT_TRANSLATED_ROOT"
          echo "BLOG_ROOT=$BLOG_ROOT"
          # Build the ServiceWorker first
          yarn build:sw
          yarn build:prepare

          # (July 15, 2021) This is a temporary solution. This should become an
          # integrated part of 'build:prepare'.
          # See https://github.com/mdn/yari/issues/4217
          yarn tool popularities

          yarn tool sync-translated-content

          # Build using one process per locale.
          # Note: We have 4 cores, but 9 processes is a reasonable number.
          for locale in en-us es fr ja ko pt-br ru zh-cn zh-tw; do
            yarn build --locale $locale 2>&1 | sed "s/^/[$locale] /"   &
            pids+=($!)
          done

          for pid in "${pids[@]}"; do
            wait $pid
          done

          du -sh client/build

          # Generate sitemap index file
          yarn build --sitemap-index

          # Build the blog
          yarn build:blog

          # Build the curriculum
          yarn build:curriculum

          # Generate whatsdeployed files.
          yarn tool whatsdeployed --output client/build/_whatsdeployed/code.json
          yarn tool whatsdeployed $CONTENT_ROOT --output client/build/_whatsdeployed/content.json
          yarn tool whatsdeployed $CONTENT_TRANSLATED_ROOT --output client/build/_whatsdeployed/translated-content.json

      - name: Update search index
        if: ${{ ! vars.SKIP_BUILD }}
        env:
          DEPLOYER_ELASTICSEARCH_URL: ${{ secrets.DEPLOYER_STAGE_ELASTICSEARCH_URL }}
        run: |
          cd deployer
          poetry run deployer search-index ../client/build

      - name: Authenticate with GCP
        if: ${{ ! vars.SKIP_BUILD }}
        uses: google-github-actions/auth@v2
        with:
          token_format: access_token
          service_account: deploy-stage-content@${{ secrets.GCP_PROJECT_NAME }}.iam.gserviceaccount.com
          workload_identity_provider: projects/${{ secrets.WIP_PROJECT_ID }}/locations/global/workloadIdentityPools/github-actions/providers/github-actions

      - name: Setup gcloud
        if: ${{ ! vars.SKIP_BUILD }}
        uses: google-github-actions/setup-gcloud@v2

      - name: Sync build
        if: ${{ ! vars.SKIP_BUILD }}
        run: |-
          gsutil -q -m -h "Cache-Control: public, max-age=3600" cp -r client/build/static gs://${{ vars.GCP_BUCKET_NAME }}/main/
          gsutil -q -m -h "Cache-Control: public, max-age=3600" rsync -cdrj html,json,txt -y "^static/" client/build gs://${{ vars.GCP_BUCKET_NAME }}/main

      - name: Authenticate with GCP
        if: ${{ ! vars.SKIP_FUNCTION }}
        uses: google-github-actions/auth@v2
        with:
          token_format: access_token
          service_account: deploy-stage-nonprod-mdn-ingre@${{ secrets.GCP_PROJECT_NAME }}.iam.gserviceaccount.com
          workload_identity_provider: projects/${{ secrets.WIP_PROJECT_ID }}/locations/global/workloadIdentityPools/github-actions/providers/github-actions

      - name: Setup gcloud
        if: ${{ ! vars.SKIP_FUNCTION }}
        uses: google-github-actions/setup-gcloud@v2
        with:
          install_components: "beta"

      - name: Generate redirects map
        if: ${{ ! vars.SKIP_FUNCTION }}
        working-directory: cloud-function
        env:
          CONTENT_ROOT: ${{ github.workspace }}/mdn/content/files
          CONTENT_TRANSLATED_ROOT: ${{ github.workspace }}/mdn/translated-content/files
        run: |
          npm ci
          npm run build-redirects

      - name: Deploy Function
        if: ${{ ! vars.SKIP_FUNCTION }}
        run: |-
          for region in europe-west1 us-west1 asia-east1; do
            gcloud beta functions deploy mdn-nonprod-stage-$region \
            --gen2 \
            --runtime=nodejs18 \
            --region=$region \
            --source=cloud-function \
            --trigger-http \
            --allow-unauthenticated \
            --entry-point=mdnHandler \
            --concurrency=100 \
            --min-instances=1 \
            --max-instances=100 \
            --memory=2GB \
            --timeout=120s \
            --set-env-vars="ORIGIN_MAIN=developer.allizom.org" \
            --set-env-vars="ORIGIN_LIVE_SAMPLES=live.mdnyalp.dev" \
            --set-env-vars="ORIGIN_PLAY=mdnyalp.dev" \
            --set-env-vars="SOURCE_CONTENT=https://storage.googleapis.com/${{ vars.GCP_BUCKET_NAME }}/main/" \
            --set-env-vars="SOURCE_API=https://api.developer.allizom.org/" \
            --set-env-vars="BSA_ENABLED=true" \
            --set-env-vars="SENTRY_DSN=${{ secrets.SENTRY_DSN_CLOUD_FUNCTION }}" \
            --set-env-vars="SENTRY_ENVIRONMENT=stage" \
            --set-env-vars="SENTRY_TRACES_SAMPLE_RATE=${{ vars.SENTRY_TRACES_SAMPLE_RATE }}" \
            --set-env-vars="SENTRY_RELEASE=${{ github.sha }}" \
            --set-secrets="KEVEL_SITE_ID=projects/${{ secrets.GCP_PROJECT_NAME }}/secrets/stage-kevel-site-id/versions/latest" \
            --set-secrets="KEVEL_NETWORK_ID=projects/${{ secrets.GCP_PROJECT_NAME }}/secrets/stage-kevel-network-id/versions/latest" \
            --set-secrets="SIGN_SECRET=projects/${{ secrets.GCP_PROJECT_NAME }}/secrets/stage-sign-secret/versions/latest" \
            --set-secrets="BSA_ZONE_KEYS=projects/${{ secrets.GCP_PROJECT_NAME }}/secrets/stage-bsa-zone-keys/versions/latest" \
            2>&1 | sed "s/^/[$region] /" &
            pids+=($!)
          done

          for pid in "${pids[@]}"; do
            wait $pid
          done

      - name: Update AI Help index
        run: yarn ai-help update-index
        env:
          CONTENT_ROOT: ${{ github.workspace }}/mdn/content/files
          OPENAI_KEY: ${{ secrets.OPENAI_KEY }}
          SUPABASE_SERVICE_ROLE_KEY: ${{ secrets.SUPABASE_SERVICE_ROLE_KEY }}
          SUPABASE_URL: ${{ secrets.SUPABASE_URL }}

      - name: Update AI Help index with macros
        run: yarn ai-help-macros update-index
        env:
          OPENAI_KEY: ${{ secrets.OPENAI_KEY }}
          PG_URI: ${{ secrets.PG_URI }}

      - name: Slack Notification
        if: failure()
        uses: rtCamp/action-slack-notify@v2
        env:
          SLACK_CHANNEL: mdn-notifications
          SLACK_COLOR: ${{ job.status }}
          SLACK_ICON: https://avatars.slack-edge.com/2020-11-17/1513880588420_fedd7f0e9456888e69ff_96.png
          SLACK_TITLE: "Stage"
          SLACK_MESSAGE: "Build failed :collision:"
          SLACK_FOOTER: "Powered by stage-build.yml"
          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

      - name: Invalidate CDN
        if: ${{ github.event.inputs.invalidate }}
        run: gcloud compute url-maps invalidate-cdn-cache ${{ secrets.GCP_LOAD_BALANCER_NAME }} --path "/*" --async