Implement fast bazel target lookup with index caching

.github/workflows/bazel.yml

@@ -71,6 +71,11 @@ on:
         required: false
         type: number
         default: 1
+      cache-name:
+        description: Name for cache restore (restores {name}.gz with key {name}-)
+        required: false
+        type: string
+        default: ''
 
 jobs:
   bazel:
@@ -100,6 +105,13 @@ jobs:
         run: git pull origin "$GIT_REF"
         env:
           GIT_REF: ${{ github.ref }}
+      - name: Restore cache
+        if: inputs.cache-name != ''
+        uses: actions/cache/restore@v4
+        with:
+          path: ${{ inputs.cache-name }}
+          key: ${{ inputs.cache-name }}-
+          restore-keys: ${{ inputs.cache-name }}-
       - name: Free space
         if: inputs.os != 'windows'
         run: ./scripts/github-actions/free-disk-space.sh

.github/workflows/ci-build-index.yml

@@ -12,7 +12,7 @@ jobs:
     uses: ./.github/workflows/bazel.yml
     with:
       name: Build Test Index
-      os: macos
+      os: ubuntu
       run: ./go bazel:build_test_index bazel-test-target-index
       artifact-name: bazel-test-target-index
       artifact-path: bazel-test-target-index
@@ -21,6 +21,7 @@ jobs:
     name: Cache Index
     needs: build
     runs-on: ubuntu-latest
+    permissions: {}
     steps:
       - name: Download index
         uses: actions/download-artifact@v4

.github/workflows/ci.yml

@@ -20,7 +20,7 @@ permissions:
 jobs:
   build-index:
     name: Build Index
-    if: github.event_name == 'schedule'
+    if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
     uses: ./.github/workflows/ci-build-index.yml
 
   check:
@@ -31,8 +31,22 @@ jobs:
     uses: ./.github/workflows/bazel.yml
     with:
       name: Check Targets
-      run: COMMIT_RANGE="${{ github.event.pull_request.base.sha || github.event.before }}...${{ github.event.pull_request.head.sha || github.sha }}" ./scripts/github-actions/check-bazel-targets.sh
+      cache-name: bazel-test-target-index
       fetch-depth: 50
+      run: |
+        if [ "${{ github.event_name }}" == "schedule" ] || \
+        [ "${{ github.event_name }}" == "workflow_call" ] || \
+        [ "${{ github.event_name }}" == "workflow_dispatch" ]; then
+          echo "Running all targets for ${{ github.event_name }} event"
+          echo "//java/... //py:* //rb/... //dotnet/... //rust/..." > bazel-targets.txt
+        else
+          BASE_SHA="${{ github.event.pull_request.base.sha || github.event.before }}"
+          if [ -n "$BASE_SHA" ]; then
+            ./go bazel:affected_targets "${BASE_SHA}..HEAD" bazel-test-target-index
+          else
+            ./go bazel:affected_targets bazel-test-target-index
+          fi
+        fi
       artifact-name: check-targets
       artifact-path: bazel-targets.txt
 
@@ -41,83 +55,62 @@ jobs:
     needs: check
     runs-on: ubuntu-latest
     outputs:
-      targets: ${{ steps.read.outputs.targets }}
+      java: ${{ steps.read.outputs.java }}
+      py: ${{ steps.read.outputs.py }}
+      rb: ${{ steps.read.outputs.rb }}
+      dotnet: ${{ steps.read.outputs.dotnet }}
+      rust: ${{ steps.read.outputs.rust }}
     steps:
       - name: Download targets
         uses: actions/download-artifact@v4
         with:
           name: check-targets
       - name: Read targets
         id: read
+        env:
+          COMMIT_MESSAGES: ${{ join(github.event.commits.*.message, ' ') }}
+          PR_TITLE: ${{ github.event.pull_request.title }}
         run: |
-          if [ -s bazel-targets.txt ]; then
-            {
-              echo "targets<<EOF"
-              cat bazel-targets.txt
-              echo "EOF"
-            } >> "$GITHUB_OUTPUT"
-          else
-            echo "targets=" >> "$GITHUB_OUTPUT"
-          fi
-
+          targets=$(cat bazel-targets.txt 2>/dev/null || echo "")
+          check_binding() {
+            local pattern=$1 tag=$2
+            if [[ "$targets" == *"$pattern"* ]] || [[ "$COMMIT_MESSAGES" == *"[$tag]"* ]] || [[ "$PR_TITLE" == *"[$tag]"* ]]; then
+              echo "$tag=true" >> "$GITHUB_OUTPUT"
+            fi
+          }
+          check_binding "//java/" "java"
+          check_binding "//py:" "py"
+          check_binding "//rb/" "rb"
+          check_binding "//dotnet/" "dotnet"
+          check_binding "//rust/" "rust"
   dotnet:
     name: .NET
     needs: read-targets
     uses: ./.github/workflows/ci-dotnet.yml
-    if: >
-      github.event_name == 'schedule' ||
-      github.event_name == 'workflow_dispatch' ||
-      github.event_name == 'workflow_call' ||
-      contains(needs.read-targets.outputs.targets, '//dotnet') ||
-      contains(join(github.event.commits.*.message), '[dotnet]') ||
-      contains(github.event.pull_request.title, '[dotnet]')
+    if: needs.read-targets.outputs.dotnet != ''
 
   java:
     name: Java
     needs: read-targets
     uses: ./.github/workflows/ci-java.yml
-    if: >
-      github.event_name == 'schedule' ||
-      github.event_name == 'workflow_dispatch' ||
-      github.event_name == 'workflow_call' ||
-      contains(needs.read-targets.outputs.targets, '//java') ||
-      contains(join(github.event.commits.*.message), '[java]') ||
-      contains(github.event.pull_request.title, '[java]')
+    if: needs.read-targets.outputs.java != ''
 
   python:
     name: Python
     needs: read-targets
     uses: ./.github/workflows/ci-python.yml
-    if: >
-      github.event_name == 'schedule' ||
-      github.event_name == 'workflow_dispatch' ||
-      github.event_name == 'workflow_call' ||
-      contains(needs.read-targets.outputs.targets, '//py') ||
-      contains(join(github.event.commits.*.message), '[py]') ||
-      contains(github.event.pull_request.title, '[py]')
+    if: needs.read-targets.outputs.py != ''
 
   ruby:
     name: Ruby
     needs: read-targets
     uses: ./.github/workflows/ci-ruby.yml
-    if: >
-      github.event_name == 'schedule' ||
-      github.event_name == 'workflow_dispatch' ||
-      github.event_name == 'workflow_call' ||
-      contains(needs.read-targets.outputs.targets, '//rb') ||
-      contains(join(github.event.commits.*.message), '[rb]') ||
-      contains(github.event.pull_request.title, '[rb]')
+    if: needs.read-targets.outputs.rb != ''
 
   rust:
     name: Rust
     needs: read-targets
     uses: ./.github/workflows/ci-rust.yml
     secrets:
       SELENIUM_CI_TOKEN: ${{ secrets.SELENIUM_CI_TOKEN }}
-    if: >
-      github.event_name == 'schedule' ||
-      github.event_name == 'workflow_dispatch' ||
-      github.event_name == 'workflow_call' ||
-      contains(needs.read-targets.outputs.targets, '//rust') ||
-      contains(join(github.event.commits.*.message), '[rust]') ||
-      contains(github.event.pull_request.title, '[rust]')
+    if: needs.read-targets.outputs.rust != ''

rake_tasks/bazel.rake

@@ -1,12 +1,52 @@
 # frozen_string_literal: true
 
 require 'json'
+require 'set'
 
-# ./go bazel:build_test_index            --> 'build/bazel-test-target-index.json'
-# ./go bazel:build_test_index index.json --> 'index.json'
+# ./go bazel:affected_targets                              --> HEAD^..HEAD with default index
+# ./go bazel:affected_targets abc123..def456               --> explicit range
+# ./go bazel:affected_targets abc123..def456 my-index      --> explicit range with custom index
+# ./go bazel:affected_targets my-index                     --> HEAD^..HEAD with custom index
+desc 'Find test targets affected by changes between revisions'
+task :affected_targets do |_task, args|
+  values = args.to_a
+  index_file = values.find { |value| File.exist?(value) }
+  range = (values - [index_file]).first || 'HEAD'
+  index_file ||= 'build/bazel-test-target-index'
+
+  base_rev, head_rev = if range.include?('..')
+                         range.split('..', 2)
+                       else
+                         ["#{range}^", range]
+                       end
+
+  puts "Commit range: #{base_rev}..#{head_rev}"
+
+  changed_files = `git diff --name-only #{base_rev} #{head_rev}`.split("\n").map(&:strip).reject(&:empty?)
+  puts "Changed files: #{changed_files.size}"
+
+  targets = if File.exist?(index_file)
+              affected_targets_with_index(changed_files, index_file)
+            else
+              puts 'No index found, using directory-based fallback'
+              affected_targets_fallback(changed_files)
+            end
+
+  if targets.empty?
+    puts 'No test targets affected'
+    File.write('bazel-targets.txt', '')
+  else
+    puts "Found #{targets.size} affected test targets"
+    File.write('bazel-targets.txt', targets.sort.join(' '))
+    targets.sort.each { |t| puts t }
+  end
+end
+
+# ./go bazel:build_test_index                    --> 'build/bazel-test-target-index'
+# ./go bazel:build_test_index my-index           --> 'my-index'
 desc 'Build test target index for faster affected target lookup'
 task :build_test_index, [:index_file] do |_task, args|
-  output = args[:index_file] || 'build/bazel-test-target-index.json'
+  output = args[:index_file] || 'build/bazel-test-target-index'
 
   index = {}
   tests = []
@@ -53,3 +93,96 @@ def bazel_label_to_package(label)
   label = label.sub(%r{^@selenium//}, '').sub(%r{^//}, '')
   label.split(':').first
 end
+
+def find_bazel_package(filepath)
+  path = File.dirname(filepath)
+  until path.empty?
+    return path if File.exist?(File.join(path, 'BUILD.bazel')) || File.exist?(File.join(path, 'BUILD'))
+    return nil if path == '.'
+
+    path = File.dirname(path)
+  end
+  nil
+end
+
+def affected_targets_with_index(changed_files, index_file)
+  puts "Using index: #{index_file}"
+  begin
+    index = JSON.parse(File.read(index_file))
+  rescue JSON::ParserError => e
+    puts "Invalid JSON in index file: #{e.message}"
+    return affected_targets_fallback(changed_files)
+  end
+
+  test_files, lib_files = changed_files.partition { |f| f.match?(%r{[_-]test\.rb$|_test\.py$|Test\.java$|Tests?\.cs$|\.test\.[jt]s$|_spec\.rb$}) }
+
+  affected = Set.new
+  affected.merge(targets_from_tests(test_files))
+
+  lib_files.each do |filepath|
+    pkg = find_bazel_package(filepath)
+    affected.merge(targets_from_lookup(pkg, index, filepath))
+  end
+
+  affected.to_a
+end
+
+def targets_from_lookup(pkg, index, filepath)
+  # ignore files not associated with bazel package
+  return [] if pkg.nil?
+
+  # Root package is empty string, not '.'
+  pkg = '' if pkg == '.'
+
+  # generate targets if package not in the index
+  test_targets = index[pkg] || query_package_dep(pkg)
+
+  # dotnet tests depend on java server, but there are no remote tests, so safe to ignore
+  filepath.start_with?('java/') ? test_targets.reject { |t| t.start_with?('//dotnet/') } : test_targets
+end
+
+def query_package_dep(pkg)
+  # Root package is empty string, not '.'
+  pkg = '' if pkg == '.'
+  puts "Package not in index, querying deps: //#{pkg}"
+  targets = []
+  Bazel.execute('query', ['--output=label'], "kind('.*_test', deps(//#{pkg}:all))") do |out|
+    targets = out.lines.map(&:strip).select { |l| l.start_with?('//') }
+  end
+  targets
+end
+
+def targets_from_tests(test_files)
+  test_files.select! { |f| File.exist?(f) }
+  return [] if test_files.empty?
+
+  query = test_files.filter_map { |f|
+    pkg = find_bazel_package(f)
+    next if pkg.nil?
+
+    # Bazel srcs often use paths relative to the package, not basenames.
+    rel = f.sub(%r{^#{Regexp.escape(pkg)}/}, '')
+    "attr(srcs, '#{rel}', //#{pkg}:*)"
+  }.join(' + ')
+
+  return [] if query.empty?
+
+  targets = []
+  Bazel.execute('query', ['--output=label'], "kind('.*_test', #{query})") do |out|
+    targets = out.lines.map(&:strip).select { |l| l.start_with?('//') }
+  end
+  targets
+end
+
+def affected_targets_fallback(changed_files)
+  targets = Set.new
+  top_level_dirs = changed_files.map { |f| f.split('/').first }.uniq
+
+  return BINDING_TARGETS.values if top_level_dirs.intersect?(%w[common rust])
+
+  top_level_dirs.each do |dir|
+    targets << BINDING_TARGETS[dir] if BINDING_TARGETS[dir]
+  end
+
+  targets.to_a
+end