Update PDFJS to v4 (IN-11761)

.github/workflows/main.yml

@@ -15,7 +15,7 @@ jobs:
       - name: Setup Node
         uses: actions/setup-node@v1
         with:
-          node-version: '14'
+          node-version: '21'
       - name: Cache Node.js modules
         uses: actions/cache@v2
         with:

package.json

@@ -7,12 +7,13 @@
    "dependencies": {
       "canvas": "^2.11.0",
       "commander": "~10.0.0",
-      "pdfjs-dist": "~3.3.122"
+      "pdfjs-dist": "^4.6.82"
    },
    "devDependencies": {
       "eslint": "^8.33.0",
       "mocha": "^10.2.0"
    },
+   "type": "module",
    "scripts": {
       "test": "mocha"
    },

src/app.js

@@ -1,10 +1,12 @@
 #!/usr/bin/env node
 
-const { readFile, writeFile } = require("fs").promises;
-const { basename, join } = require("path");
-const { getThumbnail, NodeCanvasFactory } = require("./lib");
+import { readFile, writeFile } from "fs/promises";
 
-const { Command, InvalidOptionArgumentError } = require("commander");
+import { basename, join } from "path";
+
+import { getThumbnail, NodeCanvasFactory } from "./lib.js";
+
+import { Command, InvalidOptionArgumentError } from "commander";
 
 const program = new Command();
 
@@ -37,19 +39,9 @@ program
     const canvasFactory = new NodeCanvasFactory();
     try {
       for (const file of files) {
-        const outputPath = join(
-          options.output,
-          basename(file, ".pdf") + ".jpg"
-        );
+        const outputPath = join(options.output, basename(file, ".pdf") + ".jpg");
         const contents = await readFile(file);
-        const jpegBuffer = await getThumbnail(
-          contents,
-          options.pagenum,
-          options.width,
-          options.quality,
-          options.standardFonts,
-          canvasFactory
-        );
+        const jpegBuffer = await getThumbnail(new Uint8Array(contents), options.pagenum, options.width, options.quality, options.standardFonts, canvasFactory);
 
         await writeFile(outputPath, jpegBuffer);
       }

src/lib.js

@@ -1,10 +1,8 @@
-const assert = require("assert").strict;
+import assert from "assert";
 
-const Canvas = require("canvas");
-const pdfjsLib = require("pdfjs-dist/legacy/build/pdf.js");
+import Canvas from "canvas";
 
-// todo: Use this once Node.js is upgraded to 16
-// const pdfjsLib = require('pdfjs-dist')
+import * as pdfjsLib from "pdfjs-dist/legacy/build/pdf.mjs";
 
 /* Largely copied from https://github.com/mozilla/pdf.js/blob/55f55f58594b9a6947fecaabf8ef4e3b02002023/examples/node/pdf2png/pdf2png.js#L20 */
 class NodeCanvasFactory {
@@ -37,14 +35,7 @@ class NodeCanvasFactory {
   }
 }
 
-async function getThumbnail(
-  data,
-  pageNum = 1,
-  width = 300,
-  quality = 1.0,
-  standardFonts = "/tmp/fonts/",
-  canvasFactory = new NodeCanvasFactory()
-) {
+async function getThumbnail(data, pageNum = 1, width = 300, quality = 1.0, standardFonts = "/tmp/fonts/", canvasFactory = new NodeCanvasFactory()) {
   const loadingTask = pdfjsLib.getDocument({
     data: data,
     standardFontDataUrl: standardFonts,
@@ -56,10 +47,7 @@ async function getThumbnail(
   const scale = width / viewport.width;
   viewport = page.getViewport({ scale: scale });
 
-  const canvasAndContext = canvasFactory.create(
-    viewport.width,
-    viewport.height
-  );
+  const canvasAndContext = canvasFactory.create(viewport.width, viewport.height);
   const renderContext = {
     canvasContext: canvasAndContext.context,
     viewport: viewport,
@@ -75,7 +63,4 @@ async function getThumbnail(
   });
 }
 
-module.exports = {
-  NodeCanvasFactory,
-  getThumbnail,
-};
+export { NodeCanvasFactory, getThumbnail };

test/test.js

@@ -1,17 +1,23 @@
-const assert = require("assert");
-const path = require("path");
+import assert from "assert";
 
-const { readFile, writeFile } = require("fs").promises;
+import path from "path";
+import { fileURLToPath } from 'url';
+import { dirname } from 'path';
 
-const { getThumbnail } = require("../src/lib");
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+
+import { readFile, writeFile } from "fs/promises";
+
+import { getThumbnail } from "../src/lib.js";
 
 describe("ProcessFeaturedArticle", function () {
   it("should successfully generate from wikipedia featured article", async function () {
     const data = await readFile("test/Herja.pdf");
-    const jpegBuffer = await getThumbnail(data);
+    const jpegBuffer = await getThumbnail(new Uint8Array(data));
 
     await writeFile(path.join(__dirname, "thumbnail.jpg"), jpegBuffer);
 
-    assert.strictEqual(jpegBuffer.length, 51013);
+    assert.strictEqual(jpegBuffer.length, 51133);
   });
 });