Update to pdfjs v4

requires ESM modules and dropping support for nodejs v16

.eslintrc.js → .eslintrc.cjs

@@ -5,6 +5,7 @@ module.exports = {
   extends: ["kyle"],
   rules: {
     quotes: "off",
+    "import/extensions": "off",
   },
   settings: {
     jest: { version: 29 },

.github/workflows/test.yml

@@ -10,7 +10,7 @@ jobs:
 
     strategy:
       matrix:
-        node-version: [16.17, 18.x, 20.x, 21.x]
+        node-version: [18.x, 20.x, 22.x]
 
     steps:
       - name: ⏬ Checkout code

CHANGELOG.md

@@ -7,6 +7,11 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## 4.0.0 (2024-06-08)
+
+- 💥 BREAKING CHANGE: Drop support for node v16. The minimum version is now v18
+- Updated pdfjs to v4
+
 ## 3.0.0 (2024-05-24)
 
 - 💥 BREAKING CHANGE: Drop support for node v14 and v16. The minimum version is now v16.17

README.md

@@ -10,17 +10,22 @@
 
 Useful for unit tests of PDFs
 
-Supports nodejs v16.17+, and comes with a CLI.
+Supports nodejs v18+, and comes with a CLI.
 
 ## Install
 
 ```sh
 npm install -S pdf-to-img
 ```
 
+> [!IMPORTANT]
+> You should use v4 by default. v4 requires nodejs v18 or later, and ESM modules.
+>
+> If you can't upgrade to v4 yet, you can still use v3. If you use v3, you can safely ignore `npm audit`'s [warning about pdfjs-dist](https://github.com/advisories/GHSA-wgrm-67xf-hhpq), since this library [disables `eval` by default](https://github.com/k-yle/pdf-to-img/commit/bdac3a1dcc2004c3f1fe7380bbb860086ec2746f).
+
 ## Example
 
-NodeJS:
+NodeJS (using ESM Modules):
 
 ```js
 const { promises: fs } = require("node:fs");
@@ -37,10 +42,28 @@ async function main() {
 main();
 ```
 
+If your app does not support ESM modules, just change the import:
+
+```diff
+  const { promises: fs } = require("node:fs");
+- const { pdf } = require("pdf-to-img");
+
+  async function main() {
++   const { pdf } = await import("pdf-to-img");
+    let counter = 1;
+    const document = await pdf("example.pdf", { scale: 3 });
+    for await (const image of document) {
+      await fs.writeFile(`page${counter}.png`, image);
+      counter++;
+    }
+  }
+  main();
+```
+
 Using jest (or vitest) with [jest-image-snapshot](https://npm.im/jest-image-snapshot):
 
 ```js
-const { pdf } = require("pdf-to-img");
+import { pdf } from "pdf-to-img";
 
 it("generates a PDF", async () => {
   for await (const page of await pdf("example.pdf")) {

package.json

@@ -3,8 +3,9 @@
   "version": "3.0.0",
   "author": "Kyle Hensel",
   "description": "📃📸 Converts PDFs to images in nodejs",
-  "main": "dist",
+  "exports": "./dist/index.js",
   "types": "dist/index.d.ts",
+  "type": "module",
   "license": "MIT",
   "files": [
     "dist"
@@ -28,11 +29,12 @@
     "pdf2img": "./bin/cli.mjs"
   },
   "engines": {
-    "node": ">=16.17"
+    "node": ">=18"
   },
+  "engineStrict": true,
   "dependencies": {
     "canvas": "2.11.2",
-    "pdfjs-dist": "3.2.146"
+    "pdfjs-dist": "4.2.67"
   },
   "devDependencies": {
     "@rushstack/eslint-patch": "^1.5.1",

src/index.ts

@@ -1,13 +1,11 @@
 import "./polyfill"; // do this before pdfjs
 
 import path from "node:path";
-// 🛑 inspite of esModuleInterop being on, you still need to use `import *`, and there are no typedefs
-import * as _pdfjs from "pdfjs-dist/legacy/build/pdf";
-import type { DocumentInitParameters } from "pdfjs-dist/types/src/display/api";
-import { NodeCanvasFactory } from "./canvasFactory";
-import { parseInput } from "./parseInput";
+import * as pdfjs from "pdfjs-dist/legacy/build/pdf.mjs";
+import type { DocumentInitParameters } from "pdfjs-dist/types/src/display/api.js";
+import { NodeCanvasFactory } from "./canvasFactory.js";
+import { parseInput } from "./parseInput.js";
 
-const pdfjs: typeof import("pdfjs-dist") = _pdfjs;
 const pdfjsPath = path.dirname(require.resolve("pdfjs-dist/package.json"));
 
 /** required since k-yle/pdf-to-img#58, the objects from pdfjs are weirdly structured */
@@ -46,7 +44,7 @@ export type Options = {
 /**
  * Converts a PDF to a series of images. This returns a `Symbol.asyncIterator`
  *
- * @param input Either (a) the path to a pdf file, or (b) a data url, or (c) a buffer, or (d) a ReadableStream.
+ * @param input Either (a) the path to a pdf file, or (b) a data url, or (b) a buffer, (c) a buffer, or (e) a ReadableStream.
  *
  * @example
  * ```js
@@ -68,7 +66,7 @@ export type Options = {
  * ```
  */
 export async function pdf(
-  input: string | Buffer | NodeJS.ReadableStream,
+  input: string | Uint8Array | Buffer | NodeJS.ReadableStream,
   options: Options = {}
 ): Promise<{
   length: number;
@@ -77,13 +75,15 @@ export async function pdf(
 }> {
   const data = await parseInput(input);
 
+  const canvasFactory = new NodeCanvasFactory();
   const pdfDocument = await pdfjs.getDocument({
     password: options.password, // retain for backward compatibility, but ensure settings from docInitParams overrides this and others, if given.
     standardFontDataUrl: path.join(pdfjsPath, `standard_fonts${path.sep}`),
     cMapUrl: path.join(pdfjsPath, `cmaps${path.sep}`),
     cMapPacked: true,
-    isEvalSupported: false,
     ...options.docInitParams,
+    isEvalSupported: false,
+    canvasFactory,
     data,
   }).promise;
 
@@ -102,7 +102,6 @@ export async function pdf(
 
             const viewport = page.getViewport({ scale: options.scale ?? 1 });
 
-            const canvasFactory = new NodeCanvasFactory();
             const { canvas, context } = canvasFactory.create(
               viewport.width,
               viewport.height
@@ -111,7 +110,6 @@ export async function pdf(
             await page.render({
               canvasContext: context,
               viewport,
-              canvasFactory,
             }).promise;
 
             return { done: false, value: canvas.toBuffer() };

src/parseInput.ts

@@ -1,36 +1,31 @@
-import { promises as fs } from "node:fs";
+import { readFileSync } from "node:fs";
+import { arrayBuffer } from "node:stream/consumers";
 
 const PREFIX = "data:application/pdf;base64,";
 
-async function streamToBuffer(
-  readableStream: NodeJS.ReadableStream
-): Promise<Buffer> {
-  const chunks: Buffer[] = [];
-  for await (const chunk of readableStream) {
-    chunks.push(chunk as Buffer);
-  }
-  return Buffer.concat(chunks);
-}
-
 export async function parseInput(
-  input: string | Buffer | NodeJS.ReadableStream
-): Promise<Buffer | Uint8Array> {
+  input: string | Uint8Array | Buffer | NodeJS.ReadableStream
+): Promise<Uint8Array> {
+  // Buffer is a subclass of Uint8Array, but it's not actually
+  // compatible: https://github.com/sindresorhus/uint8array-extras/issues/4
+  if (Buffer.isBuffer(input)) return Uint8Array.from(input);
+
+  if (input instanceof Uint8Array) return input;
+
   // provided with a data url or a path to a file on disk
   if (typeof input === "string") {
-    return input.startsWith(PREFIX)
-      ? Buffer.from(input.slice(PREFIX.length), "base64")
-      : new Uint8Array(await fs.readFile(input));
+    if (input.startsWith(PREFIX)) {
+      return Uint8Array.from(Buffer.from(input.slice(PREFIX.length), "base64"));
+    }
+    return new Uint8Array(readFileSync(input));
   }
 
-  // provided a buffer
-  if (Buffer.isBuffer(input)) return input;
-
   // provided a ReadableStream (or any object with an asyncIterator that yields buffer chunks)
   if (typeof input === "object" && input && Symbol.asyncIterator in input) {
-    return streamToBuffer(input);
+    return new Uint8Array(await arrayBuffer(input));
   }
 
   throw new Error(
-    "pdf-to-img received an unexpected input. Provide a path to file, a data URL, a Buffer, or a ReadableStream."
+    "pdf-to-img received an unexpected input. Provide a path to file, a data URL, a Uint8Array, a Buffer, or a ReadableStream."
   );
 }

tests/jsdom.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment jsdom
 import { promises as fs, createReadStream } from "node:fs";
 import { describe, expect, it } from "vitest";
-import { pdf } from "../src";
+import { pdf } from "../src/index.js";
 
 describe("example.pdf", () => {
   it("correctly generates a single png for the one page", async () => {
@@ -139,7 +139,7 @@ describe("invalid", () => {
       async () => pdf(1)
     ).rejects.toThrow(
       new Error(
-        "pdf-to-img received an unexpected input. Provide a path to file, a data URL, a Buffer, or a ReadableStream."
+        "pdf-to-img received an unexpected input. Provide a path to file, a data URL, a Uint8Array, a Buffer, or a ReadableStream."
       )
     );
   });

tests/node.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment node
 import { createReadStream, promises as fs } from "node:fs";
 import { describe, expect, it } from "vitest";
-import { pdf } from "../src";
+import { pdf } from "../src/index.js";
 
 describe("example.pdf in node", () => {
   it("correctly generates a single png for the one page in nodejs environment", async () => {

tsconfig.json

@@ -4,9 +4,9 @@
     "esModuleInterop": true,
     "lib": ["es2021"],
     "skipLibCheck": true,
-    "module": "commonjs",
-    "target": "es2015",
-    "moduleResolution": "node",
+    "module": "node16",
+    "target": "es2020",
+    "moduleResolution": "node16",
     "downlevelIteration": true,
     "declaration": true,
     "outDir": "temp"