From a183c26e516ac6802534e7b185305eae2fbfa745 Mon Sep 17 00:00:00 2001
From: Enrico Ros <enrico.ros@gmail.com>
Date: Fri, 17 May 2024 06:33:19 -0700
Subject: [PATCH] PDFUtils: improve PDF to image

---
 src/common/util/pdfUtils.ts | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/src/common/util/pdfUtils.ts b/src/common/util/pdfUtils.ts
index fce030487..a5165f1ac 100644
--- a/src/common/util/pdfUtils.ts
+++ b/src/common/util/pdfUtils.ts
@@ -48,15 +48,22 @@ export async function pdfToText(pdfBuffer: ArrayBuffer): Promise<string> {
 }
 
 
-type PdfPageImage = { base64Url: string, scale: number, width: number, height: number };
+interface PdfPageImage {
+  mimeType: string;
+  base64Data: string;
+  scale: number;
+  width: number;
+  height: number;
+};
 
 /**
  * Renders all pages of a PDF to images
  *
  * @param pdfBuffer The content of a PDF file
+ * @param imageMimeType The MIME type of the image to render (default 'image/jpeg')
  * @param scale The scale factor for the image resolution (default 1.5 for moderate quality)
  */
-export async function pdfToImageDataURLs(pdfBuffer: ArrayBuffer, scale = 1.5): Promise<PdfPageImage[]> {
+export async function pdfToImageDataURLs(pdfBuffer: ArrayBuffer, imageMimeType = 'image/jpeg', scale = 1.5): Promise<PdfPageImage[]> {
   const { getDocument } = await dynamicImportPdfJs();
   const pdf = await getDocument({ data: pdfBuffer }).promise;
   const images: PdfPageImage[] = [];
@@ -74,8 +81,12 @@ export async function pdfToImageDataURLs(pdfBuffer: ArrayBuffer, scale = 1.5): P
       viewport,
     }).promise;
 
+    const base64DataUrl = canvas.toDataURL(imageMimeType, 0.95);
+    const base64Data = base64DataUrl.slice(`data:${imageMimeType};base64,`.length);
+
     images.push({
-      base64Url: canvas.toDataURL('image/jpeg'),
+      mimeType: imageMimeType,
+      base64Data,
       scale,
       width: viewport.width,
       height: viewport.height,