TextSearch

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
12
13  exports.runTextSearchTest = () => {
14    const PDFNet = exports.Core.PDFNet;
15
16    const main = async () => {
17      // Relative path to the folder containing test files.
18      const inputURL = '../TestFiles/';
19      const inputFilename = 'credit card numbers.pdf'; // addimage.pdf, newsletter.pdf
20
21      try {
22        const doc = await PDFNet.PDFDoc.createFromURL(inputURL + inputFilename);
23        doc.initSecurityHandler();
24        doc.lock();
25
26        const txtSearch = await PDFNet.TextSearch.create();
27        let mode = PDFNet.TextSearch.Mode.e_whole_word + PDFNet.TextSearch.Mode.e_page_stop; // Uses both whole word and page stop
28        let pattern = 'joHn sMiTh';
29
30        txtSearch.begin(doc, pattern, mode); // searches for the "pattern" in the document while following the inputted modes.
31
32        let step = 0;
33
34        // call Run() iteratively to find all matching instances of the word 'joHn sMiTh'
35        /* eslint-disable-next-line no-constant-condition */
36        while (true) {
37          const result = await txtSearch.run();
38          let hlts;
39          if (result.code === PDFNet.TextSearch.ResultCode.e_found) {
40            if (step === 0) {
41              // Step 0: found "John Smith"
42              // note that, here, 'ambient_str' and 'highlights' are not written to,
43              // as 'e_ambient_string' and 'e_highlight' are not set.
44              console.log(result.out_str + "'s credit card number is: ");
45
46              // now switch to using regular expressions to find John's credit card number
47              mode = await txtSearch.getMode();
48              mode += PDFNet.TextSearch.Mode.e_reg_expression + PDFNet.TextSearch.Mode.e_highlight;
49              txtSearch.setMode(mode);
50              pattern = '\\d{4}-\\d{4}-\\d{4}-\\d{4}'; // or "(\\d{4}-){3}\\d{4}"
51              txtSearch.setPattern(pattern);
52
53              ++step;
54            } else if (step === 1) {
55              // step 1: found John's credit card number
56              console.log('  ' + result.out_str);
57              // note that, here, 'hlts' is written to, as 'e_highlight' has been set.
58              // output the highlight info of the credit card number.
59              hlts = result.highlights;
60              hlts.begin(doc);
61              while (await hlts.hasNext()) {
62                const highlightPageNum = await hlts.getCurrentPageNumber();
63                console.log('The current highlight is from page: ' + highlightPageNum);
64                await hlts.next();
65              }
66              // see if there is an AMEX card number
67              pattern = '\\d{4}-\\d{6}-\\d{5}';
68              txtSearch.setPattern(pattern);
69
70              ++step;
71            } else if (step === 2) {
72              // found an AMEX card number
73              console.log('\nThere is an AMEX card number:\n  ' + result.out_str);
74
75              // change mode to find the owner of the credit card; supposedly, the owner's
76              // name proceeds the number
77              mode = await txtSearch.getMode();
78              mode += PDFNet.TextSearch.Mode.e_search_up;
79              txtSearch.setMode(mode);
80              pattern = '[A-z]++ [A-z]++';
81              txtSearch.setPattern(pattern);
82
83              ++step;
84            } else if (step === 3) {
85              // found the owner's name of the AMEX card
86              console.log("Is the owner's name:\n  " + result.out_str + '?');
87
88              // add a link annotation based on the location of the found instance
89              hlts = result.highlights;
90              await hlts.begin(doc); // is await needed?
91              while (await hlts.hasNext()) {
92                const curPage = await doc.getPage(await hlts.getCurrentPageNumber());
93                const quadArr = await hlts.getCurrentQuads();
94                for (let i = 0; i < quadArr.length; ++i) {
95                  const currQuad = quadArr[i];
96                  const x1 = Math.min(Math.min(Math.min(currQuad.p1x, currQuad.p2x), currQuad.p3x), currQuad.p4x);
97                  const x2 = Math.max(Math.max(Math.max(currQuad.p1x, currQuad.p2x), currQuad.p3x), currQuad.p4x);
98                  const y1 = Math.min(Math.min(Math.min(currQuad.p1y, currQuad.p2y), currQuad.p3y), currQuad.p4y);
99                  const y2 = Math.max(Math.max(Math.max(currQuad.p1y, currQuad.p2y), currQuad.p3y), currQuad.p4y);
100
101                  const hyperLink = await PDFNet.LinkAnnot.create(doc, await PDFNet.Rect.init(x1, y1, x2, y2));
102                  await hyperLink.setAction(await PDFNet.Action.createURI(doc, 'http://www.apryse.com'));
103                  await curPage.annotPushBack(hyperLink);
104                }
105                hlts.next();
106              }
107              const docBuffer = await doc.saveMemoryBuffer(PDFNet.SDFDoc.SaveOptions.e_linearized);
108              saveBufferAsPDFDoc(docBuffer, 'credit card numbers_linked.pdf');
109              break;
110            }
111          } else if (result.code === PDFNet.TextSearch.ResultCode.e_page) {
112            // you can update your UI here, if needed
113            console.log('page end');
114          } else if (result.code === PDFNet.TextSearch.ResultCode.e_done) {
115            break;
116          }
117        }
118      } catch (err) {
119        console.log(err);
120      }
121    };
122    // add your own license key as the second parameter, e.g. PDFNet.runWithCleanup(main, 'YOUR_LICENSE_KEY')
123    PDFNet.runWithCleanup(main);
124  };
125})(window);
126// eslint-disable-next-line spaced-comment
127//# sourceURL=TextSearchTest.js