Package net.sf.regain

Examples of net.sf.regain.RegainException

243
244
245
246
247
248
249
250
251
252
253
        // NOTE: getCanonicalFile is needed to get rid of "\." is a save way
        try {
          mWorkingDir = new File(realpath).getCanonicalFile().getParentFile();
        }
        catch (IOException exc) {
          throw new RegainException("Getting the working directory of " +
                "the web server failed", exc);
        }
      } else {
        // We got no real path -> Return the current directory
        mWorkingDir = new File(".");
View Full Code Here
83
84
85
86
87
88
89
90
  public String getParameter(String name) throws RegainException {
    try {
      return mRequest.getParameter(name);
    }
    catch (IOException exc) {
      throw new RegainException("Getting parameter " + name + " failed", exc);
    }
  }
View Full Code Here
102
103
104
105
106
107
108
109
  public String[] getParameters(String name) throws RegainException {
    try {
      return mRequest.getParameters().getParameters(name);
    }
    catch (IOException exc) {
      throw new RegainException("Getting parameter " + name + " failed", exc);
    }
  }
View Full Code Here
119
120
121
122
123
124
125
126
    try {
      Parameters params = mRequest.getParameters();
      return params.getParameterNames();
    }
    catch (IOException exc) {
      throw new RegainException("Getting parameter names failed", exc);
    }
  }
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
172
173
174
175
176
177
178
179
180
181
182
183
184
    try {
      PreparatorSettings[] prepConf = config.getPreparatorSettingsList();
      mPreparatorArr = PreparatorFactory.getInstance().createPreparatorArr(prepConf);
    }
    catch (RegainException exc) {
      throw new RegainException("Creating the document preparators failed", exc);
    }

    // Create a profiler for each preparator
    mPreparatorProfilerArr = new Profiler[mPreparatorArr.length];
    for (int i = 0; i < mPreparatorProfilerArr.length; i++) {
      String name = mPreparatorArr[i].getClass().getName();
      mPreparatorProfilerArr[i] = new Profiler("Preparator " + name, "docs");
    }

    // Create the CrawlerAccessController
    String accessClass = config.getCrawlerAccessControllerClass();
    if (accessClass != null) {
      String accessJar = config.getCrawlerAccessControllerJar();
      mCrawlerAccessController = (CrawlerAccessController)
        RegainToolkit.createClassInstance(accessClass, CrawlerAccessController.class,
                                          accessJar);
     
      Properties accessControllerConfig = config.getCrawlerAccessControllerConfig();
      if (accessControllerConfig == null) {
        accessControllerConfig = new Properties();
      }
      mCrawlerAccessController.init(accessControllerConfig);
     
      mLog.info("Using crawler access controller: " + accessClass);
    }

    // Create the mUseLinkTextAsTitleReArr
    String[] useLinkTextAsTitleRegexArr = config.getUseLinkTextAsTitleRegexList();
    if (useLinkTextAsTitleRegexArr == null) {
      mUseLinkTextAsTitleReArr = new RE[0];
    } else {
      mUseLinkTextAsTitleReArr = new RE[useLinkTextAsTitleRegexArr.length];
      for (int i = 0; i < useLinkTextAsTitleRegexArr.length; i++) {
        try {
          mUseLinkTextAsTitleReArr[i] = new RE(useLinkTextAsTitleRegexArr[i]);
        }
        catch (RESyntaxException exc) {
          throw new RegainException("Regular expression of "
            + "use-link-text-as-title-pattern #" + i + " has wrong syntax '"
            + useLinkTextAsTitleRegexArr[i] + "'", exc);
        }
      }
    }
View Full Code Here
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
      preparatorProfiler.stopMeasuring(rawDocument.getLength());
    }
    catch (Throwable thr) {
      preparatorProfiler.abortMeasuring();
      throw new RegainException("Preparing " + url
        + " with preparator " + preparator.getClass().getName() + " failed", thr);
    }

    // Check the mandatory information
    if (cleanedContent == null) {
      throw new RegainException("Preparator " + preparator.getClass().getName()
        + " did not extract the content of " + url);
    }

    // Preparing succeed -> Create the document
    Document doc = createDocument(rawDocument, cleanedContent, title,
View Full Code Here
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
        // The port has changed -> Close the old socket
        try {
          mCurrentSocket.close();
        }
        catch (IOException exc) {
          throw new RegainException("Closing the old socket failed", exc);
        }
      }
     
      // Create the simpleweb connection if nessesary
      if (mSimplewebConnection == null) {
        try {
          FileContext context = new FileContext(new File("web"));
          MapperEngine engine = new MapperEngine(context);
         
          engine.load("SharedTagService", SharedTagService.class.getName());
          engine.link("*", "SharedTagService");
         
          ProtocolHandler handler = HandlerFactory.getInstance(engine);
         
          mSimplewebConnection = ConnectionFactory.getConnection(handler);
        }
        catch (Exception exc) {
          throw new RegainException("Creating simpleweb server failed", exc);
        }
      }

      mLog.info("Listening on port " + port + "...");
      try {
        mCurrentSocket = new ServerSocket(port);
      }
      catch (IOException exc) {
        throw new RegainException("Creating socket for port " + port + " failed.", exc);
      }
      mSimplewebConnection.connect(mCurrentSocket);
    }
  }
View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
  public void printEndTag(PageRequest request, PageResponse response)
    throws RegainException
  {
    Document hit = (Document) request.getContextAttribute(ATTR_CURRENT_HIT);
    if (hit == null) {
      throw new RegainException("Tag " + getTagName()
          + " must be inside a list tag!");
    }

    Integer hitIndex = (Integer) request.getContextAttribute(ATTR_CURRENT_HIT_INDEX);
View Full Code Here
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
      if (indexNameArr == null) {
        // There was no index specified -> Check whether we have default indexes
        // defined
        indexNameArr = mConfig.getDefaultIndexNameArr();
        if (indexNameArr == null) {
          throw new RegainException("Request parameter 'index' not specified and " +
              "no default index configured");
        }
      }
     
      // Get the configurations for these indexes
      configArr = new IndexConfig[indexNameArr.length];
      for (int i = 0; i < indexNameArr.length; i++) {
        configArr[i] = mConfig.getIndexConfig(indexNameArr[i]);
        if (configArr[i] == null) {
          throw new RegainException("The configuration does not contain the index '"
              + indexNameArr[i] + "'");
        }
      }

      // Store the IndexConfig in the page context
View Full Code Here
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
        // There was no index specified -> Check whether we have default
        // indexes
        // defined
        indexNameArr = mConfig.getDefaultIndexNameArr();
        if (indexNameArr == null) {
          throw new RegainException("Request parameter 'index' not specified and "
              + "no default index configured");
        }
      }

      // Get the configurations for these indexes
      List configList = new ArrayList();
      for (int i = 0; i < indexNameArr.length; i++) {
        IndexConfig index = mConfig.getIndexConfig(indexNameArr[i]);
        if (index == null) {
          throw new RegainException("The configuration does not contain the index '" + indexNameArr[i] + "'");

        }
        // If index is a parent index -> get all childs
        if (index.isParent()) {
          String[] allIndexName = mConfig.getAllIndexNameArr();
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.