feat(export/share): use better content parsing for search

apps/server/package.json

@@ -26,6 +26,7 @@
   },
   "dependencies": {
     "better-sqlite3": "12.4.1",
+    "html-to-text": "9.0.5",
     "node-html-parser": "7.0.1"
   },
   "devDependencies": {
@@ -35,8 +36,8 @@
     "@preact/preset-vite": "2.10.2",
     "@triliumnext/commons": "workspace:*",
     "@triliumnext/express-partial-content": "workspace:*",
-    "@triliumnext/turndown-plugin-gfm": "workspace:*",
     "@triliumnext/highlightjs": "workspace:*",
+    "@triliumnext/turndown-plugin-gfm": "workspace:*",
     "@types/archiver": "7.0.0",
     "@types/better-sqlite3": "7.6.13",
     "@types/cls-hooked": "4.3.9",

apps/server/src/services/export/zip/share_theme.ts

@@ -8,6 +8,7 @@ import { renderNoteForExport } from "../../../share/content_renderer";
 import type BNote from "../../../becca/entities/bnote.js";
 import type BBranch from "../../../becca/entities/bbranch.js";
 import { getShareThemeAssetDir } from "../../../routes/assets";
+import { convert as convertToText } from "html-to-text";
 
 const shareThemeAssetDir = getShareThemeAssetDir();
 
@@ -57,7 +58,9 @@ export default class ShareThemeExportProvider extends ZipExportProvider {
 
         if (note) {
             // Prepare search index.
-            searchContent = typeof content === "string" ? utils.stripTags(content) : "";
+            searchContent = typeof content === "string" ? convertToText(content, {
+                whitespaceCharacters: "\t\r\n\f\u200b\u00a0\u2002"
+            }) : "";
 
             content = renderNoteForExport(note, branch, basePath, noteMeta.notePath.slice(0, -1));
             if (typeof content === "string") {

pnpm-lock.yaml

@@ -458,6 +458,9 @@ importers:
       better-sqlite3:
         specifier: 12.4.1
         version: 12.4.1
+      html-to-text:
+        specifier: 9.0.5
+        version: 9.0.5
       node-html-parser:
         specifier: 7.0.1
         version: 7.0.1
@@ -4354,6 +4357,9 @@ packages:
   '@scarf/scarf@1.4.0':
     resolution: {integrity: sha512-xxeapPiUXdZAE3che6f3xogoJPeZgig6omHEy1rIY5WVsB3H2BHNnZH+gHG6x91SCWyQCzWGsuL2Hh3ClO5/qQ==}
 
+  '@selderee/plugin-htmlparser2@0.11.0':
+    resolution: {integrity: sha512-P33hHGdldxGabLFjPPpaTxVolMrzrcegejx+0GxjrIb9Zv48D8yAIA/QTDR2dFl7Uz7urX8aX6+5bCZslr+gWQ==}
+
   '@sideway/address@4.1.5':
     resolution: {integrity: sha512-IqO/DUQHUkPeixNQ8n0JA6102hT9CmaljNTPmQ1u8MEhBo/R4Q8eKLN/vGZxuebwOroDB4cbpjheD4+/sKFK4Q==}
 
@@ -8491,6 +8497,10 @@ packages:
     resolution: {integrity: sha512-ztqyC3kLto0e9WbNp0aeP+M3kTt+nbaIveGmUxAtZa+8iFgKLUOD4YKM5j+f3QD89bra7UeumolZHKuOXnTmeQ==}
     engines: {node: '>=8'}
 
+  html-to-text@9.0.5:
+    resolution: {integrity: sha512-qY60FjREgVZL03vJU6IfMV4GDjGBIoOyvuFdpBDIX9yTlDw0TjxVBQp+P8NvpdIXNJvfWBTNul7fsAQJq2FNpg==}
+    engines: {node: '>=14'}
+
   html-void-elements@3.0.0:
     resolution: {integrity: sha512-bEqo66MRXsUGxWHV5IP0PUiAWwoEjba4VCzg0LjFJBpchPaTfyfCKTG6bc5F8ucKec3q5y6qOdGyYTSBEvhCrg==}
 
@@ -9324,6 +9334,9 @@ packages:
     resolution: {integrity: sha512-b94GiNHQNy6JNTrt5w6zNyffMrNkXZb3KTkCZJb2V1xaEGCk093vkZ2jk3tpaeP33/OiXC+WvK9AxUebnf5nbw==}
     engines: {node: '>= 0.6.3'}
 
+  leac@0.6.0:
+    resolution: {integrity: sha512-y+SqErxb8h7nE/fiEX07jsbuhrpO9lL8eca7/Y1nuWV2moNlXhyd59iDGcRf6moVyDMbmTNzL40SUyrFU/yDpg==}
+
   leaflet-gpx@2.2.0:
     resolution: {integrity: sha512-iVUx6o0ydLn2ikYSVLuWnr0k/CDAOIUtmvQ91AI24/PXuIFIb+iEIJMHTQfvGhMCKcFrwd9ZaFYH7P/46tgGhw==}
 
@@ -10592,6 +10605,9 @@ packages:
   parse5@7.3.0:
     resolution: {integrity: sha512-IInvU7fabl34qmi9gY8XOVxhYyMyuH2xUNpb2q8/Y+7552KlejkRvqvD19nMoUW/uQGGbqNpA6Tufu5FL5BZgw==}
 
+  parseley@0.12.1:
+    resolution: {integrity: sha512-e6qHKe3a9HWr0oMRVDTRhKce+bRO8VGQR3NyVwcjwrbhMmFCX9KszEV35+rn4AdilFAq9VPxP/Fe1wC9Qjd2lw==}
+
   parseurl@1.3.3:
     resolution: {integrity: sha512-CiyeOxFT/JZyN5m0z9PfXw4SCBJ6Sygz1Dpl0wqjlhDEGGBP1GnsUVEL0p63hoG1fcj3fHynXi9NYO4nWOL+qQ==}
     engines: {node: '>= 0.8'}
@@ -10673,6 +10689,9 @@ packages:
     resolution: {integrity: sha512-nh39Mo1eGWmZS7y+mK/dQIqg7S1lp38DpRxkyoHf0ZcUs/HDc+yyTjuOtTvSMZHmfSLuSQaX945u05Y2Q6UWZg==}
     engines: {node: '>=14', npm: '>=7'}
 
+  peberminta@0.9.0:
+    resolution: {integrity: sha512-XIxfHpEuSJbITd1H3EeQwpcZbTLHc+VVr8ANI9t5sit565tsI4/xK3KWTUFE2e6QiangUkh3B0jihzmGnNrRsQ==}
+
   peek-readable@4.1.0:
     resolution: {integrity: sha512-ZI3LnwUv5nOGbQzD9c2iDG6toheuXSZP5esSHBjopsXH4dg19soufvpUGA3uohi5anFtGb2lhAVdHzH6R/Evvg==}
     engines: {node: '>=8'}
@@ -12321,6 +12340,9 @@ packages:
   secure-compare@3.0.1:
     resolution: {integrity: sha512-AckIIV90rPDcBcglUwXPF3kg0P0qmPsPXAj6BBEENQE1p5yA1xfmDJzfi1Tappj37Pv2mVbKpL3Z1T+Nn7k1Qw==}
 
+  selderee@0.11.0:
+    resolution: {integrity: sha512-5TF+l7p4+OsnP8BCCvSyZiSPc4x4//p5uPwK8TCnVPJYRmU2aYKMpOXvw8zM5a5JvuuCGN1jmsMwuU2W02ukfA==}
+
   select-hose@2.0.0:
     resolution: {integrity: sha512-mEugaLK+YfkijB4fx0e6kImuJdCIt2LxCRcbEYPqRGCs4F2ogyfZU5IAZRdjCP8JPq2AtdNoC/Dux63d9Kiryg==}
 
@@ -15081,6 +15103,8 @@ snapshots:
       '@ckeditor/ckeditor5-utils': 47.1.0
       '@ckeditor/ckeditor5-watchdog': 47.1.0
       es-toolkit: 1.39.5
+    transitivePeerDependencies:
+      - supports-color
 
   '@ckeditor/ckeditor5-dev-build-tools@43.1.0(@swc/helpers@0.5.17)(tslib@2.8.1)(typescript@5.9.3)':
     dependencies:
@@ -15245,6 +15269,8 @@ snapshots:
       '@ckeditor/ckeditor5-utils': 47.1.0
       ckeditor5: 47.1.0(patch_hash=8331a09d41443b39ea1c784daaccfeb0da4f9065ed556e7de92e9c77edd9eb41)
       es-toolkit: 1.39.5
+    transitivePeerDependencies:
+      - supports-color
 
   '@ckeditor/ckeditor5-editor-classic@47.1.0':
     dependencies:
@@ -15272,6 +15298,8 @@ snapshots:
       '@ckeditor/ckeditor5-utils': 47.1.0
       ckeditor5: 47.1.0(patch_hash=8331a09d41443b39ea1c784daaccfeb0da4f9065ed556e7de92e9c77edd9eb41)
       es-toolkit: 1.39.5
+    transitivePeerDependencies:
+      - supports-color
 
   '@ckeditor/ckeditor5-editor-multi-root@47.1.0':
     dependencies:
@@ -15524,8 +15552,6 @@ snapshots:
       '@ckeditor/ckeditor5-ui': 47.1.0
       '@ckeditor/ckeditor5-utils': 47.1.0
       ckeditor5: 47.1.0(patch_hash=8331a09d41443b39ea1c784daaccfeb0da4f9065ed556e7de92e9c77edd9eb41)
-    transitivePeerDependencies:
-      - supports-color
 
   '@ckeditor/ckeditor5-line-height@47.1.0':
     dependencies:
@@ -15768,8 +15794,6 @@ snapshots:
       '@ckeditor/ckeditor5-ui': 47.1.0
       '@ckeditor/ckeditor5-utils': 47.1.0
       ckeditor5: 47.1.0(patch_hash=8331a09d41443b39ea1c784daaccfeb0da4f9065ed556e7de92e9c77edd9eb41)
-    transitivePeerDependencies:
-      - supports-color
 
   '@ckeditor/ckeditor5-restricted-editing@47.1.0':
     dependencies:
@@ -15814,8 +15838,6 @@ snapshots:
       '@ckeditor/ckeditor5-ui': 47.1.0
       '@ckeditor/ckeditor5-utils': 47.1.0
       ckeditor5: 47.1.0(patch_hash=8331a09d41443b39ea1c784daaccfeb0da4f9065ed556e7de92e9c77edd9eb41)
-    transitivePeerDependencies:
-      - supports-color
 
   '@ckeditor/ckeditor5-slash-command@47.1.0':
     dependencies:
@@ -16024,8 +16046,6 @@ snapshots:
       '@ckeditor/ckeditor5-utils': 47.1.0
       ckeditor5: 47.1.0(patch_hash=8331a09d41443b39ea1c784daaccfeb0da4f9065ed556e7de92e9c77edd9eb41)
       es-toolkit: 1.39.5
-    transitivePeerDependencies:
-      - supports-color
 
   '@codemirror/autocomplete@6.18.6':
     dependencies:
@@ -18703,6 +18723,11 @@ snapshots:
 
   '@scarf/scarf@1.4.0': {}
 
+  '@selderee/plugin-htmlparser2@0.11.0':
+    dependencies:
+      domhandler: 5.0.3
+      selderee: 0.11.0
+
   '@sideway/address@4.1.5':
     dependencies:
       '@hapi/hoek': 9.3.0
@@ -24154,6 +24179,14 @@ snapshots:
 
   html-tags@3.3.1: {}
 
+  html-to-text@9.0.5:
+    dependencies:
+      '@selderee/plugin-htmlparser2': 0.11.0
+      deepmerge: 4.3.1
+      dom-serializer: 2.0.0
+      htmlparser2: 8.0.2
+      selderee: 0.11.0
+
   html-void-elements@3.0.0: {}
 
   html2plaintext@2.1.4:
@@ -25061,6 +25094,8 @@ snapshots:
     dependencies:
       readable-stream: 2.3.8
 
+  leac@0.6.0: {}
+
   leaflet-gpx@2.2.0: {}
 
   leaflet@1.9.4: {}
@@ -26731,6 +26766,11 @@ snapshots:
     dependencies:
       entities: 6.0.0
 
+  parseley@0.12.1:
+    dependencies:
+      leac: 0.6.0
+      peberminta: 0.9.0
+
   parseurl@1.3.3: {}
 
   path-browserify@1.0.1: {}
@@ -26789,6 +26829,8 @@ snapshots:
 
   pe-library@1.0.1: {}
 
+  peberminta@0.9.0: {}
+
   peek-readable@4.1.0: {}
 
   peek-readable@7.0.0: {}
@@ -28549,6 +28591,10 @@ snapshots:
 
   secure-compare@3.0.1: {}
 
+  selderee@0.11.0:
+    dependencies:
+      parseley: 0.12.1
+
   select-hose@2.0.0: {}
 
   selfsigned@2.4.1: