Add PDF worker build and path utilities

Introduces worker build configuration, source and path utilities for PDF worker scripts. Updates package.json to include new worker entry points and build script, and adds supporting TypeScript and declaration files for worker module handling.

Author: mehmet-kozan

.npmrc

@@ -1,6 +1,6 @@
 package-lock=false
 audit=false
 fund=false
-loglevel=warn
+#loglevel=warn
 prefer-offline=false
 save=false

bin/worker.cjs

@@ -0,0 +1,12 @@
+const { join } = require('node:path');
+
+/**
+ * Gets the absolute path to the PDF worker file.
+ *
+ * @returns {string} The absolute path to the PDF worker module (pdf.worker.mjs)
+ */
+function getPath() {
+	return join(__dirname, '../dist/node/pdf.worker.mjs');
+}
+
+exports.getPath = getPath;

bin/worker.js

@@ -0,0 +1,14 @@
+import { dirname, join } from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+
+/**
+ * Gets the absolute path to the PDF worker file.
+ *
+ * @returns {string} The absolute path to the PDF worker module (pdf.worker.mjs)
+ */
+export function getPath() {
+	return join(__dirname, '../dist/node/pdf.worker.mjs');
+}

package.json

@@ -1,6 +1,6 @@
 {
 	"name": "pdf-parse",
-	"version": "2.3.5",
+	"version": "2.3.6",
 	"type": "module",
 	"main": "dist/cjs/index.cjs",
 	"module": "dist/esm/index.js",
@@ -14,6 +14,14 @@
 		"./node": {
 			"types": "./dist/esm/index.d.ts",
 			"default": "./dist/node/index.cjs"
+		},
+		"./worker/source": {
+			"import": "./dist/worker/source.js",
+			"require": "./dist/worker/source.cjs"
+		},
+		"./worker/path": {
+			"import": "./bin/worker.js",
+			"require": "./bin/worker.cjs"
 		}
 	},
 	"description": "Pure TypeScript, cross-platform module for extracting text, images, and tabular data from PDFs. Run directly in your browser or in Node!",
@@ -56,16 +64,18 @@
 	"author": "Mehmet Kozan",
 	"files": [
 		"dist/",
-		"src/"
+		"src/",
+		"bin/"
 	],
 	"scripts": {
-		"build": "npm-run-all clean build:esm build:cjs build:cjs:rename build:node build:browser build:browser:min",
+		"build": "npm-run-all clean build:esm build:cjs build:cjs:rename build:node build:browser build:browser:min build:worker",
 		"build:esm": "tsc",
 		"build:cjs": "tsc --project tsconfig.node.json",
 		"build:cjs:rename": "node scripts/rename-cjs.js",
 		"build:node": "vite build --config vite.config.node.ts",
 		"build:browser": "vite build --config vite.config.browser.ts",
 		"build:browser:min": "vite build --config vite.config.browser.min.ts",
+		"build:worker": "vite build --config vite.config.worker.ts",
 		"clean:build": "rimraf dist",
 		"clean:site": "rimraf reports_site/test-report reports_site/coverage reports_site/live_demo/dist-browser",
 		"clean:test": "rimraf --glob dist test/**/*.txt test/**/imgs test/**/*_images",

src_worker/index.ts

@@ -0,0 +1,9 @@
+import * as WorkerUrl from 'pdfjs-dist/legacy/build/pdf.worker.mjs?url';
+
+/**
+ * Gets base64 data URL string for  worker pdf.worker.mjs script.
+ * @returns {string} The worker base64 data URL string.
+ */
+export function getSource() {
+	return WorkerUrl.default || WorkerUrl;
+}

src_worker/vite-url.d.ts

@@ -0,0 +1,4 @@
+declare module '*?url' {
+	const value: string;
+	export default value;
+}

vite.config.worker.ts

@@ -0,0 +1,23 @@
+import { defineConfig } from 'vite';
+
+export default defineConfig({
+	build: {
+		outDir: 'dist/worker',
+		emptyOutDir: false,
+		sourcemap: false,
+		minify: false,
+		lib: {
+			entry: 'src_worker/index.ts',
+			name: 'PdfParse',
+			fileName: (format) => `source.${format === 'es' ? 'js' : 'cjs'}`,
+			formats: ['es', 'cjs'],
+		},
+		rollupOptions: {
+			output: {
+				preserveModules: false,
+				compact: false,
+			},
+		},
+	},
+	esbuild: false,
+});