|
1 package org.enhydra.snapper.parsers;
2
3 import java.io.File  ;
4 import java.io.FileInputStream ;
5 import java.io.IOException ;
6 import java.io.InputStream ;
7 import java.io.StringWriter ;
8
9 import org.enhydra.snapper.api.*;
10 import org.pdfbox.pdmodel.PDDocument;
11 import org.pdfbox.pdmodel.PDDocumentInformation;
12 import org.pdfbox.util.PDFTextStripper;
13 import org.pdfbox.pdfparser.PDFParser;
14
15 public class SnapperPDFParser implements org.enhydra.snapper.api.Parser {
16 String parsedText, fileName, title, name;
17
18 public String parse(InputStream is) throws java.io.IOException { return "";}
19
20 public void parse(){
21 parsedText = null;
22 title = null;
23 name = null;
24 PDDocument document = null;
25 PDDocumentInformation info = null;
26 PDFTextStripper stripper = null;
27 FileInputStream in = null;
28 try
29 {
30 in = new FileInputStream (new File (fileName)); PDFParser parser = new PDFParser(in);
32 parser.parse();
33 document = parser.getPDDocument();
34
35 stripper = new PDFTextStripper();
36 String pdfFile = null;
37 String textFile = null;
38
39 if( document.isEncrypted() )
40 {
41 try{
42 ParserManager.logger.debug("Document "+ fileName + " is encrypted!");
43 } catch (Exception ex) {
44 System.out.println("Document "+ fileName + " is encrypted!");
45 }
46 document.decrypt("");
47 }
48
49 info = document.getDocumentInformation();
50
51 title = info.getTitle();
52 StringWriter writer = new StringWriter ();
53 stripper.writeText(document, writer);
54 parsedText = writer.getBuffer().toString();
55
56 document.close();
57
58 in.close();
59 in = null;
60 writer.close();
61 writer = null;
62 stripper = null;
63
64 document = null;
65 parser = null;
66 info = null;
67
68 } catch (Exception ex){
69
70 if (document != null) {
71 try{
72 document.close();
73 }
74 catch (IOException e){}
75 }
76 document = null;
77 stripper = null;
78 info=null;
79 try{
80 ParserManager.logger.debug("***** File could not be parsed: " + fileName);
81 } catch (Exception e) {
82 System.out.println("***** File could not be parsed: " + fileName);
83 }
84 }
85 catch (Throwable ex){
86
87 document = null;
88 stripper = null;
89 info=null;
90 try{
91 ParserManager.logger.debug("***** File could not be parsed: " + fileName);
92 } catch (Exception e) {
93 System.out.println("***** File could not be parsed: " + fileName);
94 }
95 }
96 }
97
98
99 public void setFileName(String fileName){
100 this.fileName = fileName;
101 }
102
103 public String getParsedText() {
104 return parsedText;
105 }
106
107 public String getTitle() {
108 return title;
109 }
110 public String getName() {
111 return title;
112 }
113
114 } |
Popular Tags |
Java API By Example, From Geeks To Geeks.