Add Hugo Algolia indexing

.github/workflows/build.yml

@@ -76,12 +76,15 @@ jobs:
         with:
           node-version: ${{ env.NODE_VERSION }}
           cache: npm
+          cache-dependency-path: tools/package-lock.json
 
       - name: Install Node.js dependencies
         run: npm ci
+        working-directory: tools
 
       - name: Test Algolia indexing
         run: npm run test:algolia
+        working-directory: tools
 
       - name: Verify Hugo modules
         run: hugo mod verify
@@ -91,6 +94,7 @@ jobs:
 
       - name: Validate Algolia records
         run: npm run algolia:index -- --dry-run
+        working-directory: tools
 
       - name: Install XML validation tools
         run: sudo apt-get update && sudo apt-get install --yes libxml2-utils

.github/workflows/update-algolia.yml

@@ -1,39 +1,79 @@
 name: Update the Algolia search index
+
 on:
   workflow_dispatch:
   schedule:
     - cron: "0 2 * * *"
 
+permissions:
+  contents: read
+
+concurrency:
+  group: update-algolia
+  cancel-in-progress: false
+
+env:
+  GO_VERSION: "1.26.3"
+  HUGO_VERSION: "0.163.3"
+  NODE_VERSION: "24.18.0"
+  ALGOLIA_APP_ID: "LIT6P0EW26"
+  ALGOLIA_INDEX_NAME: "hugo"
+
 jobs:
-  build:
+  index:
+    name: Build and publish index
     runs-on: ubuntu-latest
+    timeout-minutes: 30
+
     steps:
       - name: Checkout repository
-        uses: actions/checkout@v6
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
         with:
-          submodules: true
+          fetch-depth: 0
+          submodules: false
           lfs: false
-          path: website
-      - uses: ruby/setup-ruby@v1
+
+      - name: Set up Go
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6
+        with:
+          go-version: ${{ env.GO_VERSION }}
+          cache-dependency-path: go.sum
+
+      - name: Set up Hugo
+        uses: peaceiris/actions-hugo@2752ce1d29631191ea3f27c23495fa06139a5b78 # v3
         with:
-          bundler-cache: true
-          working-directory: website
-      - name: Check recent commit
-        id: checkcommit
-        working-directory: website
-        run: |
-          LATEST_COMMIT=$(git log -1 --format=%ct)
-          CURRENT_TIME=$(date +%s)
-          TIME_DIFF=$((CURRENT_TIME - LATEST_COMMIT))
-          if [ $TIME_DIFF -gt 86400 ]; then
-            echo "No commit within 24 hours."
-            echo "recent_commit=false" >> "$GITHUB_OUTPUT"
-          else
-            echo "recent_commit=true" >> "$GITHUB_OUTPUT"
-          fi
-      - name: Build website and update Algolia index
-        if: ${{ github.event_name == 'workflow_dispatch' || steps.checkcommit.outputs.recent_commit == 'true' }}
-        working-directory: website
+          hugo-version: ${{ env.HUGO_VERSION }}
+          extended: true
+
+      - name: Set up Node.js
+        uses: actions/setup-node@249970729cb0ef3589644e2896645e5dc5ba9c38 # v6
+        with:
+          node-version: ${{ env.NODE_VERSION }}
+          cache: npm
+          cache-dependency-path: tools/package-lock.json
+
+      - name: Install Node.js dependencies
+        run: npm ci
+        working-directory: tools
+
+      - name: Test Algolia indexing
+        run: npm run test:algolia
+        working-directory: tools
+
+      - name: Verify Hugo modules
+        run: hugo mod verify
+
+      - name: Build Hugo Algolia export
+        run: hugo --gc --minify --cleanDestinationDir --environment production
+
+      - name: Validate generated Algolia records
+        run: npm run algolia:index -- --dry-run
+        working-directory: tools
+
+      - name: Replace Algolia records
         env:
-          JEKYLL_ENV: production
-        run: ALGOLIA_API_KEY=${{ secrets.ALGOLIA_API_KEY }} bundle exec jekyll algolia
+          ALGOLIA_APP_ID: ${{ env.ALGOLIA_APP_ID }}
+          ALGOLIA_WRITE_API_KEY: ${{ secrets.ALGOLIA_API_KEY }}
+          ALGOLIA_INDEX_NAME: ${{ env.ALGOLIA_INDEX_NAME }}
+        run: npm run algolia:index
+        working-directory: tools

config/_default/hugo.toml

@@ -26,7 +26,7 @@ precice_v2_citations_url = "https://scholar.google.com/scholar?oi=bibs&hl=en&cit
 
 [params.algolia]
 application_id = "LIT6P0EW26"
-index_name = "jekyll"
+index_name = "hugo"
 search_only_api_key = "760ed6be3e165d08b4a798ac4aa82fe4"
 nodes_to_index = "p,code,table"
 max_record_size = 20000

layouts/index.algolia.json

@@ -0,0 +1,31 @@
+{{- $pages := slice -}}
+{{- range site.Pages -}}
+  {{- if and (or (eq .Kind "page") (eq .Kind "section")) (ne (.Param "search") "exclude") -}}
+    {{- $pages = $pages | append . -}}
+  {{- end -}}
+{{- end -}}
+{{- $pages = sort $pages "RelPermalink" -}}
+{
+  "schemaVersion": 1,
+  "nodesToIndex": {{ site.Params.algolia.nodes_to_index | jsonify }},
+  "maxRecordSize": {{ site.Params.algolia.max_record_size }},
+  "pages": [
+    {{- range $index, $page := $pages }}
+      {{- if $index }},{{ end }}
+      {{- /* Keep the search export on the same rendered-content path as page layouts. */ -}}
+      {{- $content := partial "compat_content.html" $page -}}
+      {
+        "title": {{ $page.Title | jsonify }},
+        "url": {{ $page.RelPermalink | jsonify }},
+        "html": {{ $content | jsonify }},
+        "plain": {{ $content | plainify | jsonify }},
+        "section": {{ $page.Section | jsonify }},
+        "kind": {{ $page.Kind | jsonify }},
+        "tags": {{ $page.Params.tags | jsonify }},
+        "categories": {{ $page.Params.categories | jsonify }},
+        "keywords": {{ $page.Params.keywords | jsonify }},
+        "date": {{ if ne ($page.Lastmod.Format "2006") "0001" }}{{ $page.Lastmod.Unix }}{{ else }}null{{ end }}
+      }
+    {{- end }}
+  ]
+}

test/algolia-index.test.mjs

@@ -0,0 +1,90 @@
+import assert from "node:assert/strict";
+import test from "node:test";
+import {
+  createRecords,
+  parseCliArgs,
+  recordByteSize,
+  validateSearchExport,
+} from "../tools/algolia-index.mjs";
+
+const searchExport = {
+  schemaVersion: 1,
+  nodesToIndex: "p,code,table",
+  maxRecordSize: 20_000,
+  pages: [
+    {
+      title: "Example guide",
+      url: "/docs/example/",
+      html: `
+        <h2 id="install">Install</h2>
+        <p>Use <code>precice-config</code> to configure the adapter.</p>
+        <table><tbody><tr><td>Version</td><td>3.4</td></tr></tbody></table>
+        <script>ignored content</script>
+      `,
+      plain: "Use precice-config to configure the adapter. Version 3.4",
+      section: "docs",
+      kind: "page",
+      tags: ["adapter"],
+      categories: [],
+      keywords: "configuration",
+      date: 1_784_198_400,
+    },
+  ],
+};
+
+test("creates Jekyll-compatible records with heading and anchor metadata", () => {
+  const records = createRecords(searchExport);
+
+  assert.equal(records.length, 3);
+  assert.equal(records[0].title, "Example guide");
+  assert.equal(records[0].url, "/docs/example/");
+  assert.equal(records[0].anchor, "install");
+  assert.deepEqual(records[0].headings, ["Install"]);
+  assert.match(records[0].html, /^<p>/);
+  assert.equal(records[1].content, "precice-config");
+  assert.equal(records[2].content, "Version3.4");
+  assert.equal(records[0].custom_ranking.heading, 80);
+  assert.match(records[0].objectID, /^[a-f0-9]{64}$/);
+  assert.equal(records.some((record) => record.content.includes("ignored content")), false);
+});
+
+test("creates deterministic object IDs", () => {
+  const first = createRecords(searchExport).map((record) => record.objectID);
+  const second = createRecords(searchExport).map((record) => record.objectID);
+
+  assert.deepEqual(first, second);
+});
+
+test("excludes callouts so a tutorial result starts with its introduction", () => {
+  const calloutExport = structuredClone(searchExport);
+  calloutExport.nodesToIndex = "p";
+  calloutExport.pages[0].html = `
+    <div class="alert alert-info" role="alert"><p>Note: Download the case files.</p></div>
+    <p>This tutorial introduces the partitioned heat-conduction case.</p>
+  `;
+  calloutExport.pages[0].plain = "Note: Download the case files. This tutorial introduces the partitioned heat-conduction case.";
+
+  const records = createRecords(calloutExport);
+
+  assert.equal(records.length, 1);
+  assert.equal(records[0].content, "This tutorial introduces the partitioned heat-conduction case.");
+  assert.equal(records[0].content.includes("Download the case files"), false);
+});
+
+test("splits oversized records without exceeding the configured limit", () => {
+  const longExport = structuredClone(searchExport);
+  longExport.nodesToIndex = "p";
+  longExport.pages[0].html = `<p>${"word ".repeat(400)}</p>`;
+  longExport.pages[0].plain = "word ".repeat(400);
+
+  const records = createRecords(longExport, { maxRecordSize: 500 });
+
+  assert.ok(records.length > 1);
+  assert.ok(records.every((record) => recordByteSize(record) <= 500));
+});
+
+test("rejects invalid Hugo exports and command-line arguments", () => {
+  assert.throws(() => validateSearchExport({ schemaVersion: 2 }), /Unsupported Algolia export schema/);
+  assert.throws(() => parseCliArgs(["--batch-size", "0"]), /positive integer/);
+  assert.throws(() => parseCliArgs(["--unexpected"]), /Unknown option/);
+});