unpdf 0.12.1 → 0.12.2

package/dist/index.cjs

@@ -54,10 +54,10 @@ async function interopDefault(m) {
 
 async function extractImages$1(data, pageNumber) {
   const pdf = isPDFDocumentProxy(data) ? data : await getDocumentProxy(data);
-  const page = await pdf.getPage(pageNumber);
   if (pageNumber < 1 || pageNumber > pdf.numPages) {
     throw new Error(`Invalid page number. Must be between 1 and ${pdf.numPages}.`);
   }
+  const page = await pdf.getPage(pageNumber);
   const operatorList = await page.getOperatorList();
   const { OPS } = await getResolvedPDFJS();
   const images = [];
@@ -68,7 +68,22 @@ async function extractImages$1(data, pageNumber) {
     }
     const imageKey = operatorList.argsArray[i][0];
     const image = await page.objs.get(imageKey);
-    images.push(image.data);
+    if (!image || !image.data || !image.width || !image.height) {
+      continue;
+    }
+    const { width, height, data: data2 } = image;
+    const calculatedChannels = data2.length / (width * height);
+    if (![1, 3, 4].includes(calculatedChannels)) {
+      continue;
+    }
+    const channels = calculatedChannels;
+    images.push({
+      data: data2,
+      width,
+      height,
+      channels,
+      key: imageKey
+    });
   }
   return images;
 }

package/dist/index.d.cts

@@ -3,7 +3,34 @@ import { Canvas } from 'canvas';
 import { DocumentInitParameters, PDFDocumentProxy } from './types/src/display/api';
 import * as _PDFJS from './types/src/pdf';
 
-declare function extractImages$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
+interface ExtractedImageObject {
+    data: Uint8ClampedArray;
+    width: number;
+    height: number;
+    channels: 1 | 3 | 4;
+    key: string;
+}
+/**
+ * Extracts images from a specific page of a PDF document, including necessary metadata,
+ * such as width, height, and calculated color channels.
+ *
+ * This version calculates channels based on image data length, width, and height,
+ * as the `kind` property provided by PDF.js might not reliably indicate the actual
+ * channel count of the raw pixel data (e.g., returning RGBA data even when kind is 3).
+ *
+ * @example
+ * const imagesData = await extractImages(pdf, pageNum)
+ *
+ * for (const imgData of imagesData) {
+ *   const imageIndex = totalImagesProcessed + 1;
+ *   await sharp(imgData.data, {
+ *     raw: { width: imgData.width, height: imgData.height, channels: imgData.channels }
+ *   })
+ *     .png()
+ *     .toFile(`${imageIndex}.png`);
+ * }
+ */
+declare function extractImages$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number): Promise<ExtractedImageObject[]>;
 declare function renderPageAsImage$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number, options?: {
     canvas?: () => Promise<typeof canvas>;
     /** @default 1 */

package/dist/index.d.mts

@@ -3,7 +3,34 @@ import { Canvas } from 'canvas';
 import { DocumentInitParameters, PDFDocumentProxy } from './types/src/display/api';
 import * as _PDFJS from './types/src/pdf';
 
-declare function extractImages$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
+interface ExtractedImageObject {
+    data: Uint8ClampedArray;
+    width: number;
+    height: number;
+    channels: 1 | 3 | 4;
+    key: string;
+}
+/**
+ * Extracts images from a specific page of a PDF document, including necessary metadata,
+ * such as width, height, and calculated color channels.
+ *
+ * This version calculates channels based on image data length, width, and height,
+ * as the `kind` property provided by PDF.js might not reliably indicate the actual
+ * channel count of the raw pixel data (e.g., returning RGBA data even when kind is 3).
+ *
+ * @example
+ * const imagesData = await extractImages(pdf, pageNum)
+ *
+ * for (const imgData of imagesData) {
+ *   const imageIndex = totalImagesProcessed + 1;
+ *   await sharp(imgData.data, {
+ *     raw: { width: imgData.width, height: imgData.height, channels: imgData.channels }
+ *   })
+ *     .png()
+ *     .toFile(`${imageIndex}.png`);
+ * }
+ */
+declare function extractImages$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number): Promise<ExtractedImageObject[]>;
 declare function renderPageAsImage$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number, options?: {
     canvas?: () => Promise<typeof canvas>;
     /** @default 1 */

package/dist/index.d.ts

@@ -3,7 +3,34 @@ import { Canvas } from 'canvas';
 import { DocumentInitParameters, PDFDocumentProxy } from './types/src/display/api';
 import * as _PDFJS from './types/src/pdf';
 
-declare function extractImages$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
+interface ExtractedImageObject {
+    data: Uint8ClampedArray;
+    width: number;
+    height: number;
+    channels: 1 | 3 | 4;
+    key: string;
+}
+/**
+ * Extracts images from a specific page of a PDF document, including necessary metadata,
+ * such as width, height, and calculated color channels.
+ *
+ * This version calculates channels based on image data length, width, and height,
+ * as the `kind` property provided by PDF.js might not reliably indicate the actual
+ * channel count of the raw pixel data (e.g., returning RGBA data even when kind is 3).
+ *
+ * @example
+ * const imagesData = await extractImages(pdf, pageNum)
+ *
+ * for (const imgData of imagesData) {
+ *   const imageIndex = totalImagesProcessed + 1;
+ *   await sharp(imgData.data, {
+ *     raw: { width: imgData.width, height: imgData.height, channels: imgData.channels }
+ *   })
+ *     .png()
+ *     .toFile(`${imageIndex}.png`);
+ * }
+ */
+declare function extractImages$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number): Promise<ExtractedImageObject[]>;
 declare function renderPageAsImage$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, pageNumber: number, options?: {
     canvas?: () => Promise<typeof canvas>;
     /** @default 1 */

package/dist/index.mjs

@@ -52,10 +52,10 @@ async function interopDefault(m) {
 
 async function extractImages$1(data, pageNumber) {
   const pdf = isPDFDocumentProxy(data) ? data : await getDocumentProxy(data);
-  const page = await pdf.getPage(pageNumber);
   if (pageNumber < 1 || pageNumber > pdf.numPages) {
     throw new Error(`Invalid page number. Must be between 1 and ${pdf.numPages}.`);
   }
+  const page = await pdf.getPage(pageNumber);
   const operatorList = await page.getOperatorList();
   const { OPS } = await getResolvedPDFJS();
   const images = [];
@@ -66,7 +66,22 @@ async function extractImages$1(data, pageNumber) {
     }
     const imageKey = operatorList.argsArray[i][0];
     const image = await page.objs.get(imageKey);
-    images.push(image.data);
+    if (!image || !image.data || !image.width || !image.height) {
+      continue;
+    }
+    const { width, height, data: data2 } = image;
+    const calculatedChannels = data2.length / (width * height);
+    if (![1, 3, 4].includes(calculatedChannels)) {
+      continue;
+    }
+    const channels = calculatedChannels;
+    images.push({
+      data: data2,
+      width,
+      height,
+      channels,
+      key: imageKey
+    });
   }
   return images;
 }

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "unpdf",
   "type": "module",
-  "version": "0.12.1",
+  "version": "0.12.2",
   "packageManager": "pnpm@9.12.1",
   "description": "Utilities to work with PDFs in Node.js, browser and workers",
   "author": "Johann Schopplich <hello@johannschopplich.com>",