#58 - Experimental PDF support

src/components/ModalInFile.svelte

@@ -27,7 +27,7 @@
 
   let groupedOffsets: number[] = []
   let selectedIndex = 0
-  let note: ResultNote | null = null
+  let note: ResultNote | undefined
   let query: Query
 
   onMount(() => {

src/pdf-parser.ts

@@ -0,0 +1,18 @@
+import PDFJs from 'pdfjs-dist'
+import pdfjsWorker from 'pdfjs-dist/build/pdf.worker.entry'
+import type { TextItem } from 'pdfjs-dist/types/src/display/api'
+import type { TFile } from 'obsidian'
+
+PDFJs.GlobalWorkerOptions.workerSrc = pdfjsWorker
+
+// https://stackoverflow.com/a/59929946
+export async function getPdfText(file: TFile): Promise<string> {
+  const data = await app.vault.readBinary(file)
+  const doc = await PDFJs.getDocument(data).promise
+  const pageTexts = Array.from({ length: doc.numPages }, async (v, i) => {
+    const page = await doc.getPage(i + 1)
+    const content = await page.getTextContent()
+    return (content.items as TextItem[]).map(token => token.str).join('')
+  })
+  return (await Promise.all(pageTexts)).join('')
+}

src/search.ts

@@ -19,7 +19,7 @@ import {
   wait,
 } from './utils'
 import type { Query } from './query'
-import { settings  } from './settings'
+import { settings } from './settings'
 import {
   removeNoteFromCache,
   getNoteFromCache,
@@ -32,6 +32,7 @@ import {
   saveNotesCacheToFile,
   isCacheOutdated,
 } from './notes'
+import { getPdfText } from './pdf-parser'
 
 let minisearchInstance: MiniSearch<IndexedNote>
 let isIndexChanged: boolean
@@ -326,8 +327,13 @@ export async function addToIndex(file: TAbstractFile): Promise<void> {
       throw new Error(`${file.basename} is already indexed`)
     }
 
-    // Fetch content from the cache to index it as-is
-    const content = removeDiacritics(await app.vault.cachedRead(file))
+    let content
+    if (file.path.endsWith('.pdf')) {
+      content = removeDiacritics(await getPdfText(file as TFile))
+    } else {
+      // Fetch content from the cache to index it as-is
+      content = removeDiacritics(await app.vault.cachedRead(file))
+    }
 
     // Make the document and index it
     const note: IndexedNote = {

src/types-obsidian.d.ts

@@ -0,0 +1,24 @@
+import type { MetadataCache, ViewState, Vault } from 'obsidian'
+
+declare module 'obsidian' {
+  interface MetadataCache {
+    isUserIgnored?(path: string): boolean
+  }
+
+  interface FrontMatterCache {
+    aliases?: string[] | string
+    tags?: string[] | string
+  }
+
+  interface ViewState {
+    state?: {
+      file?: string
+    }
+  }
+
+  interface Vault {
+    getConfig(string): unknown
+  }
+}
+
+

src/types.d.ts

@@ -1,22 +1 @@
-import type { MetadataCache, ViewState, Vault } from 'obsidian'
-
-declare module 'obsidian' {
-  interface MetadataCache {
-    isUserIgnored?(path: string): boolean
-  }
-
-  interface FrontMatterCache {
-    aliases?: string[] | string
-    tags?: string[] | string
-  }
-
-  interface ViewState {
-    state?: {
-      file?: string
-    }
-  }
-
-  interface Vault {
-    getConfig(string): unknown
-  }
-}
+declare module 'pdfjs-dist/build/pdf.worker.entry';

src/utils.ts

@@ -174,7 +174,7 @@ export function getCtrlKeyLabel(): 'ctrl' | '⌘' {
 
 export function isFileIndexable(path: string): boolean {
   return (
-    path.endsWith('.md') ||
+    path.endsWith('.md') || path.endsWith('.pdf') ||
     settings.indexedFileTypes.some(t => path.endsWith(`.${t}`))
   )
 }