Package org.apache.tika.sax.xpath

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

171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
            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

        //potentially use a wrapper handler for parsing, but we still need the SolrContentHandler for getting the document.
View Full Code Here
70
71
72
73
74
75
76
77
78
79
80
                metadata, property);
    }
   
    private static ContentHandler getMeta(
            ContentHandler ch, Metadata md, Property property, String element) {
        Matcher matcher = new CompositeMatcher(
                META_XPATH.parse("//meta:" + element),
                META_XPATH.parse("//meta:" + element + "//text()"));
        ContentHandler branch =
            new MatchingContentHandler(new MetadataHandler(md, property), matcher);
        return new TeeContentHandler(ch, branch);
View Full Code Here
80
81
82
83
84
85
86
87
88
89
90
        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()"));
        // 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),
View Full Code Here
92
93
94
95
96
97
98
99
100
101
102
        return new TeeContentHandler(ch, branch);
    }

    @Deprecated 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 AttributeMetadataHandler(META_NS, attribute, md, name), matcher);
        return new TeeContentHandler(ch, branch);
    }
View Full Code Here
100
101
102
103
104
105
106
107
108
109
110
              new AttributeMetadataHandler(META_NS, attribute, md, name), matcher);
        return new TeeContentHandler(ch, branch);
    }
    private static ContentHandler getStatistic(
          ContentHandler ch, Metadata md, Property property, String attribute) {
      Matcher matcher =
          META_XPATH.parse("//meta:document-statistic/@meta:"+attribute);
      ContentHandler branch = new MatchingContentHandler(
            new AttributeMetadataHandler(META_NS, attribute, md, property), matcher);
      return new TeeContentHandler(ch, branch);
  }
View Full Code Here
34
35
36
37
38
39
40
41
42
43
44
    private static final XPathParser DC_XPATH = new XPathParser(
            "dc", "http://purl.org/dc/elements/1.1/");

    private static DefaultHandler 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
92
93
94
95
96
97
98
99
100
101
102
103
104
        // Prepare the HTML content handler that generates proper
        // XHTML events to records relevant document metadata
        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));
View Full Code Here
35
36
37
38
39
40
41
42
43
44
45
    private static final XPathParser META_XPATH = new XPathParser(
            "meta", "urn:oasis:names:tc:opendocument:xmlns:meta:1.0");

    private static DefaultHandler 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
45
46
47
48
49
50
51
52
53
54
55
        return new TeeContentHandler(ch, branch);
    }

    private static DefaultHandler 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
40
41
42
43
44
45
46
47
48
49
50
    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
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.