Using PDFjs directly from Obsidian

This commit is contained in:
Simon Cambier
2022-09-30 20:49:49 +02:00
parent 30b77fa4d8
commit 3f9d3c4f80
3 changed files with 6 additions and 28 deletions

View File

@@ -37,8 +37,7 @@
},
"dependencies": {
"@vanakat/plugin-api": "^0.1.0",
"minisearch": "^5.0.0",
"pdfjs-dist": "^2.16.105"
"minisearch": "^5.0.0"
},
"pnpm": {
"overrides": {

23
pnpm-lock.yaml generated
View File

@@ -19,7 +19,6 @@ specifiers:
jest: ^27.5.1
minisearch: ^5.0.0
obsidian: latest
pdfjs-dist: ^2.16.105
prettier: ^2.7.1
prettier-plugin-svelte: ^2.7.0
svelte: ^3.50.1
@@ -31,7 +30,6 @@ specifiers:
dependencies:
'@vanakat/plugin-api': 0.1.0
minisearch: 5.0.0
pdfjs-dist: 2.16.105
devDependencies:
'@babel/preset-env': 7.19.0
@@ -2438,10 +2436,6 @@ packages:
webidl-conversions: 5.0.0
dev: true
/dommatrix/1.0.3:
resolution: {integrity: sha512-l32Xp/TLgWb8ReqbVJAFIvXmY7go4nTxxlWiAFyhoQw9RKEOHBZNnyGvJWqDVSPmq3Y9HlM4npqF/T6VMOXhww==}
dev: false
/electron-to-chromium/1.4.247:
resolution: {integrity: sha512-FLs6R4FQE+1JHM0hh3sfdxnYjKvJpHZyhQDjc2qFq/xFvmmRt/TATNToZhrcGUFzpF2XjeiuozrA8lI0PZmYYw==}
dev: true
@@ -3926,18 +3920,6 @@ packages:
resolution: {integrity: sha512-gDKb8aZMDeD/tZWs9P6+q0J9Mwkdl6xMV8TjnGP3qJVJ06bdMgkbBlLU8IdfOsIsFz2BW1rNVT3XuNEl8zPAvw==}
engines: {node: '>=8'}
/pdfjs-dist/2.16.105:
resolution: {integrity: sha512-J4dn41spsAwUxCpEoVf6GVoz908IAA3mYiLmNxg8J9kfRXc2jxpbUepcP0ocp0alVNLFthTAM8DZ1RaHh8sU0A==}
peerDependencies:
worker-loader: ^3.0.8
peerDependenciesMeta:
worker-loader:
optional: true
dependencies:
dommatrix: 1.0.3
web-streams-polyfill: 3.2.1
dev: false
/picocolors/1.0.0:
resolution: {integrity: sha512-1fygroTLlHu66zi26VoTDv8yRgm0Fccecssto+MhsZ0D/DGW2sm8E8AjW7NU5VVTRt5GxbeZ5qBuJr+HyLYkjQ==}
dev: true
@@ -4577,11 +4559,6 @@ packages:
makeerror: 1.0.12
dev: true
/web-streams-polyfill/3.2.1:
resolution: {integrity: sha512-e0MO3wdXWKrLbL0DgGnUV7WHVuw9OUvL4hjgnPkIeEvESk74gAITi5G606JtZPp39cd8HA9VQzCIvA49LpPN5Q==}
engines: {node: '>= 8'}
dev: false
/webidl-conversions/5.0.0:
resolution: {integrity: sha512-VlZwKPCkYKxQgeSbH5EyngOmRp7Ww7I9rQLERETtf5ofd9pGeswWiOtogpEO850jziPRarreGxn5QIiTqpb2wA==}
engines: {node: '>=8'}

View File

@@ -1,12 +1,14 @@
import PDFJs from 'pdfjs-dist'
import pdfjsWorker from 'pdfjs-dist/build/pdf.worker.entry'
// 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'
import {loadPdfJs} from "obsidian";
PDFJs.GlobalWorkerOptions.workerSrc = pdfjsWorker
let PDFJs: any = null
// https://stackoverflow.com/a/59929946
export async function getPdfText(file: TFile): Promise<string> {
PDFJs = PDFJs ?? await loadPdfJs()
const data = await app.vault.readBinary(file)
const doc = await PDFJs.getDocument(data).promise
const pageTexts = Array.from({ length: doc.numPages }, async (v, i) => {