restructured project (shared-operations)

shared-operations/functions/shared/createSubDocument.js

@@ -0,0 +1,16 @@
+export async function createSubDocument(pdfDoc, pagesToExtractArray, PDFLib) {
+    const subDocument = await PDFLib.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/shared/detectEmptyPages.js

@@ -0,0 +1,61 @@
+import { getImagesOnPage } from "./getImagesOnPage.js";
+
+export async function detectEmptyPages(snapshot, whiteThreashold, PDFJS, OpenCV) {
+    const pdfDoc = await PDFJS.getDocument(snapshot).promise;
+
+    const emptyPages = [];
+    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) {
+        const textContent = await page.getTextContent();
+        return textContent.items.length === 0;
+    }
+
+    async function areImagesBlank(page, threshold) {
+        const images = await getImagesOnPage(page, PDFJS);
+        for (const image of images) {
+            if(!isImageBlank(image, threshold))
+                return false;
+        }
+        return true;
+    }
+    
+    function isImageBlank(image, threshold) {
+        const src = new OpenCV.cv.Mat(image.width, image.height, OpenCV.cv.CV_8UC4);
+        src.data.set(image.data);
+        // Convert the image to grayscale
+        const gray = new OpenCV.cv.Mat();
+        OpenCV.cv.cvtColor(src, gray, OpenCV.cv.COLOR_RGBA2GRAY);
+    
+        // Calculate the mean value of the grayscale image
+        const meanValue = OpenCV.cv.mean(gray);
+    
+        // Free memory
+        src.delete();
+        gray.delete();
+    
+        // Check if the mean value is below the threshold
+        if (meanValue[0] <= threshold) {
+            return true;
+        } else {
+            return false;
+        }
+    }
+}

shared-operations/functions/shared/getImagesOnPage.js

@@ -0,0 +1,11 @@
+export async function getImagesOnPage(page, PDFJS) {
+    const ops = await page.getOperatorList();
+    const images = [];
+    for (var j=0; j < ops.fnArray.length; j++) {
+        if (ops.fnArray[j] == PDFJS.OPS.paintJpegXObject || ops.fnArray[j] == PDFJS.OPS.paintImageXObject) {
+            const image = page.objs.get(ops.argsArray[j][0]);
+            images.push(image);
+        }
+    }
+    return images;
+}