Package net.sf.regain

Examples of net.sf.regain.RegainException

63
64
65
66
67
68
69
70
71
    mUrlRegexAsString = regex;

    try {
      mUrlRegex = Pattern.compile(regex);
    } catch (Exception ex) {
      throw new RegainException("Regular expression of URL matcher has a " +
          "wrong syntax: '" + regex + "'", ex);
    }
  }
View Full Code Here
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
      int exitCode;
      try {
        exitCode = proc.waitFor();
      } catch (InterruptedException exc) {
        throw new RegainException("Waiting for termination of process failed: "
          + commandArr[0], exc);
      }

      if (mLog.isDebugEnabled()) {
        double duration = (double) (System.currentTimeMillis() - startTime) / 1000.0;

        NumberFormat format = NumberFormat.getInstance();
        format.setMinimumFractionDigits(2);
        format.setMaximumFractionDigits(2);

        mLog.debug("..." + toCommand(commandArr) + " finished ("
            + format.format(duration) + " secs)");
      }

      if (exitCode != 0) {
        throw new RegainException("Native command exited with exit code "
            + exitCode + ": '" + toCommand(commandArr) + "'");
      }

      String[] asArr = new String[list.size()];
      list.toArray(asArr);
      return asArr;
    }
    catch (IOException exc) {
      throw new RegainException("Executing native command failed: '"
          + toCommand(commandArr) + "'", exc);
    }
    finally {
      if (in != null) {
        try { in.close(); } catch (IOException exc) {}
View Full Code Here
289
290
291
292
293
294
295
296
297
298
299
    }
    catch (RedirectException exc) {
      throw exc;
    }
    catch (IOException exc) {
      throw new RegainException("Could not load Document with HTTP", exc);
    }
    finally {
      if (in != null) {
        try { in.close(); } catch (Exception exc) {}
      }
View Full Code Here
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
   * @return byte[] The content of file
   * @throws RegainException in case of problems while loading
   */
  public static byte[] loadFile(File file) throws RegainException {
    if (file.isDirectory()) {
      throw new RegainException("Can't load a directory: "
        + file.getAbsolutePath());
    }

    FileInputStream in = null;
    ByteArrayOutputStream out = null;
    try {
      in = new FileInputStream(file);
      out = new ByteArrayOutputStream((int) file.length());

      RegainToolkit.pipe(in, out);

      return out.toByteArray();
    }
    catch (IOException exc) {
      throw new RegainException("Loading file failed " + file.getAbsolutePath(), exc);
    }
    finally {
      if (out != null) {
        try { out.close(); } catch (IOException exc) {}
      }
View Full Code Here
356
357
358
359
360
361
362
363
364
365
366
      RegainToolkit.pipe(inputStream, out);

      return out.toByteArray();
    }
    catch (IOException exc) {
      throw new RegainException("Loading inputstream failed ", exc);
    }
    finally {
      if (out != null) {
        try { out.close(); } catch (IOException exc) {}
      }
View Full Code Here
738
739
740
741
742
743
744
745
      } catch (IllegalStateException ex) {
        // No match found
        return "";
      }
    } else {
      throw new RegainException("URL is unparsable. url: " + completeUrl);
    }
  }
View Full Code Here
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
      } else {
        setCleanedContent("");
      }

    } catch (TagException ex) {
      throw new RegainException("Error handling audio file: " + rawDocument.getUrl(), ex);

    } catch (ReadOnlyFileException ex) {
      throw new RegainException("Error handling audio file: " + rawDocument.getUrl(), ex);

    } catch (InvalidAudioFrameException ex) {
      throw new RegainException("Error handling audio file: " + rawDocument.getUrl(), ex);

    } catch (IOException ex) {
      throw new RegainException("Error handling audio file: " + rawDocument.getUrl(), ex);
    }
  }
View Full Code Here
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
        StandardDecryptionMaterial sdm = new StandardDecryptionMaterial("");
        pdfDocument.openProtection(sdm);
        AccessPermission ap = pdfDocument.getCurrentAccessPermission();

        if (!ap.canExtractContent()) {
          throw new RegainException("Document is encrypted and can't be opened: " + url);
        }
      }

      // Extract the text with a utility class
      PDFTextStripper stripper = new PDFTextStripper();
      stripper.setSuppressDuplicateOverlappingText(false);
      stripper.setSortByPosition(true);
      stripper.setStartPage(1);
      stripper.setEndPage(Integer.MAX_VALUE);

      setCleanedContent(stripper.getText(pdfDocument).replaceAll("visiblespace", " "));

      // extract annotations
      StringBuilder annotsResult = new StringBuilder();
      List allPages = pdfDocument.getDocumentCatalog().getAllPages();
      for (int i = 0; i < allPages.size(); i++) {
        int pageNum = i + 1;
        PDPage page = (PDPage) allPages.get(i);
        List<PDAnnotation> annotations = page.getAnnotations();
        if (annotations.size() < 1) {
          continue;
        }
        mLog.debug("Total annotations = " + annotations.size());
        mLog.debug("\nProcess Page " + pageNum + "...");
        for (PDAnnotation annotation : annotations) {
          if (annotation.getContents() != null && annotation.getContents().length() > 0) {
            annotsResult.append(annotation.getContents());
            annotsResult.append(" ");
            mLog.debug("Text from annotation: " + annotation.getContents());
          }
        }
      }
      if (annotsResult.length() > 0) {
        setCleanedContent(getCleanedContent() + " Annotations " + annotsResult.toString());
      }

      // Get the meta data
      PDDocumentInformation info = pdfDocument.getDocumentInformation();
      StringBuilder metaData = new StringBuilder();
      metaData.append("p.");
      metaData.append(Integer.toString(pdfDocument.getNumberOfPages()));
      metaData.append(" ");

      // Check if fields are null
      if (info.getAuthor() != null) {
        metaData.append(info.getAuthor());
        metaData.append(" ");
      }
      if (info.getSubject() != null) {
        metaData.append(info.getSubject());
        metaData.append(" ");
      }
      if (info.getKeywords() != null) {
        metaData.append(info.getKeywords());
        metaData.append(" ");
      }

      if (info.getTitle() != null) {
        setTitle(info.getTitle());
      }

      setCleanedMetaData(metaData.toString());
      if (mLog.isDebugEnabled()) {
        mLog.debug("Extracted meta data ::" + getCleanedMetaData()
                + ":: from " + rawDocument.getUrl());
      }

    } catch (CryptographyException exc) {
      throw new RegainException("Error decrypting document: " + url, exc);

    } catch (BadSecurityHandlerException exc) {
      // They didn't supply a password and the default of "" was wrong.
      throw new RegainException("Document is encrypted: " + url, exc);

    } catch (IOException exc) {
      throw new RegainException("Error reading document: " + url, exc);

    } finally {
      if (stream != null) {
        try {
          stream.close();
View Full Code Here
223
224
225
226
227
228
229
230
      // Dokument schlie�en (ohne Speichern)
      doc.close(new Variant(false));
    }
    catch (ComFailException exc) {
      throw new RegainException("Using COM failed.", exc);
    }
  }
View Full Code Here
302
303
304
305
306
307
308
309
310
311
312
        // Word schlie�en
        mWordApplication.quit();
        mLog.info("Closed MS Word");
      }
      catch (Throwable thr) {
        throw new RegainException("Using COM failed.", thr);
      }
      finally {
        // Alle Ressourcen des COM-Threads freigeben
        ComThread.Release();
      }
View Full Code Here
TOP

Related Classes of net.sf.regain.RegainException

  • net.sf.regain.crawler.config.RegexUrlMatcher
  • net.sf.regain.crawler.config.XmlCrawlerConfig
  • net.sf.regain.crawler.Crawler
  • net.sf.regain.crawler.CrawlerToolkit
  • net.sf.regain.crawler.document.AbstractPreparator
  • net.sf.regain.crawler.document.DocumentFactory
  • net.sf.regain.crawler.document.PreparatorFactory
  • net.sf.regain.crawler.document.RawDocument
  • net.sf.regain.crawler.IndexWriterManager
  • net.sf.regain.crawler.preparator.AbstractJacobMsOfficePreparator
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.