3.0.0

package.json

@@ -1,33 +1,35 @@
 {
   "name": "@pushrocks/smartpdf",
-  "version": "2.0.7",
+  "version": "3.0.0",
   "private": false,
   "description": "create pdfs on the fly",
   "main": "dist_ts/index.js",
   "typings": "dist_ts/index.d.ts",
+  "type": "module",
   "author": "Lossless GmbH",
   "license": "MIT",
   "scripts": {
     "test": "(tstest test/ --web)",
-    "build": "(tsbuild --web)"
+    "build": "(tsbuild --web --allowimplicitany)"
   },
   "devDependencies": {
-    "@gitzone/tsbuild": "^2.1.25",
-    "@gitzone/tsrun": "^1.2.12",
-    "@gitzone/tstest": "^1.0.52",
-    "@pushrocks/tapbundle": "^3.2.14",
-    "@types/node": "^14.14.37",
-    "tslint": "^6.1.3",
-    "tslint-config-prettier": "^1.18.0"
+    "@gitzone/tsbuild": "^2.1.60",
+    "@gitzone/tsrun": "^1.2.31",
+    "@gitzone/tstest": "^1.0.69",
+    "@pushrocks/tapbundle": "^5.0.3",
+    "@types/node": "^17.0.23"
   },
   "dependencies": {
-    "@pushrocks/smartfile": "^8.0.9",
-    "@pushrocks/smartnetwork": "^2.0.4",
-    "@pushrocks/smartpromise": "^3.1.3",
-    "@pushrocks/smartpuppeteer": "^1.0.21",
+    "@pushrocks/smartfile": "^9.0.6",
+    "@pushrocks/smartnetwork": "^2.0.14",
+    "@pushrocks/smartpath": "^5.0.5",
+    "@pushrocks/smartpromise": "^3.1.7",
+    "@pushrocks/smartpuppeteer": "^2.0.0",
     "@pushrocks/smartunique": "^3.0.3",
-    "@types/express": "^4.17.11",
-    "express": "^4.17.1"
+    "@types/express": "^4.17.13",
+    "express": "^4.17.3",
+    "pdf-merger-js": "^3.4.0",
+    "pdf2json": "^2.0.0"
   },
   "files": [
     "ts/**/*",

test/test.ts

@@ -1,11 +1,11 @@
 import { expect, tap } from '@pushrocks/tapbundle';
-import * as smartpdf from '../ts/index';
+import * as smartpdf from '../ts/index.js';
 
 let testSmartPdf: smartpdf.SmartPdf;
 
 tap.test('should create a valid instance of smartpdf', async () => {
   testSmartPdf = new smartpdf.SmartPdf();
-  expect(testSmartPdf).to.be.instanceof(smartpdf.SmartPdf);
+  expect(testSmartPdf).toBeInstanceOf(smartpdf.SmartPdf);
 });
 
 tap.test('should start the instance', async () => {
@@ -13,7 +13,11 @@ tap.test('should start the instance', async () => {
 });
 
 tap.test('should create a pdf from html string', async () => {
-  await testSmartPdf.getPdfResultForHtmlString('hi');
+  await testSmartPdf.getA4PdfResultForHtmlString('hi');
+});
+
+tap.test('should create a pdf from html string', async () => {
+  await testSmartPdf.getA4PdfResultForHtmlString('hi');
 });
 
 tap.test('should create a pdf from website as A4', async () => {
@@ -25,18 +29,28 @@ tap.test('should create a pdf from website as single page PDF', async () => {
 });
 
 tap.test('should create a valid PDFResult', async () => {
-  const pdfResult = await testSmartPdf.getFullWebsiteAsSinglePdf(
-    'https://maintainedby.lossless.com'
-  );
-  expect(pdfResult.buffer).to.be.instanceOf(Buffer);
-  const fs = await import('fs');
+  const writePDfToDisk = async (urlArg: string, fileName: string) => {
+    const pdfResult = await testSmartPdf.getFullWebsiteAsSinglePdf(urlArg);
+    expect(pdfResult.buffer).toBeInstanceOf(Buffer);
+    const fs = await import('fs');
 
-  if (!fs.existsSync('.nogit/')) {
-    fs.mkdirSync('.nogit/');
-  }
-  fs.writeFileSync('.nogit/sample.pdf', pdfResult.buffer);
+    if (!fs.existsSync('.nogit/')) {
+      fs.mkdirSync('.nogit/');
+    }
+    fs.writeFileSync(`.nogit/${fileName}`, pdfResult.buffer);
+  };
+  await writePDfToDisk('https://maintainedby.lossless.com/', '1.pdf')
+  await writePDfToDisk('https://rendertron.lossless.one/render/https://fitnessloft.de/impressum/', '2.pdf')
 });
 
+tap.test('should combine pdfs', async () => {
+  const fs = await import('fs');
+  const buffer1 = fs.readFileSync('.nogit/1.pdf');
+  const buffer2 = fs.readFileSync('.nogit/2.pdf');
+  fs.writeFileSync(`.nogit/combined.pdf`, await testSmartPdf.mergePdfBuffers([buffer1, buffer2]));
+
+})
+
 tap.test('should be able to close properly', async () => {
   await testSmartPdf.stop();
 });

ts/index.ts

@@ -6,4 +6,4 @@ declare global {
 }
 
 // normal
-export * from './smartpdf.classes.smartpdf';
+export * from './smartpdf.classes.smartpdf.js';

ts/interfaces/index.ts

@@ -1 +1 @@
-export * from './interface.pdfresult';
+export * from './interface.pdfresult.js';

ts/interfaces/interface.pdfresult.ts

@@ -1,5 +1,8 @@
 export interface IPdfResult {
   name: string;
   id: string;
+  metadata: {
+    textExtraction: string;
+  };
   buffer: Buffer;
 }

ts/smartpdf.classes.pdfcandidate.ts

@@ -1,8 +1,11 @@
-import * as plugins from './smartpdf.plugins';
+import * as plugins from './smartpdf.plugins.js';
 
 export class PdfCandidate {
-  pdfId = plugins.smartunique.shortId();
-  doneDeferred = plugins.smartpromise.defer();
+  public htmlString: string;
+  public pdfId = plugins.smartunique.shortId();
+  public doneDeferred = plugins.smartpromise.defer();
 
-  constructor(public htmlString) {}
+  constructor(htmlStringArg: string) {
+    this.htmlString = htmlStringArg;
+  }
 }

ts/smartpdf.classes.smartpdf.ts

@@ -1,11 +1,11 @@
-import * as plugins from './smartpdf.plugins';
-import * as paths from './smartpdf.paths';
+import * as plugins from './smartpdf.plugins.js';
+import * as paths from './smartpdf.paths.js';
 import { Server } from 'http';
-import { PdfCandidate } from './smartpdf.classes.pdfcandidate';
+import { PdfCandidate } from './smartpdf.classes.pdfcandidate.js';
 
-declare const document;
+declare const document: any;
 
-import * as interfaces from './interfaces';
+import * as interfaces from './interfaces/index.js';
 
 export class SmartPdf {
   htmlServerInstance: Server;
@@ -19,7 +19,7 @@ export class SmartPdf {
     this._readyDeferred = new plugins.smartpromise.Deferred();
   }
 
-  async start(headlessBrowserArg?) {
+  async start(headlessBrowserArg?: plugins.smartpuppeteer.puppeteer.Browser) {
     // lets set the external browser in case one is provided
     this.headlessBrowser = headlessBrowserArg;
     // setup puppeteer
@@ -63,11 +63,15 @@ export class SmartPdf {
   /**
    * returns a pdf for a given html string;
    */
-  async getPdfResultForHtmlString(htmlStringArg: string): Promise<interfaces.IPdfResult> {
+  async getA4PdfResultForHtmlString(htmlStringArg: string): Promise<interfaces.IPdfResult> {
     await this._readyDeferred.promise;
     const pdfCandidate = new PdfCandidate(htmlStringArg);
     this._candidates[pdfCandidate.pdfId] = pdfCandidate;
     const page = await this.headlessBrowser.newPage();
+    await page.setViewport({
+      width: 794,
+      height: 1122,
+    });
     const response = await page.goto(`http://localhost:3210/${pdfCandidate.pdfId}`, {
       waitUntil: 'networkidle2',
     });
@@ -80,10 +84,10 @@ export class SmartPdf {
     }
 
     const pdfBuffer = await page.pdf({
-      format: 'A4',
+      width: 794,
+      height: 1122,
       printBackground: true,
       displayHeaderFooter: false,
-      preferCSSPageSize: true,
     });
     await page.close();
     delete this._candidates[pdfCandidate.pdfId];
@@ -92,12 +96,19 @@ export class SmartPdf {
     return {
       id: pdfCandidate.pdfId,
       name: `${pdfCandidate.pdfId}.js`,
+      metadata: {
+        textExtraction: await this.extractTextFromPdfBuffer(pdfBuffer),
+      },
       buffer: pdfBuffer,
     };
   }
 
   async getPdfResultForWebsite(websiteUrl: string): Promise<interfaces.IPdfResult> {
     const page = await this.headlessBrowser.newPage();
+    await page.setViewport({
+      width: 1980,
+      height: 1200,
+    });
     await page.emulateMediaType('screen');
     const response = await page.goto(websiteUrl, { waitUntil: 'networkidle2' });
     const pdfId = plugins.smartunique.shortId();
@@ -108,45 +119,84 @@ export class SmartPdf {
       };
     });
     const pdfBuffer = await page.pdf({
-      format: 'A4',
-      height: documentWidth,
+      height: documentHeight,
       width: documentWidth,
       printBackground: true,
       displayHeaderFooter: false,
-      preferCSSPageSize: true,
     });
     await page.close();
     return {
       id: pdfId,
       name: `${pdfId}.js`,
+      metadata: {
+        textExtraction: await this.extractTextFromPdfBuffer(pdfBuffer),
+      },
       buffer: pdfBuffer,
     };
   }
 
-  async getFullWebsiteAsSinglePdf(websiteUrl: string) {
+  async getFullWebsiteAsSinglePdf(websiteUrl: string): Promise<interfaces.IPdfResult> {
     const page = await this.headlessBrowser.newPage();
+    await page.setViewport({
+      width: 1920,
+      height: 1200,
+    });
     page.emulateMediaType('screen');
     const response = await page.goto(websiteUrl, { waitUntil: 'networkidle2' });
     const pdfId = plugins.smartunique.shortId();
     const { documentHeight, documentWidth } = await page.evaluate(() => {
       return {
-        documentHeight: document.height,
-        documentWidth: document.width,
+        documentHeight: document.body.scrollHeight,
+        documentWidth: document.body.clientWidth,
       };
     });
+    await page.setViewport({
+      width: 1920,
+      height: documentHeight,
+    });
     const pdfBuffer = await page.pdf({
-      format: 'A4',
-      height: documentWidth,
-      width: documentWidth,
+      height: documentHeight,
+      width: 1920,
       printBackground: true,
       displayHeaderFooter: false,
-      preferCSSPageSize: true,
+      scale: 1,
+      pageRanges: '1'
     });
     await page.close();
     return {
       id: pdfId,
       name: `${pdfId}.js`,
+      metadata: {
+        textExtraction: await this.extractTextFromPdfBuffer(pdfBuffer),
+      },
       buffer: pdfBuffer,
     };
   }
+
+  public async mergePdfBuffers(pdfBuffers: Buffer[]): Promise<Buffer> {
+    const merger = new plugins.pdfMerger();
+    for (const buffer of pdfBuffers) {
+      merger.add(buffer);
+    }
+    return merger.saveAsBuffer();
+  }
+
+  public async extractTextFromPdfBuffer(pdfBufferArg: Buffer): Promise<string> {
+    const deferred = plugins.smartpromise.defer<string>();
+    const pdfParser: any = new plugins.pdf2json();
+    pdfParser.on('pdfParser_dataReady', (pdfData: any) => {
+      let finalText = ''
+      for (const page of pdfData.Pages) {
+        for(const text of page.Texts) {
+          for (const letter of text.R) {
+            finalText = finalText + letter.T;
+          }
+          
+        };
+      }
+      deferred.resolve(finalText);
+    });
+    pdfParser.parseBuffer(pdfBufferArg);
+    return deferred.promise;
+  }
 }

ts/smartpdf.paths.ts

@@ -1,3 +1,3 @@
-import * as plugins from './smartpdf.plugins';
+import * as plugins from './smartpdf.plugins.js';
 
-export const packageDir = plugins.path.join(__dirname, '../');
+export const packageDir = plugins.path.join(plugins.smartpath.get.dirnameFromImportMetaUrl(import.meta.url), '../');

ts/smartpdf.plugins.ts

@@ -7,13 +7,17 @@ export { http, path };
 // @pushrocks
 import * as smartfile from '@pushrocks/smartfile';
 import * as smartpromise from '@pushrocks/smartpromise';
+import * as smartpath from '@pushrocks/smartpath';
 import * as smartpuppeteer from '@pushrocks/smartpuppeteer';
 import * as smartnetwork from '@pushrocks/smartnetwork';
 import * as smartunique from '@pushrocks/smartunique';
 
-export { smartfile, smartpromise, smartpuppeteer, smartunique, smartnetwork };
+export { smartfile, smartpromise, smartpath, smartpuppeteer, smartunique, smartnetwork };
 
 // thirdparty
+import pdfMerger from 'pdf-merger-js';
+// @ts-ignore
+import pdf2json from 'pdf2json';
 import express from 'express';
 
-export { express };
+export { pdfMerger, pdf2json, express };

tsconfig.json

@@ -0,0 +1,8 @@
+{
+  "compilerOptions": {
+    "experimentalDecorators": true,
+    "target": "ES2022",
+    "module": "ES2022",
+    "moduleResolution": "nodenext"
+  }
+}

tslint.json

@@ -1,17 +0,0 @@
-{
-  "extends": ["tslint:latest", "tslint-config-prettier"],
-  "rules": {
-    "semicolon": [true, "always"],
-    "no-console": false,
-    "ordered-imports": false,
-    "object-literal-sort-keys": false,
-    "member-ordering": {
-      "options":{
-        "order": [
-          "static-method"
-        ]
-      }
-    }
-  },
-  "defaultSeverity": "warning"
-}