Build the catalog image from a clean checkout

.github/workflows/deploy-catalog.yaml

@@ -1,4 +1,4 @@
-name: Deploy catalog to ECR
+name: "Catalog: build & push"
 
 on:
   push:
@@ -8,61 +8,87 @@ on:
       - '.github/workflows/deploy-catalog.yaml'
       - 'catalog/**'
       - 'shared/**'
+  pull_request:
+    paths:
+      - '.github/workflows/deploy-catalog.yaml'
+      - 'catalog/**'
+      - 'shared/**'
 
 jobs:
-  deploy-catalog-ecr:
+  build:
     runs-on: ubuntu-latest
-    defaults:
-      run:
-        working-directory: catalog
-    # These permissions are needed to interact with GitHub's OIDC Token endpoint.
     permissions:
-      id-token: write
       contents: read
     steps:
       - uses: actions/checkout@v7
-      - uses: actions/setup-node@v7
+      - uses: docker/setup-buildx-action@v4
+      - name: Build image
+        uses: docker/build-push-action@v7
+        with:
+          context: .
+          file: catalog/Dockerfile
+          tags: catalog:${{ github.sha }}
+          push: false
+          # On master, export a tarball for the push jobs below; on pull
+          # requests the build itself is the check and nothing consumes it.
+          outputs: ${{ github.event_name == 'push' && 'type=docker,dest=/tmp/catalog-image.tar' || '' }}
+          # One cache scope for both events: master builds warm the cache PRs
+          # read (the reverse is impossible — PR caches are isolated).
+          # ignore-error: fork PRs can't write the cache; worst case anywhere
+          # is a colder next build.
+          cache-from: type=gha,scope=catalog
+          cache-to: type=gha,mode=max,scope=catalog,ignore-error=true
+      - uses: actions/upload-artifact@v7
+        if: github.event_name == 'push'
         with:
-          node-version-file: 'catalog/package.json'
-          cache: 'npm'
-          cache-dependency-path: 'catalog/package-lock.json'
-      - run: npm ci
-      - run: npm run build
-      - name: Configure AWS credentials from Prod account
-        uses: aws-actions/configure-aws-credentials@v6
+          name: catalog-image
+          path: /tmp/catalog-image.tar
+          retention-days: 1
+
+  push:
+    if: github.event_name == 'push'
+    needs: build
+    runs-on: ubuntu-latest
+    # These permissions are needed to interact with GitHub's OIDC Token endpoint.
+    permissions:
+      id-token: write
+    strategy:
+      # Targets are independent, so a failed one is retried alone via "Re-run
+      # failed jobs" — necessary for mp, which can't be re-pushed under the
+      # same tag, so any re-run replaying its successful push would fail.
+      fail-fast: false
+      matrix:
+        include:
+          - target: prod
+            role: arn:aws:iam::730278974607:role/github/GitHub-Quilt
+            region: us-east-1
+            repository: quiltdata/catalog
+          - target: mp
+            role: arn:aws:iam::730278974607:role/github/GitHub-Quilt
+            region: us-east-1
+            registry_id: "709825985650"
+            repository: quilt-data/quilt-payg-catalog
+          - target: govcloud
+            role: arn:aws-us-gov:iam::313325871032:role/github/GitHub-Quilt
+            region: us-gov-east-1
+            repository: quiltdata/catalog
+    steps:
+      - uses: actions/download-artifact@v8
         with:
-          role-to-assume: arn:aws:iam::730278974607:role/github/GitHub-Quilt
-          aws-region: us-east-1
-      - name: Login to Prod ECR
-        id: login-prod-ecr
-        uses: aws-actions/amazon-ecr-login@v2
-      - name: Login to MP ECR
-        id: login-mp-ecr
-        uses: aws-actions/amazon-ecr-login@v2
+          name: catalog-image
+          path: /tmp
+      - run: docker load -i /tmp/catalog-image.tar
+      - uses: aws-actions/configure-aws-credentials@v6
         with:
-          registries: 709825985650
-      - name: Configure AWS credentials from GovCloud account
-        uses: aws-actions/configure-aws-credentials@v6
+          role-to-assume: ${{ matrix.role }}
+          aws-region: ${{ matrix.region }}
+      - uses: aws-actions/amazon-ecr-login@v2
+        id: ecr
         with:
-          role-to-assume: arn:aws-us-gov:iam::313325871032:role/github/GitHub-Quilt
-          aws-region: us-gov-east-1
-      - name: Login to GovCloud ECR
-        id: login-govcloud-ecr
-        uses: aws-actions/amazon-ecr-login@v2
-      - name: Build and push Docker image to Prod, MP and GovCloud ECR
+          registries: ${{ matrix.registry_id }}
+      - name: Tag and push
         env:
-          ECR_REGISTRY_PROD: ${{ steps.login-prod-ecr.outputs.registry }}
-          ECR_REGISTRY_GOVCLOUD: ${{ steps.login-govcloud-ecr.outputs.registry }}
-          ECR_REGISTRY_MP: ${{ steps.login-mp-ecr.outputs.registry }}
-          ECR_REPOSITORY: quiltdata/catalog
-          ECR_REPOSITORY_MP: quilt-data/quilt-payg-catalog
-          IMAGE_TAG: ${{ github.sha }}
+          IMAGE: ${{ steps.ecr.outputs.registry }}/${{ matrix.repository }}:${{ github.sha }}
         run: |
-          docker buildx build \
-            -t $ECR_REGISTRY_PROD/$ECR_REPOSITORY:$IMAGE_TAG \
-            -t $ECR_REGISTRY_GOVCLOUD/$ECR_REPOSITORY:$IMAGE_TAG \
-            -t $ECR_REGISTRY_MP/$ECR_REPOSITORY_MP:$IMAGE_TAG \
-            .
-          docker push $ECR_REGISTRY_PROD/$ECR_REPOSITORY:$IMAGE_TAG
-          docker push $ECR_REGISTRY_GOVCLOUD/$ECR_REPOSITORY:$IMAGE_TAG
-          docker push $ECR_REGISTRY_MP/$ECR_REPOSITORY_MP:$IMAGE_TAG
+          docker tag catalog:${{ github.sha }} "$IMAGE"
+          docker push "$IMAGE"

catalog/Dockerfile

@@ -1,5 +1,19 @@
+# Build context is the repo root, not catalog/ — the webpack build resolves
+# modules from ../shared (see internals/webpack/webpack.base.js).
+FROM node:26-trixie-slim@sha256:715e55e4b84e4bb0ff48e49b398a848f08e55daed8eb6a0ea1839ae53bc57583 AS build
+
+WORKDIR /src/catalog
+
+# Manifests only, so editing app sources doesn't invalidate `npm ci`.
+COPY catalog/package.json catalog/package-lock.json ./
+RUN --mount=type=cache,target=/root/.npm npm ci
+
+COPY shared /src/shared
+COPY catalog ./
+RUN npm run build
+
 FROM amazonlinux:2023.11.20260406.2
 MAINTAINER Quilt Data, Inc. support@quilt.bio
 
 ENV LC_ALL=C.UTF-8
 ENV LANG=C.UTF-8
@@ -13,16 +27,15 @@
     ln -sf /dev/stderr /var/log/nginx/error.log
 
 # Set up nginx
-COPY nginx.conf /etc/nginx/nginx.conf
-COPY nginx-web.conf /etc/nginx/conf.d/default.conf
+COPY catalog/nginx.conf /etc/nginx/nginx.conf
+COPY catalog/nginx-web.conf /etc/nginx/conf.d/default.conf
 
 ARG NGINX_STATIC_DIR=/usr/share/nginx/html
-# Copy pre-built catalog assets to nginx
 RUN rm -rf $NGINX_STATIC_DIR
-COPY build $NGINX_STATIC_DIR
+COPY --from=build /src/catalog/build $NGINX_STATIC_DIR
 
 # Copy config file
-COPY config.json.tmpl config.json.tmpl
+COPY catalog/config.json.tmpl config.json.tmpl
 
 RUN ln -s /tmp/config.json $NGINX_STATIC_DIR/config.json && \
     ln -s /tmp/config.js $NGINX_STATIC_DIR/config.js
@@ -32,6 +45,6 @@
 
 # Substitute environment variables into config.json and generate config.js based on that before starting nginx.
 # Note: use "exec" because otherwise the shell will catch Ctrl-C and other signals.
 CMD envsubst < config.json.tmpl > /tmp/config.json \
     && echo "window.QUILT_CATALOG_CONFIG = `cat /tmp/config.json`" > /tmp/config.js \
     && exec nginx -g 'daemon off;'

catalog/Dockerfile.dockerignore

@@ -0,0 +1,17 @@
+# Scoped to catalog/Dockerfile (BuildKit reads <dockerfile>.dockerignore), so
+# other images' build contexts are unaffected.
+#
+# The context is the repo root, but the build only needs catalog/ and shared/.
+# Allow-list rather than deny-list: a new top-level directory stays out by
+# default instead of silently bloating the context.
+*
+!catalog
+!shared
+
+# Local artifacts that would otherwise leak in from a dirty working tree.
+# catalog/build especially: a stale one must not shadow the in-image build.
+catalog/node_modules
+catalog/build
+catalog/coverage
+catalog/stats.json
+**/.DS_Store