feat: defineUnPDFConfig and custom PDF.js mod

README.md

@@ -10,6 +10,7 @@ This library is also intended as a modern alternative to the unmaintained [`pdf-
 
 - 🏗️ Conditional exports for Browser, Node and worker environments
 - 💬 Extract text from PDFs
+- 🧱 Use custom PDF.js build
 
 ## Installation
 
@@ -38,8 +39,30 @@ const { totalPages, info, metadata, text } = await decodePDFText(
 )
 ```
 
+## Config
+
+```ts
+interface UnPDFConfiguration {
+  /**
+   * By default, UnPDF will use the latest version of PDF.js. If you want to
+   * use an older version or the legacy build, set a promise that resolves to
+   * the PDF.js module.
+   *
+   * @example
+   * () => import('pdfjs-dist/legacy/build/pdf.js')
+   */
+  pdfjs?: () => typeof PDFJS
+}
+```
+
 ## Methods
 
+### `defineUnPDFConfig`
+
+```ts
+function defineUnPDFConfig({ pdfjs }: UnPDFConfiguration): Promise<void>
+```
+
 ### `decodePDFText`
 
 ```ts

src/index.node.ts

@@ -2,6 +2,8 @@ import { decodePDFText as _decodePDFText } from './text'
 import { getImagesFromPage as _getImagesFromPage } from './image'
 import { resolvePDFJSImports } from './utils'
 
+export { defineUnPDFConfig } from './utils'
+
 export const decodePDFText: typeof _decodePDFText = async (...args) => {
   await resolvePDFJSImports()
   return await _decodePDFText(...args)

src/index.web.ts

@@ -2,6 +2,8 @@ import { decodePDFText as _decodePDFText } from './text'
 import { getImagesFromPage as _getImagesFromPage } from './image'
 import { resolvePDFJSImports } from './utils'
 
+export { defineUnPDFConfig } from './utils'
+
 export const decodePDFText: typeof _decodePDFText = async (...args) => {
   await resolvePDFJSImports()
   return await _decodePDFText(...args)

src/index.worker.ts

@@ -1,6 +1,8 @@
 import type { decodePDFText as _decodePDFText } from './text'
 import type { getImagesFromPage as _getImagesFromPage } from './image'
 
+export { defineUnPDFConfig } from './utils'
+
 export const decodePDFText: typeof _decodePDFText = async () => {
   throw new Error('Not implemented in worker context yet')
 }

src/types.ts

@@ -1,10 +1,20 @@
+import type * as PDFJS from 'pdfjs-dist'
+
 export interface PDFContent {
   totalPages: number
   info?: Record<string, any>
   metadata?: any
   text: string | string[]
 }
 
-export type Prettify<T> = {
-  [K in keyof T]: T[K];
-} & Record<never, never>
+export interface UnPDFConfiguration {
+  /**
+   * By default, UnPDF will use the latest version of PDF.js. If you want to
+   * use an older version or the legacy build, set a promise that resolves to
+   * the PDF.js module.
+   *
+   * @example
+   * () => import('pdfjs-dist/legacy/build/pdf.js')
+   */
+  pdfjs?: () => Promise<typeof PDFJS>
+}

src/utils.ts

@@ -1,4 +1,5 @@
 import type * as PDFJS from 'pdfjs-dist'
+import type { UnPDFConfiguration } from './types'
 
 let resolvedModule: typeof PDFJS | undefined
 
@@ -14,6 +15,17 @@ export async function getDocumentProxy(data: ArrayBuffer) {
   return pdf
 }
 
+export async function defineUnPDFConfig({ pdfjs }: UnPDFConfiguration) {
+  if (pdfjs) {
+    try {
+      resolvedModule = await pdfjs()
+    }
+    catch (error) {
+      throw new Error('Resolving the PDF.js module failed. Please check your configuration.')
+    }
+  }
+}
+
 export function getResolvedPDFJS() {
   return resolvedModule!
 }
@@ -27,8 +39,6 @@ export async function resolvePDFJSImports() {
     resolvedModule = mod
   }
   catch (error) {
-    throw new Error(
-      'PDF.js is not available. Please add the package as a dependency.',
-    )
+    throw new Error('PDF.js is not available. Please add the package as a dependency.')
   }
 }