Text Search with Regular Expression Sample Code

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  exports.runTextSearchTest = () => {
8    const PDFNet = exports.Core.PDFNet;
9
10    const main = async () => {
11      // Relative path to the folder containing test files.
12      const inputURL = '../TestFiles/';
13      const inputFilename = 'credit card numbers.pdf'; // addimage.pdf, newsletter.pdf
14
15      try {
16        const doc = await PDFNet.PDFDoc.createFromURL(inputURL + inputFilename);
17        doc.initSecurityHandler();
18        doc.lock();
19
20        const txtSearch = await PDFNet.TextSearch.create();
21        let mode = PDFNet.TextSearch.Mode.e_whole_word + PDFNet.TextSearch.Mode.e_page_stop; // Uses both whole word and page stop
22        let pattern = 'joHn sMiTh';
23
24        txtSearch.begin(doc, pattern, mode); // searches for the "pattern" in the document while following the inputted modes.
25
26        let step = 0;
27
28        // call Run() iteratively to find all matching instances of the word 'joHn sMiTh'
29        /* eslint-disable-next-line no-constant-condition */
30        while (true) {
31          const result = await txtSearch.run();
32          let hlts;
33          if (result.code === PDFNet.TextSearch.ResultCode.e_found) {
34            if (step === 0) {
35              // Step 0: found "John Smith"
36              // note that, here, 'ambient_str' and 'highlights' are not written to,
37              // as 'e_ambient_string' and 'e_highlight' are not set.
38              console.log(result.out_str + "'s credit card number is: ");
39
40              // now switch to using regular expressions to find John's credit card number
41              mode = await txtSearch.getMode();
42              mode += PDFNet.TextSearch.Mode.e_reg_expression + PDFNet.TextSearch.Mode.e_highlight;
43              txtSearch.setMode(mode);
44              pattern = '\\d{4}-\\d{4}-\\d{4}-\\d{4}'; // or "(\\d{4}-){3}\\d{4}"
45              txtSearch.setPattern(pattern);
46
47              ++step;
48            } else if (step === 1) {
49              // step 1: found John's credit card number
50              console.log('  ' + result.out_str);
51              // note that, here, 'hlts' is written to, as 'e_highlight' has been set.
52              // output the highlight info of the credit card number.
53              hlts = result.highlights;
54              hlts.begin(doc);
55              while (await hlts.hasNext()) {
56                const highlightPageNum = await hlts.getCurrentPageNumber();
57                console.log('The current highlight is from page: ' + highlightPageNum);
58                await hlts.next();
59              }
60              // see if there is an AMEX card number
61              pattern = '\\d{4}-\\d{6}-\\d{5}';
62              txtSearch.setPattern(pattern);
63
64              ++step;
65            } else if (step === 2) {
66              // found an AMEX card number
67              console.log('\nThere is an AMEX card number:\n  ' + result.out_str);
68
69              // change mode to find the owner of the credit card; supposedly, the owner's
70              // name proceeds the number
71              mode = await txtSearch.getMode();
72              mode += PDFNet.TextSearch.Mode.e_search_up;
73              txtSearch.setMode(mode);
74              pattern = '[A-z]++ [A-z]++';
75              txtSearch.setPattern(pattern);
76
77              ++step;
78            } else if (step === 3) {
79              // found the owner's name of the AMEX card
80              console.log("Is the owner's name:\n  " + result.out_str + '?');
81
82              // add a link annotation based on the location of the found instance
83              hlts = result.highlights;
84              await hlts.begin(doc); 
85              while (await hlts.hasNext()) {
86                const curPage = await doc.getPage(await hlts.getCurrentPageNumber());
87                const quadArr = await hlts.getCurrentQuads();
88                for (let i = 0; i < quadArr.length; ++i) {
89                  const currQuad = quadArr[i];
90                  const x1 = Math.min(Math.min(Math.min(currQuad.p1x, currQuad.p2x), currQuad.p3x), currQuad.p4x);
91                  const x2 = Math.max(Math.max(Math.max(currQuad.p1x, currQuad.p2x), currQuad.p3x), currQuad.p4x);
92                  const y1 = Math.min(Math.min(Math.min(currQuad.p1y, currQuad.p2y), currQuad.p3y), currQuad.p4y);
93                  const y2 = Math.max(Math.max(Math.max(currQuad.p1y, currQuad.p2y), currQuad.p3y), currQuad.p4y);
94
95                  const hyperLink = await PDFNet.LinkAnnot.create(doc, await PDFNet.Rect.init(x1, y1, x2, y2));
96                  await hyperLink.setAction(await PDFNet.Action.createURI(doc, 'http://www.apryse.com'));
97                  await curPage.annotPushBack(hyperLink);
98                }
99                hlts.next();
100              }
101              const docBuffer = await doc.saveMemoryBuffer(PDFNet.SDFDoc.SaveOptions.e_linearized);
102              saveBufferAsPDFDoc(docBuffer, 'credit card numbers_linked.pdf');
103              break;
104            }
105          } else if (result.code === PDFNet.TextSearch.ResultCode.e_page) {
106            // you can update your UI here, if needed
107            console.log('page end');
108          } else if (result.code === PDFNet.TextSearch.ResultCode.e_done) {
109            break;
110          }
111        }
112      } catch (err) {
113        console.log(err);
114      }
115    };
116    // add your own license key as the second parameter, e.g. PDFNet.runWithCleanup(main, 'YOUR_LICENSE_KEY')
117    PDFNet.runWithCleanup(main);
118  };
119})(window);
120// eslint-disable-next-line spaced-comment
121//# sourceURL=TextSearchTest.js