ElementReader

1//---------------------------------------------------------------------------------------
2// Copyright (c) 2001-2023 by Apryse Software Inc. All Rights Reserved.
3// Consult legal.txt regarding legal and license information.
4//---------------------------------------------------------------------------------------
5
6(exports => {
7
8
9
10
11  exports.runElementReaderTest = () => {
12    const PDFNet = exports.Core.PDFNet;
13
14    const ProcessElements = async reader => {
15      // Read page contents
16      for (let element = await reader.next(); element !== null; element = await reader.next()) {
17        const temp = await element.getType();
18        switch (temp) {
19          case PDFNet.Element.Type.e_path: // Process path data...
20            {
21              const data = await element.getPathData();
22              /* eslint-disable @typescript-eslint/no-unused-vars */
23              const operators = data.operators;
24              const points = data.points;
25              /* eslint-enable @typescript-eslint/no-unused-vars */
26            }
27            break;
28          case PDFNet.Element.Type.e_text: // Process text strings...
29            {
30              const data = await element.getTextString();
31              console.log(data);
32            }
33            break;
34          case PDFNet.Element.Type.e_form: // Process form XObjects
35            reader.formBegin();
36            await ProcessElements(reader);
37            reader.end();
38            break;
39          default:
40        }
41      }
42    };
43
44    const main = async () => {
45      console.log('-------------------------------------------------');
46      console.log('Sample 1 - Extract text data from all pages in the document.');
47      console.log('Opening the input pdf...');
48      const ret = 0;
49
50      // Relative path to the folder containing test files.
51      const inputUrl = '../TestFiles/';
52
53      const doc = await PDFNet.PDFDoc.createFromURL(inputUrl + 'newsletter.pdf'); // await if there is ret that we care about.
54      doc.initSecurityHandler();
55      doc.lock();
56
57      // eslint-disable-next-line @typescript-eslint/no-unused-vars
58      const pgnum = await doc.getPageCount();
59      const pageReader = await PDFNet.ElementReader.create();
60      const itr = await doc.getPageIterator(1);
61
62      // Read every page
63      for (itr; await itr.hasNext(); itr.next()) {
64        const curritr = await itr.current();
65        pageReader.beginOnPage(curritr);
66        await ProcessElements(pageReader);
67        pageReader.end();
68      }
69
70      console.log('Done.');
71      return ret;
72    };
73
74    // add your own license key as the second parameter, e.g. PDFNet.runWithCleanup(main, 'YOUR_LICENSE_KEY')
75    PDFNet.runWithCleanup(main);
76  };
77})(window);
78// eslint-disable-next-line spaced-comment
79//# sourceURL=ElementReaderTest.js