From b2f9d57f57fb3d110112e4ce86fb9891e1fda1d9 Mon Sep 17 00:00:00 2001
From: Mark van Ments <35492184+MarkvanMents@users.noreply.github.com>
Date: Wed, 30 Sep 2026 17:25:36 +0200
Subject: [PATCH 1/2] Revert "Revert "Add anchors to Algolia indexes (#11976)"
 (#11993)"

This reverts commit 7c0957a208a0502b06bc74ed840a55af45e6a274.
---
 layouts/_default/list.mxdocsalgolia.json | 70 +++++++++++++++++++++++-
 1 file changed, 67 insertions(+), 3 deletions(-)

diff --git a/layouts/_default/list.mxdocsalgolia.json b/layouts/_default/list.mxdocsalgolia.json
index 5637f341c96..9ff0a7eac44 100644
--- a/layouts/_default/list.mxdocsalgolia.json
+++ b/layouts/_default/list.mxdocsalgolia.json
@@ -1,6 +1,26 @@
 {{- /*
-Algolia generation based on this blog: https://web.archive.org/web/20201109041339/https://forestry.io/blog/search-with-algolia-in-hugo/
-Define Global variables
+Algolia search index generator — based on https://web.archive.org/web/20201109041339/https://forestry.io/blog/search-with-algolia-in-hugo/
+
+Outputs a JSON array of Algolia records, one per content chunk per page.
+
+Pages included: all non-draft, non-private pages of type "docs" or "swagger"
+that are descendants of the current section.
+
+How records are created:
+1. Each page's rendered HTML is split on heading tags (h2–h6), paragraph,
+   list item, table cell, alert div, and code block tags. This produces one
+   chunk per distinct piece of content.
+2. As chunks are processed, heading context (h2–h6 text and id anchors) is
+   tracked and updated. When a higher-level heading is encountered, all
+   deeper heading levels are reset.
+3. Heading chunks themselves are never written as records. Only non-heading
+   chunks produce a record, carrying the current heading context as fields.
+4. Each record includes: objectID (unique per chunk), title, h2–h6 heading
+   text, text (plain content), url (page URL with #anchor of deepest heading
+   in scope), unique_hierarchy (breadcrumb string), slug, description,
+   content_type, type, mendix_version, menu_order, time_stamp, weight.position,
+   and weight.tag_name (p / li / th / td / code / alert).
+5. Records with empty text are excluded.
 */ -}}
 {{- $firstloop := true -}}
 {{- /* Use a local scratchpad - Page scratchpad already contains other stuff */ -}}
@@ -59,6 +79,11 @@ Define Global variables
         {{- $h4 := "" }}
         {{- $h5 := "" }}
         {{- $h6 := "" }}
+        {{- $h2anchor := "" }}
+        {{- $h3anchor := "" }}
+        {{- $h4anchor := "" }}
+        {{- $h5anchor := "" }}
+        {{- $h6anchor := "" }}
         {{- $text := "" }}
         {{- range $processSlice := $contentSlices  -}}
             {{- /* sliceNumber tells us where we are in the file and gives us a way of making each hit unique */ -}}
@@ -71,24 +96,50 @@ Define Global variables
                     {{- /* Set up the heading level we are in */ -}}
                     {{- if eq $startSlice "<h2" -}}
                         {{- $h2 = strings.TrimSpace (htmlUnescape ($processSlice | plainify)) -}}
+                        {{- /* Extract the id attribute from the heading element for use as a URL anchor */ -}}
+                        {{- $anchorMatch := findRE `id="([^"]+)"` $processSlice 1 -}}
+                        {{- $h2anchor = "" -}}
+                        {{- range $anchorMatch -}}{{- $h2anchor = strings.TrimPrefix `id="` (strings.TrimSuffix `"` .) -}}{{- end -}}
                         {{- $h3 = "" -}}
                         {{- $h4 = "" -}}
                         {{- $h5 = "" -}}
                         {{- $h6 = "" -}}
+                        {{- $h3anchor = "" -}}
+                        {{- $h4anchor = "" -}}
+                        {{- $h5anchor = "" -}}
+                        {{- $h6anchor = "" -}}
                     {{- else if eq $startSlice "<h3" -}}
                         {{- $h3 = strings.TrimSpace (htmlUnescape ($processSlice | plainify)) -}}
+                        {{- $anchorMatch := findRE `id="([^"]+)"` $processSlice 1 -}}
+                        {{- $h3anchor = "" -}}
+                        {{- range $anchorMatch -}}{{- $h3anchor = strings.TrimPrefix `id="` (strings.TrimSuffix `"` .) -}}{{- end -}}
                         {{- $h4 = "" -}}
                         {{- $h5 = "" -}}
                         {{- $h6 = "" -}}
+                        {{- $h4anchor = "" -}}
+                        {{- $h5anchor = "" -}}
+                        {{- $h6anchor = "" -}}
                     {{- else if eq $startSlice "<h4" -}}
                         {{- $h4 = strings.TrimSpace (htmlUnescape ($processSlice | plainify)) -}}
+                        {{- $anchorMatch := findRE `id="([^"]+)"` $processSlice 1 -}}
+                        {{- $h4anchor = "" -}}
+                        {{- range $anchorMatch -}}{{- $h4anchor = strings.TrimPrefix `id="` (strings.TrimSuffix `"` .) -}}{{- end -}}
                         {{- $h5 = "" -}}
                         {{- $h6 = "" -}}
+                        {{- $h5anchor = "" -}}
+                        {{- $h6anchor = "" -}}
                     {{- else if eq $startSlice "<h5" -}}
                         {{- $h5 = strings.TrimSpace (htmlUnescape ($processSlice | plainify)) -}}
+                        {{- $anchorMatch := findRE `id="([^"]+)"` $processSlice 1 -}}
+                        {{- $h5anchor = "" -}}
+                        {{- range $anchorMatch -}}{{- $h5anchor = strings.TrimPrefix `id="` (strings.TrimSuffix `"` .) -}}{{- end -}}
                         {{- $h6 = "" -}}
+                        {{- $h6anchor = "" -}}
                     {{- else if eq $startSlice "<h6" -}}
                         {{- $h6 = strings.TrimSpace (htmlUnescape ($processSlice | plainify)) -}}
+                        {{- $anchorMatch := findRE `id="([^"]+)"` $processSlice 1 -}}
+                        {{- $h6anchor = "" -}}
+                        {{- range $anchorMatch -}}{{- $h6anchor = strings.TrimPrefix `id="` (strings.TrimSuffix `"` .) -}}{{- end -}}
                     {{- else -}}
                         {{- /*
                              * Get all the information for this piece of text (not a heading) and label it uniquely
@@ -148,7 +199,20 @@ Define Global variables
                         {{- end -}}
                         {{- $scratch.SetInMap $uniqueHit "weight.tag_name" $tagName -}}
                         {{- /* .RelPermalink gives the page address relative to the baseURL. It reflects any changes from the url: tag in the front matter */ -}}
-                        {{- $scratch.SetInMap $uniqueHit "url" $pageDot.RelPermalink -}}
+                        {{- /* Append the deepest available heading anchor so search results link directly to the relevant section */ -}}
+                        {{- $anchor := "" -}}
+                        {{- with $h6anchor -}}{{- $anchor = $h6anchor -}}{{- else -}}
+                            {{- with $h5anchor -}}{{- $anchor = $h5anchor -}}{{- else -}}
+                                {{- with $h4anchor -}}{{- $anchor = $h4anchor -}}{{- else -}}
+                                    {{- with $h3anchor -}}{{- $anchor = $h3anchor -}}{{- else -}}
+                                        {{- with $h2anchor -}}{{- $anchor = $h2anchor -}}{{- end -}}
+                                    {{- end -}}
+                                {{- end -}}
+                            {{- end -}}
+                        {{- end -}}
+                        {{- $pageURL := $pageDot.RelPermalink -}}
+                        {{- with $anchor -}}{{- $pageURL = printf "%s#%s" $pageDot.RelPermalink $anchor -}}{{- end -}}
+                        {{- $scratch.SetInMap $uniqueHit "url" $pageURL -}}
                         {{- $scratch.SetInMap $uniqueHit "slug" $pageDot.File.TranslationBaseName -}}
                         {{- $scratch.SetInMap $uniqueHit "unique_hierarchy" $uniqueHierarchy -}}
                     {{- end -}}

From 8c0c5b24dac14f729124d9cb6ab8f963cf65d0ac Mon Sep 17 00:00:00 2001
From: MarkvanMents <mark.van.ments@mendix.com>
Date: Thu, 1 Oct 2026 13:32:39 +0200
Subject: [PATCH 2/2] Add page_url field to Algolia records for deduplication

The url field now includes a heading anchor (e.g. /refguide/entities/#entity-types),
which broke deduplication of multiple hits from the same page. The new
page_url field carries the bare page URL without any anchor, providing a
stable key for deduplicating hits per page in Algolia.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---
 layouts/_default/list.mxdocsalgolia.json | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/layouts/_default/list.mxdocsalgolia.json b/layouts/_default/list.mxdocsalgolia.json
index 9ff0a7eac44..c90a2f7524c 100644
--- a/layouts/_default/list.mxdocsalgolia.json
+++ b/layouts/_default/list.mxdocsalgolia.json
@@ -19,7 +19,8 @@ How records are created:
    text, text (plain content), url (page URL with #anchor of deepest heading
    in scope), unique_hierarchy (breadcrumb string), slug, description,
    content_type, type, mendix_version, menu_order, time_stamp, weight.position,
-   and weight.tag_name (p / li / th / td / code / alert).
+   weight.tag_name (p / li / th / td / code / alert), and page_url (bare page
+   URL without anchor, for deduplicating multiple hits from the same page).
 5. Records with empty text are excluded.
 */ -}}
 {{- $firstloop := true -}}
@@ -213,6 +214,8 @@ How records are created:
                         {{- $pageURL := $pageDot.RelPermalink -}}
                         {{- with $anchor -}}{{- $pageURL = printf "%s#%s" $pageDot.RelPermalink $anchor -}}{{- end -}}
                         {{- $scratch.SetInMap $uniqueHit "url" $pageURL -}}
+                        {{- /* page_url is the bare page URL without any anchor, for deduplicating multiple hits from the same page */ -}}
+                        {{- $scratch.SetInMap $uniqueHit "page_url" $pageDot.RelPermalink -}}
                         {{- $scratch.SetInMap $uniqueHit "slug" $pageDot.File.TranslationBaseName -}}
                         {{- $scratch.SetInMap $uniqueHit "unique_hierarchy" $uniqueHierarchy -}}
                     {{- end -}}