Package it.unimi.dsi.parser.callback

Examples of it.unimi.dsi.parser.callback.ComposedCallbackBuilder

125
126
127
128
129
130
131
132
133
134
135
136
137
138
  }

  private void init() {
    this.parser = new BulletParser();
   
    ComposedCallbackBuilder composedBuilder = new ComposedCallbackBuilder();
    composedBuilder.add( this.textExtractor = new TextExtractor() );
    composedBuilder.add( this.anchorExtractor = new AnchorExtractor( maxPreAnchor, maxAnchor, maxPostAnchor ) );
    parser.setCallback( composedBuilder.compose() );

    Object o;
    try {
      o = defaultMetadata.get( PropertyBasedDocumentFactory.MetadataKeys.WORDREADER );
      wordReader = o == null ? new FastBufferedReader() : ObjectParser.fromSpec( o.toString(), WordReader.class, MG4JClassParser.PACKAGE );
View Full Code Here
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
 
  private void init() {
    /** The HTML bullet parser */
    this.parser = new BulletParser(HTMLFactory.INSTANCE);

    ComposedCallbackBuilder composedBuilder = new ComposedCallbackBuilder();

    composedBuilder
        .add(this.textExtractor = new StructuredTextExtractor());
    parser.setCallback(composedBuilder.compose());

    this.wordReader = new FastBufferedReader();
    text = new char[DEFAULT_BUFFER_SIZE];

    if (collectionType == null)
View Full Code Here
827
828
829
830
831
832
833
834
835
836
837
838
839
840
                if (w.isHTMLResponse()) {
                    //System.out.println(w.getHTMLContent());

                    // See how the parsed content looks like
                    BulletParser parser = new BulletParser(TRECParsingFactory.INSTANCE);
                    ComposedCallbackBuilder composedBuilder = new ComposedCallbackBuilder();
                    StructuredTextExtractor textExtractor = new StructuredTextExtractor();
                    composedBuilder.add(textExtractor);
                    parser.setCallback(composedBuilder.compose());
                    parser.parse(w.getHTMLContent().toCharArray());
                    System.out.println(textExtractor.getText());
                }
            }
            in.close();
View Full Code Here
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
        LOGGER.info("Initialising the TREC document factory");

        // The parser is a SGML BulletParser with TREC vocabulary
        this.parser = new BulletParser(TRECParsingFactory.INSTANCE);

        ComposedCallbackBuilder composedBuilder = new ComposedCallbackBuilder();

        composedBuilder.add(this.textExtractor = new StructuredTextExtractor());

        this.textExtractor.ignore(
                TRECParsingFactory.ELEMENT_DOCNO,
                TRECParsingFactory.ELEMENT_FILEID,
                TRECParsingFactory.ELEMENT_FIRST,
                TRECParsingFactory.ELEMENT_SECOND
        );
        parser.setCallback(composedBuilder.compose());

        this.wordReader = new FastBufferedReader();
        text = new char[DEFAULT_BUFFER_SIZE];
    }
View Full Code Here
TOP

Related Classes of it.unimi.dsi.parser.callback.ComposedCallbackBuilder

  • it.unimi.dsi.mg4j.document.HtmlDocumentFactory
  • net.bpiwowar.mg4j.extensions.trec.TRECDocumentFactory
  • net.bpiwowar.mg4j.extensions.utils.HTMLDocumentFactory
  • net.bpiwowar.mg4j.extensions.warc.WarcRecord
    • 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.