Package org.apache.tika.sax.xpath

Examples of org.apache.tika.sax.xpath.Matcher

37
38
39
40
41
42
43
44
45
46
47
    // eg <meta:user-defined meta:name="Info1">Text1</meta:user-defined> becomes custom:Info1=Text1
    public static final String USER_DEFINED_METADATA_NAME_PREFIX = "custom:";

    private static ContentHandler getMeta(
            ContentHandler ch, Metadata md, String name, String element) {
        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
47
48
49
50
51
52
53
54
55
56
57
        return new TeeContentHandler(ch, branch);
    }

    private static ContentHandler getUserDefined(
            ContentHandler ch, Metadata md) {
        Matcher matcher = new CompositeMatcher(
                META_XPATH.parse("//meta:user-defined/@meta:name"),
                META_XPATH.parse("//meta:user-defined//text()"));
        ContentHandler branch = new MatchingContentHandler(
              new AttributeDependantMetadataHandler(md, "meta:name", USER_DEFINED_METADATA_NAME_PREFIX),
              matcher);
View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
        return new TeeContentHandler(ch, branch);
    }

    private static ContentHandler getStatistic(
            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
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
            serializer.setOutputFormat(new OutputFormat("Text", "UTF-8", true));
          } else {
            serializer = new XMLSerializer(writer, new OutputFormat("XML", "UTF-8", true));
          }
          if (xpathExpr != null) {
            Matcher matcher =
                    PARSER.parse(xpathExpr);
            serializer.startDocument();//The MatchingContentHandler does not invoke startDocument.  See http://tika.markmail.org/message/kknu3hw7argwiqin
            parsingHandler = new MatchingContentHandler(serializer, matcher);
          } else {
            parsingHandler = serializer;
          }
        } else if (xpathExpr != null) {
          Matcher matcher =
                  PARSER.parse(xpathExpr);
          parsingHandler = new MatchingContentHandler(handler, matcher);
        } //else leave it as is

        try{
View Full Code Here
34
35
36
37
38
39
40
41
42
43
44
    private static final XPathParser META_XPATH = new XPathParser(
            "meta", "urn:oasis:names:tc:opendocument:xmlns:meta:1.0");

    private static ContentHandler getMeta(
            ContentHandler ch, Metadata md, String name, String element) {
        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
44
45
46
47
48
49
50
51
52
53
54
        return new TeeContentHandler(ch, branch);
    }

    private static ContentHandler getStatistic(
            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
37
38
39
40
41
42
43
44
45
46
47
    // eg <meta:user-defined meta:name="Info1">Text1</meta:user-defined> becomes custom:Info1=Text1
    public static final String USER_DEFINED_METADATA_NAME_PREFIX = "custom:";

    private static ContentHandler getMeta(
            ContentHandler ch, Metadata md, String name, String element) {
        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
47
48
49
50
51
52
53
54
55
56
57
        return new TeeContentHandler(ch, branch);
    }

    private static ContentHandler getUserDefined(
            ContentHandler ch, Metadata md) {
        Matcher matcher = new CompositeMatcher(
                META_XPATH.parse("//meta:user-defined/@meta:name"),
                META_XPATH.parse("//meta:user-defined//text()"));
        ContentHandler branch = new MatchingContentHandler(
              new AttributeDependantMetadataHandler(md, "meta:name", USER_DEFINED_METADATA_NAME_PREFIX),
              matcher);
View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
        return new TeeContentHandler(ch, branch);
    }

    private static ContentHandler getStatistic(
            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
33
34
35
36
37
38
39
40
41
42
43
    private static final XPathParser DC_XPATH = new XPathParser(
            "dc", "http://purl.org/dc/elements/1.1/");

    private static ContentHandler getDublinCore(
            ContentHandler ch, Metadata md, String name, String element) {
        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
TOP

Related Classes of org.apache.tika.sax.xpath.Matcher

  • com.cloudera.cdk.morphline.solrcell.SolrCellBuilder$SolrCell
  • org.apache.solr.handler.extraction.ExtractingDocumentLoader
  • org.apache.tika.example.ContentHandlerExample
  • org.apache.tika.parser.html.HtmlParser
  • org.apache.tika.parser.odf.OpenDocumentMetaParser
  • org.apache.tika.parser.opendocument.OpenOfficeMetaParser
  • org.apache.tika.parser.xml.DcXMLParser
  • org.kitesdk.morphline.solrcell.SolrCellBuilder$SolrCell
    • 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.