Package org.apache.tika.mime

Examples of org.apache.tika.mime.MediaType

375
376
377
378
379
380
381
382
383
            log.warn("Encounterd illegal formatted mediaType '{}'  -> will try " +
                "to detect the mediaType based on the parsed content!",
                blob.getMimeType());
            return null;
        } else {
            return new MediaType(mediaTypeArray[0], mediaTypeArray[1],
                blob.getParameter());
        }
    }
View Full Code Here
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
      String encoding = null;
      if(inputType == null) {
        Metadata metadata = new Metadata();
        metadata.set(Metadata.RESOURCE_NAME_KEY, sourceName);
        DefaultDetector detector = new DefaultDetector();
        MediaType type = detector.detect(is, metadata);
   
        ContentHandler contenthandler = new BodyContentHandler();
        AutoDetectParser parser = new AutoDetectParser();
        try {
          parser.parse(is, contenthandler, metadata);
        } catch (SAXException | TikaException e) {
          logger.error("Unable to parse stream: " + e.getMessage());
          throw new KarmaException("Unable to parse stream: "
              + e.getMessage());
        }
        MediaTypeRegistry registry = MimeTypes.getDefaultMimeTypes()
            .getMediaTypeRegistry();
        registry.addSuperType(new MediaType("text", "csv"), new MediaType(
            "text", "plain"));
        MediaType parsedType = MediaType.parse(metadata
            .get(Metadata.CONTENT_TYPE));
   
        if (registry.isSpecializationOf(registry.normalize(type), registry
            .normalize(parsedType).getBaseType())) {
          metadata.set(Metadata.CONTENT_TYPE, type.toString());
View Full Code Here
626
627
628
629
630
631
632
633
634
635
636
        for (MediaType type : registry.getTypes()) {
            System.out.println(type);
            for (MediaType alias : registry.getAliases(type)) {
                System.out.println("  alias:     " + alias);
            }
            MediaType supertype = registry.getSupertype(type);
            if (supertype != null) {
                System.out.println("  supertype: " + supertype);
            }
            Parser p = parsers.get(type);
            if (p != null) {
View Full Code Here
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
            if (name == null) {
                name = "file" + count++;
            }

            MediaType contentType = detector.detect(inputStream, metadata);

            if (name.indexOf('.')==-1 && contentType!=null) {
                try {
                    name += config.getMimeRepository().forName(
                            contentType.toString()).getExtension();
                } catch (MimeTypeException e) {
                    e.printStackTrace();
                }
            }
View Full Code Here
67
68
69
70
71
72
73
74
75
76
77
        TikaInputStream tis = TikaInputStream.cast(input);
        if (tis != null) {
            try {
                ZipFile zip = new ZipFile(tis.getFile());
                try {
                    MediaType type = detectOpenDocument(zip);
                    if (type == null) {
                        type = detectOfficeOpenXML(zip, tis);
                    }
                    if (type == null) {
                        type = detectIWork(zip);
View Full Code Here
91
92
93
94
95
96
97
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
            // It's OOXML
            TikaInputStream stream = TikaInputStream.get(
                    new DocumentInputStream((DocumentEntry) ooxml));
            try {
                ZipContainerDetector detector = new ZipContainerDetector();
                MediaType type = detector.detect(stream, new Metadata());
                handleEmbeddedResource(stream, null, type.toString(), xhtml, true);
                return;
            } finally {
                stream.close();
            }
        } catch(FileNotFoundException e) {
            // It's regular OLE2
        }

       // Need to dump the directory out to a new temp file, so
       //  it's stand along
       POIFSFileSystem newFS = new POIFSFileSystem();
       copy(dir, newFS.getRoot());

       File tmpFile = File.createTempFile("tika", ".ole2");
       try {
           FileOutputStream out = new FileOutputStream(tmpFile);
           newFS.writeFilesystem(out);
           out.close();

           // What kind of document is it?
           Metadata metadata = new Metadata();
           POIFSDocumentType type = POIFSDocumentType.detectType(dir);

           TikaInputStream embedded;

           if (type==POIFSDocumentType.OLE10_NATIVE) {
               Entry entry = dir.getEntry(Ole10Native.OLE10_NATIVE);
               ByteArrayOutputStream bos = new ByteArrayOutputStream();
               IOUtils.copy(new DocumentInputStream((DocumentEntry) entry), bos);
               byte[] data = bos.toByteArray();

               try {
                    Ole10Native ole = new Ole10Native(data, 0);
                    byte[] dataBuffer = ole.getDataBuffer();

                    metadata.set("resourceName", dir.getName() + '/' + ole.getLabel());

                    embedded = TikaInputStream.get(dataBuffer);
               } catch (Ole10NativeException ex) {
                 embedded = TikaInputStream.get(data);
               }
           } else {
               metadata.set(Metadata.CONTENT_TYPE, type.getType().toString());
               metadata.set(Metadata.RESOURCE_NAME_KEY, dir.getName() + '.' + type.getExtension());

               embedded = TikaInputStream.get(tmpFile);
           }

           try {
View Full Code Here
111
112
113
114
115
116
117
118
119
120
121
122
123
        CharsetDetector detector = new CharsetDetector();
        String incomingCharset = metadata.get(Metadata.CONTENT_ENCODING);
        String incomingType = metadata.get(Metadata.CONTENT_TYPE);
        if (incomingCharset == null && incomingType != null) {
            // TIKA-341: Use charset in content-type
            MediaType mt = MediaType.parse(incomingType);
            if (mt != null) {
                String charset = mt.getParameters().get("charset");
                if ((charset != null) && Charset.isSupported(charset)) {
                    incomingCharset = charset;
                }
            }
        }
View Full Code Here
79
80
81
82
83
84
85
86
87
88
89
90
91
        CharsetDetector detector = new CharsetDetector();
        String incomingCharset = metadata.get(Metadata.CONTENT_ENCODING);
        String incomingType = metadata.get(Metadata.CONTENT_TYPE);
        if (incomingCharset == null && incomingType != null) {
            // TIKA-341: Use charset in content-type
            MediaType mt = MediaType.parse(incomingType);
            if (mt != null) {
                incomingCharset = mt.getParameters().get("charset");
            }
        }

        if (incomingCharset != null) {
            detector.setDeclaredEncoding(incomingCharset);
View Full Code Here
117
118
119
120
121
122
123
124
125
126
127
128
        TemporaryResources tmp = new TemporaryResources();
        try {
            TikaInputStream tis = TikaInputStream.get(stream, tmp);

            // Automatically detect the MIME type of the document
            MediaType type = detector.detect(tis, metadata);
            metadata.set(Metadata.CONTENT_TYPE, type.toString());

            // TIKA-216: Zip bomb prevention
            SecureContentHandler sch = new SecureContentHandler(handler, tis);
            try {
                // Parse the document
View Full Code Here
333
334
335
336
337
338
339
340
341
342
343
                NodeList mimes = node.getElementsByTagName("mime");
                if (mimes.getLength() > 0) {
                    Set<MediaType> types = new HashSet<MediaType>();
                    for (int j = 0; j < mimes.getLength(); j++) {
                        String mime = getText(mimes.item(j));
                        MediaType type = MediaType.parse(mime);
                        if (type != null) {
                            types.add(type);
                        } else {
                            throw new TikaException(
                                    "Invalid media type name: " + mime);
View Full Code Here
TOP

Related Classes of org.apache.tika.mime.MediaType

  • bixo.utils.HttpUtils
  • com.cloudera.cdk.morphline.solrcell.SolrCellBuilder$SolrCell
  • com.cloudera.cdk.morphline.tika.DetectMimeTypeBuilder$DetectMimeType
  • com.digitalpebble.behemoth.tika.TikaProcessor
  • controllers.CodeApp
  • dk.i2m.drupal.util.EntityUtils
  • edu.isi.karma.rdf.GenericRDFGenerator
  • io.lumify.tikaMimeType.TikaMimeTypeMapper
  • org.apache.cxf.jaxrs.ext.search.tika.TikaContentExtractor
  • org.apache.jackrabbit.oak.http.AcceptHeader
    • 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.