Package org.apache.tika.sax

Examples of org.apache.tika.sax.TeeContentHandler

50
51
52
53
54
55
56
57
58
59
60
     */
    public void parse(
            InputStream stream, ContentHandler handler, Metadata metadata)
            throws IOException, SAXException, TikaException {
        ContentHandler body = new BodyContentHandler();
        super.parse(stream, new TeeContentHandler(handler, body), metadata);

        String content = body.toString();
        metadata.set("fulltext", content);

        int length = Math.min(content.length(), 500);
View Full Code Here
38
39
40
41
42
43
44
        Matcher matcher = new CompositeMatcher(
                DC_XPATH.parse("//dc:" + element),
                DC_XPATH.parse("//dc:" + element + "//text()"));
        ContentHandler branch =
            new MatchingContentHandler(new MetadataHandler(md, name), matcher);
        return new TeeContentHandler(ch, branch);
    }
View Full Code Here
39
40
41
42
43
44
45
        Matcher matcher = new CompositeMatcher(
                META_XPATH.parse("//meta:" + element),
                META_XPATH.parse("//meta:" + element + "//text()"));
        ContentHandler branch =
            new MatchingContentHandler(new MetadataHandler(md, name), matcher);
        return new TeeContentHandler(ch, branch);
    }
View Full Code Here
48
49
50
51
52
53
54
            ContentHandler ch, Metadata md, String name, String attribute) {
        Matcher matcher =
            META_XPATH.parse("//meta:document-statistic/@meta:" + attribute);
        ContentHandler branch =
            new MatchingContentHandler(new MetadataHandler(md, name), matcher);
        return new TeeContentHandler(ch, branch);
    }
View Full Code Here
95
96
97
98
99
100
101
102
103
104
105
        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        XPathParser xpath = new XPathParser(null, "");
        Matcher body = xpath.parse("/HTML/BODY//node()");
        Matcher title = xpath.parse("/HTML/HEAD/TITLE//node()");
        Matcher meta = xpath.parse("/HTML/HEAD/META//node()");
        handler = new TeeContentHandler(
                new MatchingContentHandler(getBodyHandler(xhtml), body),
                new MatchingContentHandler(getTitleHandler(metadata), title),
                new MatchingContentHandler(getMetaHandler(metadata), meta));

        // Parse the HTML document
View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
            }
        };
        Metadata metadata = new Metadata();
        InputStream stream = getStream("test-documents/testHTML.html");
        try {
            parser.parse(stream, new TeeContentHandler(body, link), metadata);
        } finally {
            stream.close();
        }

        assertEquals(
View Full Code Here
68
69
70
71
72
73
74
75
76
77
78
                        }
                    }
                }
            };
            new HtmlParser().parse(
                    stream, new TeeContentHandler(body, link),
                    metadata, new ParseContext());
        } finally {
            stream.close();
        }
View Full Code Here
89
90
91
92
93
94
95
96
97
98
99
   
    BoilerpipeContentHandler mainContentHandler = new BoilerpipeContentHandler(mainContentBuffer);
    BodyContentHandler bodyHandler = new BodyContentHandler(bodyBuffer);
    LinkContentHandler linkHandler = new LinkContentHandler();
   
    TeeContentHandler parallelHandler = new TeeContentHandler(xmlHandler, mainContentHandler, bodyHandler, linkHandler );

    InputStream instream = entity.obtainContent();
    try {
      parser.parse(instream, parallelHandler, metadata, new ParseContext());
     
View Full Code Here
86
87
88
89
90
91
92
93
94
95
96
   
    BoilerpipeContentHandler mainContentHandler = new BoilerpipeContentHandler(mainContentBuffer);
    BodyContentHandler bodyHandler = new BodyContentHandler(bodyBuffer);
    LinkContentHandler linkHandler = new LinkContentHandler();
   
    TeeContentHandler parallelHandler = new TeeContentHandler(xmlHandler, mainContentHandler, bodyHandler, linkHandler );

    InputStream instream = entity.obtainContent();
    try {
      parser.parse(instream, parallelHandler, metadata, new ParseContext());
     
View Full Code Here
234
235
236
237
238
239
240
241
242
243
244
        ContentHandler parsingHandler = handler;
        StringWriter debugWriter = null;
        if (LOG.isTraceEnabled()) {
          debugWriter = new StringWriter();
          ContentHandler serializer = new XMLSerializer(debugWriter, new OutputFormat("XML", "UTF-8", true));
          parsingHandler = new TeeContentHandler(parsingHandler, serializer);
        }

        // String xpathExpr = "/xhtml:html/xhtml:body/xhtml:div/descendant:node()";
        if (xpathExpr != null) {
          Matcher matcher = PARSER.parse(xpathExpr);
View Full Code Here
TOP

Related Classes of org.apache.tika.sax.TeeContentHandler

  • bixo.parser.TikaCallable
  • com.cloudera.cdk.morphline.solrcell.SolrCellBuilder$SolrCell
  • com.digitalpebble.storm.crawler.bolt.ParserBolt
  • com.tamingtext.tika.TikaTest
  • org.apache.droids.tika.TikaDocumentParser
  • org.apache.droids.tika.TikaHtmlParser
  • org.apache.tika.fork.ForkParser
  • org.apache.tika.gui.TikaGUI
  • org.apache.tika.parser.html.HtmlParser
  • org.apache.tika.parser.html.HtmlParserTest
    • 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.