Package org.apache.tika.mime

Examples of org.apache.tika.mime.MediaType

38
39
40
41
42
43
44
45
46
47
48
49
            }
        }

        DefaultDetector detector = new DefaultDetector();
        Metadata metadata = new Metadata();
        MediaType mediaType = detector.detect(new BufferedInputStream(in), metadata);
        mimeType = mediaType.toString();
        if (mimeType != null) {
            return mimeType;
        }

        return "application/octet-stream";
View Full Code Here
30
31
32
33
34
35
36
37
38
39
40
41
42
        return result;
    }
   
    public static String getCharsetFromContentType(String contentType) {
        String result = "";
        MediaType mt = MediaType.parse(contentType);
        if (mt != null) {
            String charset = mt.getParameters().get("charset");
            if (charset != null) {
                result = charset;
            }
        }
       
View Full Code Here
20
21
22
23
24
25
26
27
28
29
30
31
public class HttpUtils {
   
    public static String getMimeTypeFromContentType(String contentType) {
        String result = "";
        MediaType mt = MediaType.parse(contentType);
        if (mt != null) {
            result = mt.getType() + "/" + mt.getSubtype();
        }
       
        return result;
    }
View Full Code Here
75
76
77
78
79
80
81
82
83
84
85
86
87
        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
95
96
97
98
99
100
101
102
103
104
105
106
107
        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
110
111
112
113
114
115
116
117
118
119
120
121
122
        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
83
84
85
86
87
88
89
90
91
92
93
94
95
        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
85
86
87
88
89
90
91
92
93
94
95
96
            throws IOException, SAXException, TikaException {
        // We need (reliable!) mark support for type detection before parsing
        stream = new BufferedInputStream(stream);

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

        // TIKA-216: Zip bomb prevention
        CountingInputStream count = new CountingInputStream(stream);
        SecureContentHandler secure = new SecureContentHandler(handler, count);
View Full Code Here
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
    try {
      Charset charset = reader.getCharset();
      String mediaType = metadata.get(Metadata.CONTENT_TYPE);
      String name = metadata.get(Metadata.RESOURCE_NAME_KEY);
      if (mediaType != null && name != null) {
        MediaType type = MediaType.parse(mediaType);
        metadata.set(Metadata.CONTENT_TYPE, type.toString());
        metadata.set(Metadata.CONTENT_ENCODING, charset.name());

        StringBuilder out = new StringBuilder();
        String line;
        int nbLines =  0;
        while ((line = reader.readLine()) != null) {
            out.append(line);
            String author = parserAuthor(line);
            if (author != null) {
              metadata.add(TikaCoreProperties.CREATOR, author);
            }
            nbLines ++;
        }
        metadata.set("LoC", String.valueOf(nbLines));

        Renderer renderer = getRenderer(type.toString());
        String codeAsHtml = renderer.highlight(name, out.toString(), charset.name(), false);
        char[] charArray = codeAsHtml.toCharArray();
        handler.startDocument();
        handler.characters(charArray, 0, charArray.length);
        handler.endDocument();
View Full Code Here
111
112
113
114
115
116
117
118
119
120
121
    }

  }

  private Renderer getRenderer(String mimeType) {
    MediaType mt = MediaType.parse(mimeType);
    String type = TYPES_TO_RENDERER.get(mt);
    if (type == null) {
      throw new RuntimeException("unparseable content type " + mimeType);
    }
    return XhtmlRendererFactory.getRenderer(type);
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.