Package org.apache.tika

Examples of org.apache.tika.Tika

64
65
66
67
68
69
70
71
72
73
74
  /* our log stream */
  private static final Logger LOG = LoggerFactory.getLogger(MimeUtil.class.getName());

  public MimeUtil(Configuration conf) {
    tika = new Tika();
    ObjectCache objectCache = ObjectCache.get(conf);
    MimeTypes mimeTypez = (MimeTypes) objectCache.getObject(MimeTypes.class
        .getName());
    if (mimeTypez == null) {
      try {
View Full Code Here
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
        || (type != null && type.getName().equals(MimeTypes.OCTET_STREAM))) {
      // If no mime-type header, or cannot find a corresponding registered
      // mime-type, then guess a mime-type from the url pattern
      try {
        TikaConfig tikaConfig = TikaConfig.getDefaultConfig();
        Tika tika = new Tika(tikaConfig);
        retType = tika.detect(url) != null ? tika.detect(url) : null;
      } catch (Exception e) {
        String message = "Problem loading default Tika configuration";
        LOG.error(message, e);
        throw new RuntimeException(e);
      }
    } else {
        retType = type.getName();
    }

    // if magic is enabled use mime magic to guess if the mime type returned
    // from the magic guess is different than the one that's already set so far
    // if it is, and it's not the default mime type, then go with the mime type
    // returned by the magic
    if (this.mimeMagic) {
      magicType = tika.detect(data);

      // Deprecated in Tika 1.0 See https://issues.apache.org/jira/browse/NUTCH-1230
      //MimeType magicType = this.mimeTypes.getMimeType(data);
      if (magicType != null && !magicType.equals(MimeTypes.OCTET_STREAM)
          && !magicType.equals(MimeTypes.PLAIN_TEXT)
View Full Code Here
103
104
105
106
107
108
109
110
111
112
113
    this.base = url.toString();
    this.file = file;
    this.conf = conf;
   
    MIME = new MimeUtil(conf);
    tika = new Tika();

    if (!"file".equals(url.getProtocol()))
      throw new FileException("Not a file url:" + url);

    if (File.LOG.isTraceEnabled()) {
View Full Code Here
123
124
125
126
127
128
129
130
131
132
133
              s.close();
              partialDoc.setFullText(text);                                     
            }
            else { // not HTML, send to tika instead
              if (null == _tika) {
                _tika = new Tika();
              }
              Metadata metadata = new Metadata();
              text = _tika.parseToString(urlStream, metadata);
              partialDoc.setFullText(text);         
              TextExtractorTika.addMetadata(partialDoc, metadata);
View Full Code Here
329
330
331
332
333
334
335
        }//TESTED
       
      }//TESTED
    }//(end if has options)
   
    _tika = new Tika(TikaConfig.getDefaultConfig().getDetector(), autoDetectParser);
  }//TESTED (apart from unused number option configuration)
View Full Code Here
1175
1176
1177
1178
1179
1180
1181
1182
        }//TESTED
       
      }//TESTED
    }//(end if has options)
   
    _tika = new Tika(TikaConfig.getDefaultConfig().getDetector(), autoDetectParser);
   
  }//TESTED (apart from unused number option configuration)
View Full Code Here
130
131
132
133
134
135
136
137
138
139
140
141
142
143
     * @param bytes for validation
     * @throws ImageFormatException invalid format image processing error
     */
    public void validateImageFormat(byte[] bytes) throws ImageFormatException {
        Validate.notNull(bytes, "Incoming byte array cannot be null");
        Tika tika = new Tika();
        InputStream input = new ByteArrayInputStream(bytes);
        try {
            String type = tika.detect(input);
            if (!VALID_IMAGE_TYPES.contains(type)) {
                LOGGER.debug("Wrong file extension. May be only {}", VALID_IMAGE_EXTENSIONS);
                throw new ImageFormatException(VALID_IMAGE_EXTENSIONS);
            }
        } catch (IOException e) {
View Full Code Here
148
149
150
151
152
153
154
155
156
157
158
159
160
     * @throws ImageProcessException image conversion problem.
     */
    public BufferedImage convertByteArrayToImage(byte[] bytes) throws ImageProcessException {
        BufferedImage result;
        BufferedInputStream bis = new BufferedInputStream(new ByteArrayInputStream(bytes));
        Tika tika = new Tika();
        try {
            String type = tika.detect(bis);
            if (type.contains(ImageService.ICO_TYPE)) {
                result = ICODecoder.read(bis).get(0);
            } else {
                result = ImageIO.read(bis);
            }
View Full Code Here
103
104
105
106
107
108
109
110
111
112
113
    this.base = url.toString();
    this.file = file;
    this.conf = conf;
   
    MIME = new MimeUtil(conf);
    tika = new Tika();

    if (!"file".equals(url.getProtocol()))
      throw new FileException("Not a file url:" + url);

    if (File.LOG.isTraceEnabled()) {
View Full Code Here
63
64
65
66
67
68
69
70
71
72
73
  /* our log stream */
  private static final Logger LOG = LoggerFactory.getLogger(MimeUtil.class.getName());

  public MimeUtil(Configuration conf) {
    tika = new Tika();
    ObjectCache objectCache = ObjectCache.get(conf);
    MimeTypes mimeTypez = (MimeTypes) objectCache.getObject(MimeTypes.class
        .getName());
    if (mimeTypez == null) {
      try {
View Full Code Here
TOP

Related Classes of org.apache.tika.Tika

  • ch.entwine.weblounge.kernel.site.SiteServlet
  • com.digitalpebble.storm.crawler.bolt.ParserBolt
  • com.gentics.cr.file.ResolvableFileBean
  • com.gitblit.servlet.RawServlet
  • com.ikanow.infinit.e.harvest.extraction.document.file.FileHarvester
  • com.ikanow.infinit.e.harvest.extraction.text.boilerpipe.TextExtractorBoilerpipe
  • com.ikanow.infinit.e.harvest.extraction.text.legacy.TextExtractorTika
  • com.javajing.webjar.FileUploadServlet
  • com.trsst.Command
  • com.trsst.ui.AppServlet
    • 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.