removeBlankPages. backend only

shared-operations/package.json

@@ -11,6 +11,7 @@
   "dependencies": {
     "@stirling-tools/joi": "github:Stirling-Tools/joi",
     "buffer": "^6.0.3",
+    "canvas": "^2.11.2",
     "i18next-resources-to-backend": "^1.2.0",
     "image-js": "^0.35.5",
     "memfs": "^4.9.2",

shared-operations/public/locales/removeBlankPages/en.json

@@ -0,0 +1,10 @@
+{
+    "friendlyName": "Remove Blank Pages",
+    "description": "Remove pages without .",
+    "values": {
+        "whiteThreashold": {
+            "friendlyName": "Image White Threashold",
+            "description": "0-255 the tolerance for near white images. 0 - Almost white is detected as empty. 255 - No images are detected as empty."
+        }
+    }
+}

shared-operations/src/functions/common/getImagesOnPage.ts

@@ -3,6 +3,8 @@ import { PDFPageProxy } from "pdfjs-dist/types/src/display/api";
 
 import * as PDFJS from "pdfjs-dist";
 
+import { createCanvas } from "canvas";
+
 export interface PDFJSImage {
     width: number;
     height: number;
@@ -16,7 +18,9 @@ export async function getImagesOnPage(page: PDFPageProxy): Promise<PDFJSImage[]>
     const images: PDFJSImage[] = [];
     for (let j=0; j < ops.fnArray.length; j++) {
         if (ops.fnArray[j] == PDFJS.OPS.paintImageXObject) {
-            const image = page.objs.get(ops.argsArray[j][0]) as PDFJSImage;
+            const image: PDFJSImage = page.objs.get(ops.argsArray[j][0])
+
+            console.log("Image: ", image);
             images.push(image);
         }
     }

shared-operations/src/functions/removeBlankPages.ts

@@ -1,23 +1,58 @@
+import { Operator, Progress, oneToOne } from ".";
+
+import Joi from "@stirling-tools/joi";
+import { JoiPDFFileSchema } from "../wrappers/PdfFileJoi";
+
+import i18next from "i18next";
 
 import { PdfFile } from "../wrappers/PdfFile";
 import { detectEmptyPages } from "./common/detectEmptyPages";
 import { getPages } from "./common/getPagesByIndex";
 import { invertSelection } from "./common/pageIndexesUtils";
 
-export interface RemoveBlankPagesParamsType {
-    file: PdfFile;
-    whiteThreashold: number;
+export class RemoveBlankPages extends Operator {
+    static type = "removeBlankPages";
+
+    /**
+     * Validation & Localisation
+     */
+
+    protected static inputSchema = JoiPDFFileSchema.label(i18next.t("inputs.pdffile.name")).description(i18next.t("inputs.pdffile.description"));
+    protected static valueSchema = Joi.object({
+        whiteThreashold: Joi.number().min(0).max(255).required()
+            .label(i18next.t("values.whiteThreashold.friendlyName", { ns: "removeBlankPages" })).description(i18next.t("values.whiteThreashold.description", { ns: "removeBlankPages" }))
+            .example("10").example("0").example("255").required()
+    });
+    protected static outputSchema = JoiPDFFileSchema.label(i18next.t("outputs.pdffile.name")).description(i18next.t("outputs.pdffile.description"));
+
+    static schema = Joi.object({
+        input: RemoveBlankPages.inputSchema,
+        values: RemoveBlankPages.valueSchema.required(),
+        output: RemoveBlankPages.outputSchema
+    }).label(i18next.t("friendlyName", { ns: "removeBlankPages" })).description(i18next.t("description", { ns: "removeBlankPages" }));
+
+
+    /**
+     * Logic
+     */
+
+    /** Detect and remove white pages */
+    async run(input: PdfFile[], progressCallback: (state: Progress) => void): Promise<PdfFile[]> {
+        return oneToOne<PdfFile, PdfFile>(input, async (input, index, max) => {
+            const pdfDoc = await input.pdfLibDocument;
+            const pageCount = pdfDoc.getPageCount();
+
+            progressCallback({ curFileProgress: 0, operationProgress: index/max });
+            const emptyPages = await detectEmptyPages(input, this.actionValues.whiteThreashold);
+            progressCallback({ curFileProgress: 0.6, operationProgress: index/max });
+            const pagesToKeep = invertSelection(emptyPages, pageCount);
+            progressCallback({ curFileProgress: 0.3, operationProgress: index/max });
+
+            const result = await getPages(input, pagesToKeep);
+            progressCallback({ curFileProgress: 1, operationProgress: index/max });
+
+            result.filename += "_removedBlanks";
+            return result;
+        });
+    }
 }
-export async function removeBlankPages(params: RemoveBlankPagesParamsType) {
-    const { file, whiteThreashold } = params;
-    const pdfDoc = await file.pdfLibDocument;
-    const pageCount = pdfDoc.getPageCount();
-
-    const emptyPages = await detectEmptyPages(file, whiteThreashold);
-    console.debug("Empty Pages: ", emptyPages);
-    const pagesToKeep = invertSelection(emptyPages, pageCount);
-
-    const newFile = await getPages(file, pagesToKeep);
-    newFile.filename += "_removedBlanks";
-    return newFile;
-}

shared-operations/tsconfig.json

@@ -6,6 +6,9 @@
         "baseUrl": "./src",                                  /* Specify the base directory to resolve non-relative module names. */
         "paths": {
             "#pdfcpu": ["../../shared-operations/src/wasm/pdfcpu/pdfcpu-wrapper.server"],
-        }
+        },
+        "types": [
+            "vite/client"
+        ],
     }                                     /* Specify a set of entries that re-map imports to additional lookup locations. */ 
 }