ci: publish distroless variants and rebuild images weekly

Extract the release build/push and manifest jobs into a reusable
build-images.yml workflow, extend it with a distroless variant, and add
a weekly rebuild-images.yml so digest-pinned bases pick up security
updates without a manual re-pin.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

.github/workflows/build-images.yml

@@ -0,0 +1,166 @@
+name: Build and Push Images
+
+# Reusable image build. Called by release.yml when cutting a release and by
+# rebuild-images.yml on a schedule, so that published images pick up base image
+# security updates without waiting for the next release.
+
+on:
+  workflow_call:
+    inputs:
+      version:
+        description: Tag applied to the analyzer, anonymizer and CLI-adjacent images.
+        required: true
+        type: string
+      image-version:
+        description: Tag applied to the image-redactor image.
+        required: true
+        type: string
+      date-tag:
+        description: >
+          Optional immutable suffix, e.g. 20260909. When set, each image is also
+          pushed as <version>-<date-tag> so a specific rebuild can be pinned or
+          rolled back to.
+        required: false
+        default: ''
+        type: string
+
+permissions: read-all
+
+env:
+  IMAGE_REGISTRY: ghcr.io
+  IMAGE_NAMESPACE: data-privacy-stack
+
+jobs:
+  build-and-push-containers:
+    name: Build and Push ${{ matrix.image }}${{ matrix.tag_suffix }} (${{ matrix.platform }})
+    runs-on: ${{ matrix.platform == 'linux/arm64' && 'ubuntu-24.04-arm' || 'ubuntu-latest' }}
+    permissions:
+      contents: read
+      packages: write
+      id-token: write
+    strategy:
+      fail-fast: false
+      matrix:
+        image: [presidio-anonymizer, presidio-analyzer, presidio-image-redactor]
+        platform: [linux/amd64, linux/arm64]
+        variant: [default, distroless]
+        exclude:
+          # No distroless variant is published for the image-redactor: it
+          # depends on the Tesseract OCR system packages, which are not
+          # available in a distroless base image.
+          - image: presidio-image-redactor
+            variant: distroless
+        include:
+          - platform: linux/amd64
+            platform_tag: linux-amd64
+          - platform: linux/arm64
+            platform_tag: linux-arm64
+          - variant: default
+            dockerfile: Dockerfile
+            tag_suffix: ''
+          - variant: distroless
+            dockerfile: Dockerfile.distroless
+            tag_suffix: -distroless
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.0
+        with:
+          persist-credentials: false
+
+      - name: Log in to GitHub Container Registry
+        env:
+          GHCR_TOKEN: ${{ github.repository_owner == 'data-privacy-stack' && github.token || secrets.GHCR_TOKEN }}
+          GHCR_USERNAME: ${{ github.repository_owner == 'data-privacy-stack' && github.actor || vars.GHCR_USERNAME || github.actor }}
+        run: echo "${GHCR_TOKEN}" | docker login ${{ env.IMAGE_REGISTRY }} -u "${GHCR_USERNAME}" --password-stdin
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@4d04d5d9486b7bd6fa91e7baf45bbb4f8b9deedd # v4.0.0
+
+      - name: Build and Push ${{ matrix.image }}${{ matrix.tag_suffix }} for ${{ matrix.platform }}
+        run: |
+          set -eu
+          repo="${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/${{ matrix.image }}"
+          version_tag="${VERSION}"
+          if [ "${{ matrix.image }}" = "presidio-image-redactor" ]; then
+            version_tag="${IMAGE_VERSION}"
+          fi
+          suffix="${{ matrix.tag_suffix }}"
+          platform_tag="${{ matrix.platform_tag }}"
+
+          tag_args="--tag ${repo}:latest${suffix}-${platform_tag}"
+          tag_args="${tag_args} --tag ${repo}:${version_tag}${suffix}-${platform_tag}"
+          if [ -n "${DATE_TAG}" ]; then
+            tag_args="${tag_args} --tag ${repo}:${version_tag}-${DATE_TAG}${suffix}-${platform_tag}"
+          fi
+
+          # shellcheck disable=SC2086
+          docker buildx build \
+            --platform ${{ matrix.platform }} \
+            --file ./${{ matrix.image }}/${{ matrix.dockerfile }} \
+            --cache-from type=registry,ref=${repo}:latest${suffix} \
+            --cache-to type=inline \
+            ${tag_args} \
+            --attest type=sbom \
+            --attest type=provenance,mode=max \
+            --push \
+            ./${{ matrix.image }}
+        env:
+          VERSION: ${{ inputs.version }}
+          IMAGE_VERSION: ${{ inputs.image-version }}
+          DATE_TAG: ${{ inputs.date-tag }}
+
+  create-container-manifests:
+    name: Create Multi-Platform Container Manifests
+    runs-on: ubuntu-latest
+    needs: build-and-push-containers
+    permissions:
+      contents: read
+      packages: write
+    steps:
+      - name: Log in to GitHub Container Registry
+        env:
+          GHCR_TOKEN: ${{ github.repository_owner == 'data-privacy-stack' && github.token || secrets.GHCR_TOKEN }}
+          GHCR_USERNAME: ${{ github.repository_owner == 'data-privacy-stack' && github.actor || vars.GHCR_USERNAME || github.actor }}
+        run: echo "${GHCR_TOKEN}" | docker login ${{ env.IMAGE_REGISTRY }} -u "${GHCR_USERNAME}" --password-stdin
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@4d04d5d9486b7bd6fa91e7baf45bbb4f8b9deedd # v4.0.0
+
+      - name: Create multi-platform manifests
+        run: |
+          set -eu
+          images=(presidio-anonymizer presidio-analyzer presidio-image-redactor)
+          platform_tags=(linux-amd64 linux-arm64)
+
+          for image in "${images[@]}"; do
+            repo="${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/${image}"
+            version_tag="${VERSION}"
+            suffixes=("" "-distroless")
+            if [ "${image}" = "presidio-image-redactor" ]; then
+              version_tag="${IMAGE_VERSION}"
+              suffixes=("")
+            fi
+
+            for suffix in "${suffixes[@]}"; do
+              tags=("latest${suffix}" "${version_tag}${suffix}")
+              if [ -n "${DATE_TAG}" ]; then
+                tags+=("${version_tag}-${DATE_TAG}${suffix}")
+              fi
+
+              for tag in "${tags[@]}"; do
+                platform_refs=()
+                for platform_tag in "${platform_tags[@]}"; do
+                  platform_refs+=("${repo}:${tag}-${platform_tag}")
+                done
+
+                docker buildx imagetools create \
+                  --tag "${repo}:${tag}" \
+                  "${platform_refs[@]}"
+              done
+            done
+          done
+        env:
+          VERSION: ${{ inputs.version }}
+          IMAGE_VERSION: ${{ inputs.image-version }}
+          DATE_TAG: ${{ inputs.date-tag }}

.github/workflows/rebuild-images.yml

@@ -0,0 +1,71 @@
+name: Rebuild Images
+
+# Published images pin their base image by digest, so they never pick up
+# operating system security updates on their own. Rebuilding on a schedule
+# republishes the same versions on top of a freshly patched base, which is what
+# container compliance scanners expect ("redeploy the image", "update the
+# vulnerable image reference").
+#
+# Rebuilds do not change any application code: the workflow builds the current
+# state of main and republishes the `latest` and current version tags, plus an
+# immutable `<version>-<date>` tag so a specific rebuild can be pinned.
+
+on:
+  schedule:
+    # Every Monday at 05:00 UTC.
+    - cron: '0 5 * * 1'
+  workflow_dispatch:
+
+permissions: read-all
+
+jobs:
+  get-version:
+    name: Get Version Numbers
+    runs-on: ubuntu-slim
+    outputs:
+      version: ${{ steps.set-version.outputs.version }}
+      image-version: ${{ steps.set-image-version.outputs.image-version }}
+      date-tag: ${{ steps.set-date-tag.outputs.date-tag }}
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.0
+        with:
+          persist-credentials: false
+
+      - name: Extract main version
+        id: set-version
+        run: |
+          set -eu
+          ver=$(grep -m 1 version presidio-analyzer/pyproject.toml | tr -s ' ' | tr -d '"' | tr -d "'" | cut -d' ' -f3)
+          echo "version=$ver" >> $GITHUB_OUTPUT
+          echo "Main version: $ver"
+
+      - name: Extract image-redactor version
+        id: set-image-version
+        run: |
+          set -eu
+          imageVer=$(grep -m 1 version presidio-image-redactor/pyproject.toml | tr -s ' ' | tr -d '"' | tr -d "'" | cut -d' ' -f3)
+          echo "image-version=$imageVer" >> $GITHUB_OUTPUT
+          echo "Image-redactor version: $imageVer"
+
+      - name: Set date tag
+        id: set-date-tag
+        run: |
+          set -eu
+          dateTag=$(date -u +'%Y%m%d')
+          echo "date-tag=$dateTag" >> $GITHUB_OUTPUT
+          echo "Date tag: $dateTag"
+
+  rebuild-and-push-containers:
+    name: Rebuild and Push Container Images
+    needs: get-version
+    permissions:
+      contents: read
+      packages: write
+      id-token: write
+    uses: ./.github/workflows/build-images.yml
+    with:
+      version: ${{ needs.get-version.outputs.version }}
+      image-version: ${{ needs.get-version.outputs.image-version }}
+      date-tag: ${{ needs.get-version.outputs.date-tag }}
+    secrets: inherit

.github/workflows/release.yml

@@ -5,10 +5,6 @@ on:
 
 permissions: read-all
 
-env:
-  IMAGE_REGISTRY: ghcr.io
-  IMAGE_NAMESPACE: data-privacy-stack
-
 jobs:
   get-version:
     name: Get Version Numbers
@@ -114,101 +110,14 @@ jobs:
           skip-existing: true
 
   build-and-push-containers:
-    name: Build and Push ${{ matrix.image }} (${{ matrix.platform }})
+    name: Build and Push Container Images
-    runs-on: ${{ matrix.platform == 'linux/arm64' && 'ubuntu-24.04-arm' || 'ubuntu-latest' }}
     needs: get-version
     permissions:
       contents: read
       packages: write
       id-token: write
-    strategy:
+    uses: ./.github/workflows/build-images.yml
-      fail-fast: false
+    with:
-      matrix:
+      version: ${{ needs.get-version.outputs.version }}
-        image: [presidio-anonymizer, presidio-analyzer, presidio-image-redactor]
+      image-version: ${{ needs.get-version.outputs.image-version }}
-        platform: [linux/amd64, linux/arm64]
+    secrets: inherit
-        include:
-          - platform: linux/amd64
-            platform_tag: linux-amd64
-          - platform: linux/arm64
-            platform_tag: linux-arm64
-
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.0
-        with:
-          persist-credentials: false
-
-      - name: Log in to GitHub Container Registry
-        env:
-          GHCR_TOKEN: ${{ github.repository_owner == 'data-privacy-stack' && github.token || secrets.GHCR_TOKEN }}
-          GHCR_USERNAME: ${{ github.repository_owner == 'data-privacy-stack' && github.actor || vars.GHCR_USERNAME || github.actor }}
-        run: echo "${GHCR_TOKEN}" | docker login ${{ env.IMAGE_REGISTRY }} -u "${GHCR_USERNAME}" --password-stdin
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@4d04d5d9486b7bd6fa91e7baf45bbb4f8b9deedd # v4.0.0
-
-      - name: Build and Push ${{ matrix.image }} for ${{ matrix.platform }}
-        run: |
-          repo="${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/${{ matrix.image }}"
-          version_tag="${VERSION}"
-          if [ "${{ matrix.image }}" = "presidio-image-redactor" ]; then
-            version_tag="${IMAGE_VERSION}"
-          fi
-
-          docker buildx build \
-            --platform ${{ matrix.platform }} \
-            --cache-from type=registry,ref=${repo}:latest \
-            --cache-to type=inline \
-            --tag "${repo}:latest-${{ matrix.platform_tag }}" \
-            --tag "${repo}:${version_tag}-${{ matrix.platform_tag }}" \
-            --attest type=sbom \
-            --attest type=provenance,mode=max \
-            --push \
-            ./${{ matrix.image }}
-        env:
-          VERSION: ${{ needs.get-version.outputs.version }}
-          IMAGE_VERSION: ${{ needs.get-version.outputs.image-version }}
-
-  create-container-manifests:
-    name: Create Multi-Platform Container Manifests
-    runs-on: ubuntu-latest
-    needs: [get-version, build-and-push-containers]
-    permissions:
-      contents: read
-      packages: write
-    steps:
-      - name: Log in to GitHub Container Registry
-        env:
-          GHCR_TOKEN: ${{ github.repository_owner == 'data-privacy-stack' && github.token || secrets.GHCR_TOKEN }}
-          GHCR_USERNAME: ${{ github.repository_owner == 'data-privacy-stack' && github.actor || vars.GHCR_USERNAME || github.actor }}
-        run: echo "${GHCR_TOKEN}" | docker login ${{ env.IMAGE_REGISTRY }} -u "${GHCR_USERNAME}" --password-stdin
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@4d04d5d9486b7bd6fa91e7baf45bbb4f8b9deedd # v4.0.0
-
-      - name: Create multi-platform manifests
-        run: |
-          images=(presidio-anonymizer presidio-analyzer presidio-image-redactor)
-          platform_tags=(linux-amd64 linux-arm64)
-
-          for image in "${images[@]}"; do
-            repo="${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/${image}"
-            version_tag="${VERSION}"
-            if [ "${image}" = "presidio-image-redactor" ]; then
-              version_tag="${IMAGE_VERSION}"
-            fi
-
-            for tag in latest "${version_tag}"; do
-              platform_refs=()
-              for platform_tag in "${platform_tags[@]}"; do
-                platform_refs+=("${repo}:${tag}-${platform_tag}")
-              done
-
-              docker buildx imagetools create \
-                --tag "${repo}:${tag}" \
-                "${platform_refs[@]}"
-            done
-          done
-        env:
-          VERSION: ${{ needs.get-version.outputs.version }}
-          IMAGE_VERSION: ${{ needs.get-version.outputs.image-version }}