Cleaned the server-node dir

Split express into different ports for easier migration
Typed page ops that had warnings

shared-operations/functions/common/createSubDocument.ts

@@ -0,0 +1,19 @@
+
+import { PDFDocument } from 'pdf-lib';
+
+export async function createSubDocument(pdfDoc: PDFDocument, pagesToExtractArray: number[]): Promise<Uint8Array> {
+    const subDocument = await PDFDocument.create();
+
+    // Check that array max number is not larger pdf pages number
+    if(Math.max(...pagesToExtractArray) >= pdfDoc.getPageCount()) {
+        throw new Error(`The PDF document only has ${pdfDoc.getPageCount()} pages and you tried to extract page ${Math.max(...pagesToExtractArray)}`);
+    }
+
+    const copiedPages = await subDocument.copyPages(pdfDoc, pagesToExtractArray);
+
+    for (let i = 0; i < copiedPages.length; i++) {
+        subDocument.addPage(copiedPages[i]);
+    }
+
+    return subDocument.save();
+}

shared-operations/functions/common/detectEmptyPages.ts

@@ -0,0 +1,50 @@
+import { DocumentInitParameters, PDFPageProxy } from "pdfjs-dist/types/src/display/api.js";
+import PDFJS from 'pdfjs-dist';
+import { Image } from 'image-js';
+
+import { getImagesOnPage } from "./getImagesOnPage.js";
+
+export async function detectEmptyPages(snapshot: string | URL | ArrayBuffer | DocumentInitParameters, whiteThreashold: number) {
+    const pdfDoc = await PDFJS.getDocument(snapshot).promise;
+
+    const emptyPages: number[] = [];
+    for (let i = 1; i <= pdfDoc.numPages; i++) {
+        const page = await pdfDoc.getPage(i);
+        console.log("Checking page " + i);
+
+        if(!await hasText(page)) {
+            console.log(`Found text on Page ${i}, page is not empty`);
+            continue;
+        }
+
+        if(!await areImagesBlank(page, whiteThreashold)) {
+            console.log(`Found non white image on Page ${i}, page is not empty`);
+            continue;
+        }
+
+        console.log(`Page ${i} is empty.`);
+        emptyPages.push(i - 1);
+    }
+    return emptyPages;
+}
+
+async function hasText(page: PDFPageProxy): Promise<boolean> {
+    const textContent = await page.getTextContent();
+    return textContent.items.length === 0;
+}
+
+async function areImagesBlank(page: PDFPageProxy, threshold: number): Promise<boolean> {
+    const images = await getImagesOnPage(page);
+    for (const image of images) {
+        if(!await isImageBlank(image, threshold))
+            return false;
+    }
+    return true;
+}
+
+async function isImageBlank(image: string | Uint8Array | ArrayBuffer, threshold: number): Promise<boolean> {
+    var img = await Image.load(image);
+    var grey = img.grey();
+    var mean = grey.getMean();
+    return mean[0] <= threshold;
+}

shared-operations/functions/common/getImagesOnPage.ts

@@ -0,0 +1,15 @@
+
+import { PDFPageProxy } from "pdfjs-dist/types/src/display/api.js";
+import PDFJS from 'pdfjs-dist';
+
+export async function getImagesOnPage(page: PDFPageProxy) {
+    const ops = await page.getOperatorList();
+    const images: any = [];
+    for (var j=0; j < ops.fnArray.length; j++) {
+        if (ops.fnArray[j] == PDFJS.OPS.paintImageXObject) {
+            const image = page.objs.get(ops.argsArray[j][0]);
+            images.push(image);
+        }
+    }
+    return images;
+}