Merge branch 'develop'

.github/workflows/release.yml

@@ -14,17 +14,20 @@ jobs:
     runs-on: ubuntu-latest
 
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
-
-      - uses: pnpm/action-setup@v2.1.0
         with:
-          version: 7.17.0
+          fetch-depth: 0
+
+      - uses: pnpm/action-setup@v4
+        with:
+          version: 9.3.0
           run_install: true
 
       - name: Use Node.js
-        uses: actions/setup-node@v1
+        uses: actions/setup-node@v4
         with:
-          node-version: "18.x"
+          cache: 'pnpm'
+          node-version: "20.x"
 
       - name: Build
         id: build
@@ -37,6 +40,18 @@ jobs:
           ls
           echo "::set-output name=tag_name::$(git tag --sort version:refname | tail -n 1)"
 
+      - name: Generate a changelog
+        uses: orhun/git-cliff-action@v3
+        id: git-cliff
+        with:
+          config: cliff.toml
+          args: --verbose
+        env:
+          GITHUB_REPO: ${{ github.repository }}
+
+      - name: Print the changelog
+        run: cat "${{ steps.git-cliff.outputs.changelog }}"
+
       - name: Create Release
         id: create_release
         uses: actions/create-release@v1
@@ -46,7 +61,8 @@ jobs:
         with:
           tag_name: ${{ github.ref }}
           release_name: ${{ github.ref }}
-          draft: false
+          body: ${{ steps.git-cliff.outputs.changelog }}
+          draft: true
           prerelease: false
 
       - name: Upload zip file

cliff.toml

@@ -0,0 +1,12 @@
+[changelog]
+header = "Changelog"
+body = """
+{% for group, commits in commits | group_by(attribute="group") %}
+    ### {{ group | upper_first }}
+    {% for commit in commits %}
+        - {{ commit.message | upper_first }}
+    {% endfor %}
+{% endfor %}
+"""
+trim = true
+footer = "<!-- generated by git-cliff -->"

package.json

@@ -1,6 +1,6 @@
 {
   "name": "scambier.obsidian-search",
-  "version": "1.23.1",
+  "version": "1.24.0-beta.3",
   "description": "A search engine for Obsidian",
   "main": "dist/main.js",
   "scripts": {
@@ -52,6 +52,5 @@
     "overrides": {
       "moment@>=2.18.0 <2.29.4": ">=2.29.4"
     }
-  },
+  }
-  "packageManager": "pnpm@9.1.0+sha512.67f5879916a9293e5cf059c23853d571beaf4f753c707f40cb22bed5fb1578c6aad3b6c4107ccb3ba0b35be003eb621a16471ac836c87beb53f9d54bb4612724"
 }

src/__tests__/query-tests.ts

@@ -6,7 +6,10 @@ describe('The Query class', () => {
 
   it('should correctly parse string queries', () => {
     // Act
-    const query = new Query(stringQuery, { ignoreDiacritics: true })
+    const query = new Query(stringQuery, {
+      ignoreDiacritics: true,
+      ignoreArabicDiacritics: true,
+    })
 
     // Assert
     const segments = query.query.text
@@ -25,7 +28,10 @@ describe('The Query class', () => {
 
   it('should not exclude words when there is no space before', () => {
     // Act
-    const query = new Query('foo bar-baz', { ignoreDiacritics: true })
+    const query = new Query('foo bar-baz', {
+      ignoreDiacritics: true,
+      ignoreArabicDiacritics: true,
+    })
 
     // Assert
     expect(query.query.exclude.text).toHaveLength(0)
@@ -34,7 +40,10 @@ describe('The Query class', () => {
   describe('.getExactTerms()', () => {
     it('should an array of strings containg "exact" values', () => {
       // Act
-      const query = new Query(stringQuery, { ignoreDiacritics: true })
+      const query = new Query(stringQuery, {
+        ignoreDiacritics: true,
+        ignoreArabicDiacritics: true,
+      })
 
       // Assert
       expect(query.getExactTerms()).toEqual(['lorem ipsum', 'sit amet'])

src/cache-manager.ts

@@ -5,7 +5,7 @@ import {
   getAliasesFromMetadata,
   getTagsFromMetadata,
   isFileCanvas,
-  isFileFromDataloomPlugin,
+  isFileFromDataloom,
   isFileImage,
   isFileOffice,
   isFilePDF,
@@ -136,7 +136,7 @@ export class CacheManager {
     }
 
     // ** Dataloom plugin **
-    else if (isFileFromDataloomPlugin(path)) {
+    else if (isFileFromDataloom(path)) {
       try {
         const data = JSON.parse(await app.vault.cachedRead(file))
         // data is a json object, we recursively iterate the keys
@@ -230,10 +230,11 @@ export class CacheManager {
         }
       }
     }
-
+    const displayTitle = metadata?.frontmatter?.[this.plugin.settings.displayTitle] ?? ''
     const tags = getTagsFromMetadata(metadata)
     return {
       basename: file.basename,
+      displayTitle,
       content,
       /** Content without diacritics and markdown chars */
       cleanedContent: stripMarkdownCharacters(removeDiacritics(content)),

src/components/ModalVault.svelte

@@ -141,6 +141,7 @@
     }
     query = new Query(searchQuery, {
       ignoreDiacritics: plugin.settings.ignoreDiacritics,
+      ignoreArabicDiacritics: plugin.settings.ignoreArabicDiacritics,
     })
     cancelableQuery = cancelable(
       new Promise(resolve => {

src/components/ResultItemVault.svelte

@@ -3,7 +3,7 @@
   import type { ResultNote } from '../globals'
   import {
     getExtension,
-    isFileCanvas,
+    isFileCanvas, isFileExcalidraw,
     isFileImage,
     isFilePDF,
     pathWithoutFilename,
@@ -36,7 +36,7 @@
   $: cleanedContent = plugin.textProcessor.makeExcerpt(note.content, note.matches[0]?.offset ?? -1)
   $: glyph = false //cacheManager.getLiveDocument(note.path)?.doesNotExist
   $: {
-    title = note.basename
+    title = note.displayTitle || note.basename
     notePath = pathWithoutFilename(note.path)
 
     // Icons
@@ -44,11 +44,18 @@
       setIcon(elFolderPathIcon, 'folder-open')
     }
     if (elFilePathIcon) {
-      if (isFileImage(note.path)) setIcon(elFilePathIcon, 'image')
+      if (isFileImage(note.path)) {
-      else if (isFilePDF(note.path)) setIcon(elFilePathIcon, 'file-text')
+        setIcon(elFilePathIcon, 'image')
-      else if (isFileCanvas(note.path))
+      }
+      else if (isFilePDF(note.path)) {
+        setIcon(elFilePathIcon, 'file-text')
+      }
+      else if (isFileCanvas(note.path) || isFileExcalidraw(note.path)) {
         setIcon(elFilePathIcon, 'layout-dashboard')
-      else setIcon(elFilePathIcon, 'file')
+      }
+      else {
+        setIcon(elFilePathIcon, 'file')
+      }
     }
   }
 </script>

src/components/modals.ts

@@ -119,7 +119,7 @@ abstract class OmnisearchModal extends Modal {
     })
 
     // Open in background
-    this.scope.register(['Alt'], 'O', e => {
+    this.scope.register(['Ctrl'], 'O', e => {
       if (!isInputComposition()) {
         // Check if the user is still typing
         e.preventDefault()

src/globals.ts

@@ -46,6 +46,7 @@ export type DocumentRef = { path: string; mtime: number }
 export type IndexedDocument = {
   path: string
   basename: string
+  displayTitle: string
   mtime: number
 
   content: string
@@ -76,6 +77,7 @@ export type ResultNote = {
   score: number
   path: string
   basename: string
+  displayTitle: string
   content: string
   foundWords: string[]
   matches: SearchMatch[]

src/main.ts

@@ -138,7 +138,9 @@ export default class OmnisearchPlugin extends Plugin {
         })
       )
 
-      this.refreshIndexCallback = this.notesIndexer.refreshIndex.bind(this.notesIndexer)
+      this.refreshIndexCallback = this.notesIndexer.refreshIndex.bind(
+        this.notesIndexer
+      )
       addEventListener('blur', this.refreshIndexCallback)
       removeEventListener
 
@@ -272,15 +274,20 @@ export default class OmnisearchPlugin extends Plugin {
       indexingStep.set(IndexingStepType.WritingCache)
 
       // Disable settings.useCache while writing the cache, in case it freezes
-      this.settings.useCache = false
+      const cacheEnabled = this.settings.useCache
-      await saveSettings(this)
+      if (cacheEnabled && !this.settings.DANGER_forceSaveCache) {
+        this.settings.useCache = false
+        await saveSettings(this)
+      }
 
       // Write the cache
       await searchEngine.writeToCache()
 
       // Re-enable settings.caching
-      this.settings.useCache = true
+      if (cacheEnabled) {
-      await saveSettings(this)
+        this.settings.useCache = true
+        await saveSettings(this)
+      }
     }
 
     console.timeEnd('Omnisearch - Indexing total time')

src/notes-indexer.ts

@@ -4,7 +4,7 @@ import { removeAnchors } from './tools/notes'
 import type { IndexedDocument } from './globals'
 import {
   isFileCanvas,
-  isFileFromDataloomPlugin,
+  isFileFromDataloom,
   isFileImage,
   isFilePDF,
   logDebug,
@@ -51,7 +51,7 @@ export class NotesIndexer {
     return (
       this.isFilePlaintext(path) ||
       isFileCanvas(path) ||
-      isFileFromDataloomPlugin(path) ||
+      isFileFromDataloom(path) ||
       (canIndexPDF && isFilePDF(path)) ||
       (canIndexImages && isFileImage(path)) ||
       (canIndexImagesAI && isFileImage(path))
@@ -63,7 +63,7 @@ export class NotesIndexer {
       this.canIndexUnsupportedFiles() ||
       this.isFilePlaintext(path) ||
       isFileCanvas(path) ||
-      isFileFromDataloomPlugin(path)
+      isFileFromDataloom(path)
     )
   }
 
@@ -91,6 +91,7 @@ export class NotesIndexer {
     return {
       path: filename,
       basename: name,
+      displayTitle: '',
       mtime: 0,
 
       content: '',

src/search/query.ts

@@ -13,9 +13,9 @@ export class Query {
   }
   #inQuotes: string[]
 
-  constructor(text = '', options: { ignoreDiacritics: boolean }) {
+  constructor(text = '', options: { ignoreDiacritics: boolean, ignoreArabicDiacritics: boolean}) {
     if (options.ignoreDiacritics) {
-      text = removeDiacritics(text)
+      text = removeDiacritics(text, options.ignoreArabicDiacritics)
     }
     const parsed = parse(text.toLowerCase(), {
       tokenize: true,

src/search/search-engine.ts

@@ -154,8 +154,9 @@ export class SearchEngine {
         term.length <= 3 ? 0 : term.length <= 5 ? fuzziness / 2 : fuzziness,
       boost: {
         basename: settings.weightBasename,
-        directory: settings.weightDirectory,
         aliases: settings.weightBasename,
+        displayTitle: settings.weightBasename,
+        directory: settings.weightDirectory,
         headings1: settings.weightH1,
         headings2: settings.weightH2,
         headings3: settings.weightH3,
@@ -304,7 +305,12 @@ export class SearchEngine {
         const title = document?.path.toLowerCase() ?? ''
         const content = (document?.cleanedContent ?? '').toLowerCase()
         return exactTerms.every(
-          q => content.includes(q) || removeDiacritics(title).includes(q)
+          q =>
+            content.includes(q) ||
+            removeDiacritics(
+              title,
+              this.plugin.settings.ignoreArabicDiacritics
+            ).includes(q)
         )
       })
     }
@@ -434,7 +440,7 @@ export class SearchEngine {
       },
       processTerm: (term: string) =>
         (this.plugin.settings.ignoreDiacritics
-          ? removeDiacritics(term)
+          ? removeDiacritics(term, this.plugin.settings.ignoreArabicDiacritics)
           : term
         ).toLowerCase(),
       idField: 'path',

src/search/tokenizer.ts

@@ -15,36 +15,41 @@ export class Tokenizer {
    * @returns
    */
   public tokenizeForIndexing(text: string): string[] {
-    const words = this.tokenizeWords(text)
+    try {
-    let urls: string[] = []
+      const words = this.tokenizeWords(text)
-    if (this.plugin.settings.tokenizeUrls) {
+      let urls: string[] = []
-      try {
+      if (this.plugin.settings.tokenizeUrls) {
-        urls = markdownLinkExtractor(text)
+        try {
-      } catch (e) {
+          urls = markdownLinkExtractor(text)
-        logDebug('Error extracting urls', e)
+        } catch (e) {
+          logDebug('Error extracting urls', e)
+        }
       }
+
+      let tokens = this.tokenizeTokens(text, { skipChs: true })
+
+      // Split hyphenated tokens
+      tokens = [...tokens, ...tokens.flatMap(splitHyphens)]
+
+      // Split camelCase tokens into "camel" and "case
+      tokens = [...tokens, ...tokens.flatMap(splitCamelCase)]
+
+      // Add whole words (aka "not tokens")
+      tokens = [...tokens, ...words]
+
+      // Add urls
+      if (urls.length) {
+        tokens = [...tokens, ...urls]
+      }
+
+      // Remove duplicates
+      tokens = [...new Set(tokens)]
+
+      return tokens
+    } catch (e) {
+      console.error('Error tokenizing text, skipping document', e)
+      return []
     }
-
-    let tokens = this.tokenizeTokens(text, { skipChs: true })
-
-    // Split hyphenated tokens
-    tokens = [...tokens, ...tokens.flatMap(splitHyphens)]
-
-    // Split camelCase tokens into "camel" and "case
-    tokens = [...tokens, ...tokens.flatMap(splitCamelCase)]
-
-    // Add whole words (aka "not tokens")
-    tokens = [...tokens, ...words]
-
-    // Add urls
-    if (urls.length) {
-      tokens = [...tokens, ...urls]
-    }
-
-    // Remove duplicates
-    tokens = [...new Set(tokens)]
-
-    return tokens
   }
 
   /**

src/settings.ts

@@ -11,7 +11,8 @@ import {
 import { writable } from 'svelte/store'
 import { K_DISABLE_OMNISEARCH } from './globals'
 import type OmnisearchPlugin from './main'
-import { enablePrintDebug } from "./tools/utils";
+import { enablePrintDebug } from './tools/utils'
+import { debounce } from 'lodash-es'
 
 interface WeightingSettings {
   weightBasename: number
@@ -32,8 +33,12 @@ export interface OmnisearchSettings extends WeightingSettings {
   downrankedFoldersFilters: string[]
   /** Ignore diacritics when indexing files */
   ignoreDiacritics: boolean
+  ignoreArabicDiacritics: boolean
+
   /** Extensions of plain text files to index, in addition to .md */
   indexedFileTypes: string[]
+  /** Custom title field */
+  displayTitle: string
   /** Enable PDF indexing */
   PDFIndexing: boolean
   /** Enable Images indexing */
@@ -71,6 +76,7 @@ export interface OmnisearchSettings extends WeightingSettings {
   httpApiNotice: boolean
 
   DANGER_httpHost: string | null
+  DANGER_forceSaveCache: boolean
 }
 
 /**
@@ -97,6 +103,11 @@ export class SettingsTab extends PluginSettingTab {
     const { containerEl } = this
     const database = this.plugin.database
     const textExtractor = this.plugin.getTextExtractor()
+
+    const clearCacheDebounced = debounce(async () => {
+      await database.clearCache()
+    }, 1000)
+
     const aiImageAnalyzer = this.plugin.getAIImageAnalyzer()
     containerEl.empty()
 
@@ -138,18 +149,24 @@ export class SettingsTab extends PluginSettingTab {
     new Setting(containerEl)
       .setName('Indexing')
       .setHeading()
-      .setDesc(indexingDesc)
+      .setDesc(
+        htmlDescription(`⚠️ <span style="color: var(--text-accent)">Changing indexing settings will clear the cache, and requires a restart of Obsidian.</span><br/><br/>
+        ${
+          textExtractor
+            ? `👍 You have installed <a href="https://github.com/scambier/obsidian-text-extractor">Text Extractor</a>, Omnisearch can use it to index PDFs and images contents.
+            <br />Text extraction only works on desktop, but the cache can be synchronized with your mobile device.`
+            : `⚠️ Omnisearch requires <a href="https://github.com/scambier/obsidian-text-extractor">Text Extractor</a> to index PDFs and images.`
+        }`)
+      )
 
     // PDF Indexing
-    const indexPDFsDesc = new DocumentFragment()
-    indexPDFsDesc.createSpan({}, span => {
-      span.innerHTML = `Omnisearch will use Text Extractor to index the content of your PDFs.`
-    })
     new Setting(containerEl)
-      .setName(
+      .setName(`PDFs content indexing ${textExtractor ? '' : '⚠️ Disabled'}`)
-        `PDFs content indexing ${textExtractor ? '' : '⚠️ Disabled'}`
+      .setDesc(
+        htmlDescription(
+          `Omnisearch will use Text Extractor to index the content of your PDFs.`
+        )
       )
-      .setDesc(indexPDFsDesc)
       .addToggle(toggle =>
         toggle.setValue(settings.PDFIndexing).onChange(async v => {
           await database.clearCache()
@@ -160,13 +177,13 @@ export class SettingsTab extends PluginSettingTab {
       .setDisabled(!textExtractor)
 
     // Images Indexing
-    const indexImagesDesc = new DocumentFragment()
-    indexImagesDesc.createSpan({}, span => {
-      span.innerHTML = `Omnisearch will use Text Extractor to OCR your images and index their content.`
-    })
     new Setting(containerEl)
       .setName(`Images OCR indexing ${textExtractor ? '' : '⚠️ Disabled'}`)
-      .setDesc(indexImagesDesc)
+      .setDesc(
+        htmlDescription(
+          `Omnisearch will use Text Extractor to OCR your images and index their content.`
+        )
+      )
       .addToggle(toggle =>
         toggle.setValue(settings.imagesIndexing).onChange(async v => {
           await database.clearCache()
@@ -213,38 +230,49 @@ export class SettingsTab extends PluginSettingTab {
       .setDisabled(!aiImageAnalyzer)
 
     // Index filenames of unsupported files
-    const indexUnsupportedDesc = new DocumentFragment()
-    indexUnsupportedDesc.createSpan({}, span => {
-      span.innerHTML = `
-      Omnisearch can index file<strong>names</strong> of "unsupported" files, such as e.g. <pre style="display:inline">.mp4</pre>
-      or non-extracted PDFs & images.<br/>
-      "Obsidian setting" will respect the value of "Files & Links > Detect all file extensions".`
-    })
     new Setting(containerEl)
       .setName('Index paths of unsupported files')
-      .setDesc(indexUnsupportedDesc)
+      .setDesc(
+        htmlDescription(`
+      Omnisearch can index file<strong>names</strong> of "unsupported" files, such as e.g. <pre style="display:inline">.mp4</pre>
+      or non-extracted PDFs & images.<br/>
+      "Obsidian setting" will respect the value of "Files & Links > Detect all file extensions".`)
+      )
       .addDropdown(dropdown => {
         dropdown
           .addOptions({ yes: 'Yes', no: 'No', default: 'Obsidian setting' })
           .setValue(settings.unsupportedFilesIndexing)
           .onChange(async v => {
-            await database.clearCache()
+            await clearCacheDebounced()
             ;(settings.unsupportedFilesIndexing as any) = v
             await saveSettings(this.plugin)
           })
       })
 
+    // Custom display title
+    new Setting(containerEl)
+      .setName('Set frontmatter property key as title')
+      .setDesc(
+        htmlDescription(`If you have a custom property in your notes that you want to use as the title in search results.<br>
+          Leave empty to disable.`)
+      )
+      .addText(component => {
+        component.setValue(settings.displayTitle).onChange(async v => {
+          await clearCacheDebounced()
+          settings.displayTitle = v
+          await saveSettings(this.plugin)
+        })
+      })
+
     // Additional text files to index
-    const indexedFileTypesDesc = new DocumentFragment()
-    indexedFileTypesDesc.createSpan({}, span => {
-      span.innerHTML = `In addition to standard <code>md</code> files, Omnisearch can also index other <strong style="color: var(--text-accent)">PLAINTEXT</strong> files.<br/>
-      Add extensions separated by a space, without the dot. Example: "<code>txt org csv</code>".<br />
-      ⚠️ <span style="color: var(--text-accent)">Using extensions of non-plaintext files (like .pptx) WILL cause crashes,
-      because Omnisearch will try to index their content.</span>`
-    })
     new Setting(containerEl)
       .setName('Additional TEXT files to index')
-      .setDesc(indexedFileTypesDesc)
+      .setDesc(
+        htmlDescription(`In addition to standard <code>md</code> files, Omnisearch can also index other <strong style="color: var(--text-accent)">PLAINTEXT</strong> files.<br/>
+      Add extensions separated by a space, without the dot. Example: "<code>txt org csv</code>".<br />
+      ⚠️ <span style="color: var(--text-accent)">Using extensions of non-plaintext files (like .pptx) WILL cause crashes,
+      because Omnisearch will try to index their content.</span>`)
+      )
       .addText(component => {
         component
           .setValue(settings.indexedFileTypes.join(' '))
@@ -319,16 +347,13 @@ export class SettingsTab extends PluginSettingTab {
       })
 
     // Split CamelCaseWords
-    const camelCaseDesc = new DocumentFragment()
-    camelCaseDesc.createSpan({}, span => {
-      span.innerHTML = `Enable this if you want to be able to search for CamelCaseWords as separate words.<br/>        
-        ⚠️ <span style="color: var(--text-accent)">Changing this setting will clear the cache.</span><br>
-        ${needsARestart}
-        `
-    })
     new Setting(containerEl)
       .setName('Split CamelCaseWords')
-      .setDesc(camelCaseDesc)
+      .setDesc(
+        htmlDescription(`Enable this if you want to be able to search for CamelCaseWords as separate words.<br/>        
+        ⚠️ <span style="color: var(--text-accent)">Changing this setting will clear the cache.</span><br>
+        ${needsARestart}`)
+      )
       .addToggle(toggle =>
         toggle.setValue(settings.splitCamelCase).onChange(async v => {
           await database.clearCache()
@@ -469,14 +494,12 @@ export class SettingsTab extends PluginSettingTab {
       )
 
     // Show "Create note" button
-    const createBtnDesc = new DocumentFragment()
-    createBtnDesc.createSpan({}, span => {
-      span.innerHTML = `Shows a button next to the search input, to create a note.
-        Acts the same as the <code>shift ↵</code> shortcut, can be useful for mobile device users.`
-    })
     new Setting(containerEl)
       .setName('Show "Create note" button')
-      .setDesc(createBtnDesc)
+      .setDesc(
+        htmlDescription(`Shows a button next to the search input, to create a note.
+        Acts the same as the <code>shift ↵</code> shortcut, can be useful for mobile device users.`)
+      )
       .addToggle(toggle =>
         toggle.setValue(settings.showCreateButton).onChange(async v => {
           settings.showCreateButton = v
@@ -589,14 +612,14 @@ export class SettingsTab extends PluginSettingTab {
     //#region HTTP Server
 
     if (!Platform.isMobile) {
-      const httpServerDesc = new DocumentFragment()
-      httpServerDesc.createSpan({}, span => {
-        span.innerHTML = `Omnisearch can be used through a simple HTTP server (<a href="https://publish.obsidian.md/omnisearch/Public+API+%26+URL+Scheme#HTTP+Server">more information</a>).`
-      })
       new Setting(containerEl)
         .setName('API Access Through HTTP')
         .setHeading()
-        .setDesc(httpServerDesc)
+        .setDesc(
+          htmlDescription(
+            `Omnisearch can be used through a simple HTTP server (<a href="https://publish.obsidian.md/omnisearch/Public+API+%26+URL+Scheme#HTTP+Server">more information</a>).`
+          )
+        )
 
       new Setting(containerEl)
         .setName('Enable the HTTP server')
@@ -668,17 +691,14 @@ export class SettingsTab extends PluginSettingTab {
     new Setting(containerEl).setName('Danger Zone').setHeading()
 
     // Ignore diacritics
-    const diacriticsDesc = new DocumentFragment()
-    diacriticsDesc.createSpan({}, span => {
-      span.innerHTML = `Normalize diacritics in search terms. Words like "brûlée" or "žluťoučký" will be indexed as "brulee" and "zlutoucky".<br/>
-        ⚠️ <span style="color: var(--text-accent)">You probably should <strong>NOT</strong> disable this.</span><br>
-        ⚠️ <span style="color: var(--text-accent)">Changing this setting will clear the cache.</span><br>
-        ${needsARestart}
-        `
-    })
     new Setting(containerEl)
       .setName('Ignore diacritics')
-      .setDesc(diacriticsDesc)
+      .setDesc(
+        htmlDescription(`Normalize diacritics in search terms. Words like "brûlée" or "žluťoučký" will be indexed as "brulee" and "zlutoucky".<br/>
+        ⚠️ <span style="color: var(--text-accent)">You probably should <strong>NOT</strong> disable this.</span><br>
+        ⚠️ <span style="color: var(--text-accent)">Changing this setting will clear the cache.</span><br>
+        ${needsARestart}`)
+      )
       .addToggle(toggle =>
         toggle.setValue(settings.ignoreDiacritics).onChange(async v => {
           await database.clearCache()
@@ -687,6 +707,16 @@ export class SettingsTab extends PluginSettingTab {
         })
       )
 
+    new Setting(containerEl)
+      .setName('Ignore Arabic diacritics (beta)')
+      .addToggle(toggle =>
+        toggle.setValue(settings.ignoreArabicDiacritics).onChange(async v => {
+          await database.clearCache()
+          settings.ignoreArabicDiacritics = v
+          await saveSettings(this.plugin)
+        })
+      )
+
     // Disable Omnisearch
     const disableDesc = new DocumentFragment()
     disableDesc.createSpan({}, span => {
@@ -700,24 +730,38 @@ export class SettingsTab extends PluginSettingTab {
         toggle.setValue(isPluginDisabled(this.app)).onChange(async v => {
           if (v) {
             this.app.saveLocalStorage(K_DISABLE_OMNISEARCH, '1')
+            new Notice('Omnisearch - Disabled. Please restart Obsidian.')
           } else {
             this.app.saveLocalStorage(K_DISABLE_OMNISEARCH) // No value = unset
+            new Notice('Omnisearch - Enabled. Please restart Obsidian.')
           }
-          new Notice('Omnisearch - Disabled. Please restart Obsidian.')
+        })
+      )
+
+    // Force save cache
+    new Setting(containerEl)
+      .setName('Force save the cache')
+      .setDesc(
+        htmlDescription(`Omnisearch has a security feature that automatically disables cache writing if it cannot fully perform the operation.<br>
+        Use this option to force the cache to be saved, even if it causes a crash.<br>
+        ⚠️ <span style="color: var(--text-accent)">Enabling this setting could lead to crash loops</span>`)
+      )
+      .addToggle(toggle =>
+        toggle.setValue(settings.DANGER_forceSaveCache).onChange(async v => {
+          settings.DANGER_forceSaveCache = v
+          await saveSettings(this.plugin)
         })
       )
 
     // Clear cache data
     if (isCacheEnabled()) {
-      const resetCacheDesc = new DocumentFragment()
-      resetCacheDesc.createSpan({}, span => {
-        span.innerHTML = `Erase all Omnisearch cache data.
-          Use this if Omnisearch results are inconsistent, missing, or appear outdated.<br>
-          ${needsARestart}`
-      })
       new Setting(containerEl)
         .setName('Clear cache data')
-        .setDesc(resetCacheDesc)
+        .setDesc(
+          htmlDescription(`Erase all Omnisearch cache data.
+          Use this if Omnisearch results are inconsistent, missing, or appear outdated.<br>
+          ${needsARestart}`)
+        )
         .addButton(btn => {
           btn.setButtonText('Clear cache')
           btn.onClick(async () => {
@@ -725,6 +769,7 @@ export class SettingsTab extends PluginSettingTab {
           })
         })
     }
+
     //#endregion Danger Zone
   }
 
@@ -745,7 +790,9 @@ export function getDefaultSettings(app: App): OmnisearchSettings {
     hideExcluded: false,
     downrankedFoldersFilters: [] as string[],
     ignoreDiacritics: true,
+    ignoreArabicDiacritics: false,
     indexedFileTypes: [] as string[],
+    displayTitle: '',
     PDFIndexing: false,
     officeIndexing: false,
     imagesIndexing: false,
@@ -781,11 +828,20 @@ export function getDefaultSettings(app: App): OmnisearchSettings {
     verboseLogging: false,
 
     DANGER_httpHost: null,
+    DANGER_forceSaveCache: false,
   }
 }
 
 let settings: OmnisearchSettings
 
+function htmlDescription(innerHTML: string): DocumentFragment {
+  const desc = new DocumentFragment()
+  desc.createSpan({}, span => {
+    span.innerHTML = innerHTML
+  })
+  return desc
+}
+
 // /**
 //  * @deprecated
 //  */

src/tools/api.ts

@@ -88,6 +88,7 @@ export function getApi(plugin: OmnisearchPlugin) {
     async search(q: string): Promise<ResultNoteApi[]> {
       const query = new Query(q, {
         ignoreDiacritics: plugin.settings.ignoreDiacritics,
+        ignoreArabicDiacritics: plugin.settings.ignoreArabicDiacritics,
       })
       const raw = await plugin.searchEngine.getSuggestions(query)
       return mapResults(plugin, raw)

src/tools/text-processing.ts

@@ -115,7 +115,7 @@ export class TextProcessor {
     const originalText = text
     // text = text.toLowerCase().replace(new RegExp(SEPARATORS, 'gu'), ' ')
     if (this.plugin.settings.ignoreDiacritics) {
-      text = removeDiacritics(text)
+      text = removeDiacritics(text, this.plugin.settings.ignoreArabicDiacritics)
     }
     const startTime = new Date().getTime()
     let match: RegExpExecArray | null = null

src/tools/utils.ts

@@ -109,15 +109,30 @@ export function getTagsFromMetadata(metadata: CachedMetadata | null): string[] {
 /**
  * https://stackoverflow.com/a/37511463
  */
-export function removeDiacritics(str: string): string {
+export function removeDiacritics(str: string, arabic = false): string {
-  // Japanese diacritics that should be distinguished
-  const excludeDiacritics: string[] = ['\\u30FC', '\\u309A', '\\u3099']
-  const regexpExclude: string = excludeDiacritics.join('|')
-  const regexp: RegExp = new RegExp(`(?!${regexpExclude})\\p{Diacritic}`, 'gu')
-
   if (str === null || str === undefined) {
     return ''
   }
+
+  // Japanese diacritics that should be distinguished
+  const japaneseDiacritics: string[] = ['\\u30FC', '\\u309A', '\\u3099']
+  const regexpExclude: string = japaneseDiacritics.join('|')
+  const regexp: RegExp = new RegExp(`(?!${regexpExclude})\\p{Diacritic}`, 'gu')
+
+  if (arabic) {
+    // Arabic diacritics
+    // https://stackoverflow.com/a/40959537
+    str = str
+      .replace(/([^\u0621-\u063A\u0641-\u064A\u0660-\u0669a-zA-Z 0-9])/g, '')
+      .replace(/(آ|إ|أ)/g, 'ا')
+      .replace(/(ة)/g, 'ه')
+      .replace(/(ئ|ؤ)/g, 'ء')
+      .replace(/(ى)/g, 'ي')
+    for (let i = 0; i < 10; i++) {
+      str.replace(String.fromCharCode(0x660 + i), String.fromCharCode(48 + i))
+    }
+  }
+
   // Keep backticks for code blocks, because otherwise they are removed by the .normalize() function
   // https://stackoverflow.com/a/36100275
   str = str.replaceAll('`', '[__omnisearch__backtick__]')
@@ -152,7 +167,11 @@ export function isFileCanvas(path: string): boolean {
   return path.endsWith('.canvas')
 }
 
-export function isFileFromDataloomPlugin(path: string): boolean {
+export function isFileExcalidraw(path: string): boolean {
+  return path.endsWith('.excalidraw')
+}
+
+export function isFileFromDataloom(path: string): boolean {
   return path.endsWith('.loom')
 }
 
@@ -219,7 +238,7 @@ export function warnDebug(...args: any[]): void {
   printDebug(console.warn, ...args)
 }
 
-let printDebugEnabled= false
+let printDebugEnabled = false
 export function enablePrintDebug(enable: boolean): void {
   printDebugEnabled = enable
 }