Package org.apache.tika.sax

Examples of org.apache.tika.sax.TeeContentHandler

51
52
53
54
55
56
57
58
59
60
61
     */
    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
39
40
41
42
43
44
45
        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
134
135
136
137
138
139
140
141
142
143
144
            StringWriter htmlBuffer = new StringWriter();
            StringWriter textBuffer = new StringWriter();
            StringWriter xmlBuffer = new StringWriter();
            StringBuilder metadataBuffer = new StringBuilder();

            ContentHandler handler = new TeeContentHandler(
                    getHtmlHandler(htmlBuffer),
                    getTextContentHandler(textBuffer),
                    getXmlContentHandler(xmlBuffer));
            Metadata md = new Metadata();
            md.set(Metadata.RESOURCE_NAME_KEY, file.getName());
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
40
41
42
43
44
45
46
        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
49
50
51
52
53
54
55
            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
45
46
47
48
49
50
51
        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
57
58
59
60
61
62
63
                META_XPATH.parse("//meta:user-defined//text()"));
        // eg <meta:user-defined meta:name="Info1">Text1</meta:user-defined> becomes custom:Info1=Text1
        ContentHandler branch = new MatchingContentHandler(
              new AttributeDependantMetadataHandler(md, "meta:name", Metadata.USER_DEFINED_METADATA_NAME_PREFIX),
              matcher);
        return new TeeContentHandler(ch, branch);
    }
View Full Code Here
66
67
68
69
70
71
72
            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
134
135
136
137
138
139
140
141
142
143
144
            StringWriter htmlBuffer = new StringWriter();
            StringWriter textBuffer = new StringWriter();
            StringWriter xmlBuffer = new StringWriter();
            StringBuilder metadataBuffer = new StringBuilder();

            ContentHandler handler = new TeeContentHandler(
                    getHtmlHandler(htmlBuffer),
                    getTextContentHandler(textBuffer),
                    getXmlContentHandler(xmlBuffer));
            Metadata md = new Metadata();
            md.set(Metadata.RESOURCE_NAME_KEY, file.getName());
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.