WIP

.github/workflows/.yamllint

@@ -2,6 +2,8 @@
 extends: default
 
 rules:
+  document-start:
+    present: false
   line-length: disable
   truthy:
     check-keys: false

.github/workflows/docker-publish.yml

@@ -50,7 +50,6 @@ jobs:
     strategy:
       matrix:
         action-name: ${{fromJson(needs.prepare.outputs.container_names_json)}}
-        docker-platform: [linux/amd64, linux/arm64]
 
     env:
       REGISTRY: ghcr.io
@@ -90,8 +89,8 @@ jobs:
           push: true
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
-          platforms: ${{ matrix.docker-platform }}
-          provenance: false # Bug: https://github.com/docker/build-push-action/issues/820
+          platforms: linux/amd64,linux/arm64
+          # Disable "provenance" vecause of this bug: https://github.com/docker/build-push-action/issues/820
+          provenance: false
           cache-from: type=gha
           cache-to: type=gha,mode=max
-

.github/workflows/jekyll-feed-to-vespa.yml

@@ -9,6 +9,12 @@ on:
         required: false
         default: 'info'
         type: string
+      hook-post-build-script:
+        description: |
+          Script to run after Jekyll build, can be multi-line.
+          Ref: https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#jobsjob_idstepsrun
+        required: false
+        type: string
 
 defaults:
   run:
@@ -17,17 +23,18 @@ defaults:
     shell: bash
 
 jobs:
-  lint:
+  build:
     runs-on: ubuntu-latest
 
     steps:
       - uses: actions/checkout@v4
 
       - name: Build with Jekyll
-        uses: vespa-engine/gh-actions/jekyll-build-json@main
+        uses: vespa-engine/gh-actions/jekyll-build-json@fix-jekyll-build
         with:
-          log_level: ${{ github.event.inputs.log_level }}
+          log_level: ${{ inputs.log_level }}
 
-      - name: List JSON
+      - name: Hook - Post-build
+        if: ${{ inputs.hook-post-build-script != '' }}
         run: |
-          find _site -type f -name '*.json'
+          ${{ inputs.hook-post-build-script }}

jekyll-build-json/Gemfile

@@ -4,14 +4,4 @@
 source "https://rubygems.org"
 
 # Manage our dependency on the version of the github-pages gem here.
-gem "github-pages", "= 232"
-
-# Explicitly include this gem here.
-# It is not directly included in the github-pages gem list of dependencies,
-# even though it is included in the original GitHub Pages build infrastructure.
-gem "jekyll-include-cache", "= 0.2.1"
-gem "jekyll-octicons", "~> 14.2"
-gem "minima", "~> 2.0"
-
-# Work-around for webrick no longer included in Ruby 3.0 (https://github.com/jekyll/jekyll/issues/8523)
-gem "webrick"
+gem 'github-pages', group: :jekyll_plugins

jekyll-build-json/action.yml

@@ -16,12 +16,12 @@ inputs:
     default: ./_config.yml
   log_level:
     description: 'Log level for Jekyll build.'
-    required: false
+    required: true
     default: 'info'
   token:
     description: 'GitHub token'
     required: true
     default: ${{ github.token }}
 runs:
   using: 'docker'
-  image: 'docker://ghcr.io/vespa-engine/gh-actions/jekyll-build-json'
+  image: 'docker://ghcr.io/vespa-engine/gh-actions/jekyll-build-json:pr-14'

jekyll-build-json/entrypoint.sh

@@ -17,7 +17,7 @@ JEKYLL_ARGS=("-p" "/opt/jekyll/plugins")
 export JEKYLL_ENV="production"
 export JEKYLL_GITHUB_TOKEN="${INPUT_TOKEN}"
 export JEKYLL_BUILD_REVISION="${GITHUB_SHA}"
-export JEKYLL_LOG_LEVEL="${INPUT_LOG_LEVEL}"
+export JEKYLL_LOG_LEVEL="${INPUT_LOG_LEVEL:-info}"
 
 if [[ "${JEKYLL_LOG_LEVEL}" = "verbose" || "${JEKYLL_LOG_LEVEL}" = "debug" ]]; then
   JEKYLL_ARGS+=("-V")

jekyll-build-json/vespa_index_generator.rb

@@ -6,14 +6,14 @@
 
 module Jekyll
 
+    class JSONFile < Page; end
+
     class VespaIndexGenerator < Jekyll::Generator
         priority :lowest
         safe true
 
         def generate(site)
-          @site = site
-          puts "::debug::VespaIndexGenerator"
-            puts "site.config: #{(site.config).inspect}"
+            puts "::debug::VespaIndexGenerator"
 
             raise "Wrong parameter type, expected Jekyll::Site, got #{site.class}" unless site.is_a?(Jekyll::Site)
             raise "Search configuration not found" unless site.config.key?("search")
@@ -38,7 +38,7 @@ def generate(site)
                 url = page.url
                 url += 'index.html' if url[-1, 1] == '/'
 
-                text = extract_text(page)
+                text = extract_text(page.to_liquid)
                 outlinks = extract_links(page)
                 headers = extract_headers(page)
                 keywords = get_keywords(page)
@@ -61,14 +61,15 @@ def generate(site)
                 })
             end
             json = JSON.pretty_generate(operations)
-            # puts "::debug::Writing index file: #{namespace}_index.json"
-            # File.open(__dir__ + namespace + "_index.json", "w") { |f| f.write(json) }
-
-            puts "::debug::Writing index file: #{namespace}_index.json"
-            index_file = PageWithoutAFile.new(@site, __dir__, "", namespace + "_index.json")
-            index_file.content = json
-            index_file.data["layout"] = nil
-            index_file
+            json_dest = "json_indexes/"
+            full_path = "#{site.dest}/#{json_dest}"
+            FileUtils.mkdir_p(full_path) # Ensure the directory exists
+
+            fname = "#{namespace}_index.json"
+            puts "::debug::Writing JSON file for \"#{namespace}\" namespace to: #{full_path}#{fname}"
+            File.open("#{full_path}#{fname}", 'w') { |f| f.write(json) }
+
+            site.pages << Jekyll::JSONFile.new(site, site.dest, json_dest, fname)
         end
 
         def should_skip?(search_config, page)