Package org.apache.tika.mime

Examples of org.apache.tika.mime.MediaType

104
105
106
107
108
109
110
111
112
113
114
    }
   
    private static MediaType fromString(String type) {
        int splitAt = type.indexOf('/');
        if(splitAt > -1) {
            return new MediaType(
              type.substring(0,splitAt),
              type.substring(splitAt+1)
            );
        }
        return MediaType.APPLICATION_ZIP;
View Full Code Here
76
77
78
79
80
81
82
83
84
85
86
87
                context.get(ServiceLoader.class, LOADER));
        try {
            Charset charset = reader.getCharset();
            String previous = metadata.get(Metadata.CONTENT_TYPE);
            if (previous == null || previous.startsWith("text/html")) {
                MediaType type = new MediaType(MediaType.TEXT_HTML, charset);
                metadata.set(Metadata.CONTENT_TYPE, type.toString());
            }
            // deprecated, see TIKA-431
            metadata.set(Metadata.CONTENT_ENCODING, charset.name());

            // Get the HTML mapper from the parse context
View Full Code Here
74
75
76
77
78
79
80
81
82
83
84
85
86
        String charset = null;
        String head = ASCII.decode(ByteBuffer.wrap(buffer, 0, n)).toString();

        Matcher equiv = HTTP_EQUIV_PATTERN.matcher(head);
        if (equiv.find()) {
            MediaType type = MediaType.parse(equiv.group(1));
            if (type != null) {
                charset = type.getParameters().get("charset");
            }
        }
        if (charset == null) {
            // TIKA-892: HTML5 meta charset tag
            Matcher meta = META_CHARSET_PATTERN.matcher(head);
View Full Code Here
610
611
612
613
614
615
616
617
618
619
620
        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
115
116
117
118
119
120
121
122
123
124
125
     * @param mimetype
     *            the Mimetype
     * @return true if the Metadata object can be converted or false if not
     */
    public static boolean isConverterAvailable(String mimetype) {
        MediaType type = MediaType.parse( mimetype );

        if (type != null) {
            return (getConverterMap().get( type ) != null);
        }

View Full Code Here
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
            throw new IllegalArgumentException( "mimetype must not be null" );
        }

        ITikaToXMPConverter converter = null;

        MediaType type = MediaType.parse( mimetype );

        if (type != null) {
            Class<? extends ITikaToXMPConverter> clazz = getConverterMap().get( type );
            if (clazz != null) {
                try {
                    converter = clazz.newInstance();
                }
                catch (Exception e) {
                    throw new TikaException(
                            "TikaToXMP converter class cannot be instantiated for mimetype: "
                                    + type.toString(), e );
                }
            }
        }

        return converter;
View Full Code Here
113
114
115
116
117
118
119
120
121
122
123
124
125
            ais = factory.createArchiveInputStream(stream);
        } catch (ArchiveException e) {
            throw new TikaException("Unable to unpack document stream", e);
        }

        MediaType type = getMediaType(ais);
        if (!type.equals(MediaType.OCTET_STREAM)) {
            metadata.set(CONTENT_TYPE, type.toString());
        }

        // Use the delegate parser to parse the contained document
        EmbeddedDocumentExtractor extractor = context.get(
                EmbeddedDocumentExtractor.class,
View Full Code Here
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
            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
96
97
98
99
100
101
102
103
104
105
106
107
108
            cis = factory.createCompressorInputStream(stream);
        } catch (CompressorException e) {
            throw new TikaException("Unable to uncompress document stream", e);
        }

        MediaType type = getMediaType(cis);
        if (!type.equals(MediaType.OCTET_STREAM)) {
            metadata.set(CONTENT_TYPE, type.toString());
        }

        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        xhtml.startDocument();
View Full Code Here
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
       
        // Grab the file type box
        FileTypeBox fileType = getOrNull(isoFile, FileTypeBox.class);
        if (fileType != null) {
           // Identify the type
           MediaType type = MediaType.application("mp4");
           for (MediaType t : typesMap.keySet()) {
              if (typesMap.get(t).contains(fileType.getMajorBrand())) {
                 type = t;
                 break;
              }
           }
           metadata.set(Metadata.CONTENT_TYPE, type.toString());
          
           if (type.getType().equals("audio")) {
              metadata.set(XMPDM.AUDIO_COMPRESSOR, fileType.getMajorBrand().trim());
           }
        } else {
           // Some older QuickTime files lack the FileType
           metadata.set(Metadata.CONTENT_TYPE, "video/quicktime");
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.