Package org.apache.tika.parser

Examples of org.apache.tika.parser.ParseContext

300
301
302
303
304
305
306
307
308
        String test = "<html><title>Simple Content</title><body></body></html>";
        Metadata metadata = new Metadata();
        metadata.add(Metadata.CONTENT_LANGUAGE, "en");
        new HtmlParser().parse (
                new ByteArrayInputStream(test.getBytes("UTF-8")),
                new BodyContentHandler(),  metadata, new ParseContext());

        assertEquals("en", metadata.get(Metadata.CONTENT_LANGUAGE));
    }
View Full Code Here
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
            + "<title>the name is \u00e1ndre</title>"
            + "</head><body></body></html>";
        Metadata metadata = new Metadata();
        new HtmlParser().parse (
                new ByteArrayInputStream(test1.getBytes("UTF-8")),
                new BodyContentHandler(),  metadata, new ParseContext());
        assertEquals("ISO-8859-1", metadata.get(Metadata.CONTENT_ENCODING));

        // Some HTML pages have errors like ';;' versus '; ' as separator
        String test2 =
            "<html><head><meta http-equiv=\"content-type\""
            + " content=\"text/html;;charset=ISO-8859-1\" />"
            + "<title>the name is \u00e1ndre</title>"
            + "</head><body></body></html>";
        metadata = new Metadata();
        new HtmlParser().parse (
                new ByteArrayInputStream(test2.getBytes("UTF-8")),
                new BodyContentHandler(),  metadata, new ParseContext());
        assertEquals("ISO-8859-1", metadata.get(Metadata.CONTENT_ENCODING));
    }
View Full Code Here
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
            + "<body></body></html>";

        Metadata metadata = new Metadata();
        new HtmlParser().parse (
                new ByteArrayInputStream(test.getBytes("UTF-8")),
                new BodyContentHandler(),  metadata, new ParseContext());
        assertEquals("UTF-8", metadata.get(Metadata.CONTENT_ENCODING));

        metadata = new Metadata();
        metadata.set(Metadata.CONTENT_TYPE, "charset=ISO-8859-1;text/html");
        new HtmlParser().parse (
                new ByteArrayInputStream(test.getBytes("UTF-8")),
                new BodyContentHandler(),  metadata, new ParseContext());
        assertEquals("ISO-8859-1", metadata.get(Metadata.CONTENT_ENCODING));
    }
View Full Code Here
367
368
369
370
371
372
373
374
375
    public void testMetaHttpEquivWithLotsOfPreambleText() throws Exception {
        String path = "/test-documents/big-preamble.html";
        Metadata metadata = new Metadata();
        new HtmlParser().parse(
                HtmlParserTest.class.getResourceAsStream(path),
                new BodyContentHandler(),  metadata, new ParseContext());

        assertEquals("windows-1251", metadata.get(Metadata.CONTENT_ENCODING));
    }
View Full Code Here
339
340
341
342
343
344
345
     * @deprecated This method will be removed in Apache Tika 1.0.
     */
    public void parse(
            InputStream stream, ContentHandler handler, Metadata metadata)
            throws IOException, SAXException, TikaException {
        parse(stream, handler, metadata, new ParseContext());
    }
View Full Code Here
120
121
122
123
124
125
126
127
128
    private String encoding = null;

    private boolean pipeMode = true;

    public TikaCLI() throws TransformerConfigurationException {
        context = new ParseContext();
        parser = new AutoDetectParser();
        context.set(Parser.class, parser);
    }
View Full Code Here
43
44
45
46
47
48
49
     * @deprecated This method will be removed in Apache Tika 1.0.
     */
    public void parse(
            InputStream stream, ContentHandler handler, Metadata metadata)
            throws IOException, SAXException, TikaException {
        parse(stream, handler, metadata, new ParseContext());
    }
View Full Code Here
86
87
88
89
90
91
92
     * @deprecated This method will be removed in Apache Tika 1.0.
     */
    public void parse(
            InputStream stream, ContentHandler handler, Metadata metadata)
            throws IOException, SAXException, TikaException {
        parse(stream, handler, metadata, new ParseContext());
    }
View Full Code Here
196
197
198
199
200
201
202
     * @deprecated This method will be removed in Apache Tika 1.0.
     */
    public void parse(
            InputStream stream, ContentHandler handler, Metadata metadata)
    throws IOException, SAXException, TikaException {
        parse(stream, handler, metadata, new ParseContext());
    }
View Full Code Here
96
97
98
99
100
101
102
     * @deprecated This method will be removed in Apache Tika 1.0.
     */
    public void parse(
            InputStream stream, ContentHandler handler, Metadata metadata)
            throws IOException, SAXException, TikaException {
        parse(stream, handler, metadata, new ParseContext());
    }
View Full Code Here
TOP

Related Classes of org.apache.tika.parser.ParseContext

  • com.cloudera.cdk.morphline.solrcell.SolrCellBuilder$SolrCell
  • com.nidhinova.tika.server.TikaService
  • com.tamingtext.tika.TikaTest
  • cx.fbn.nevernote.threads.IndexRunner
  • gov.nara.nwts.ftappImg.jpeg.JpegExtractor
  • org.apache.chemistry.opencmis.client.parser.MetadataParserTika
  • org.apache.jackrabbit.core.query.lucene.JackrabbitParser
  • org.apache.jackrabbit.core.query.lucene.LazyTextExtractorField$ParsingTask
  • org.apache.jackrabbit.oak.plugins.index.lucene.LuceneIndexEditor
  • org.apache.nutch.parse.tika.DOMContentUtilsTest
    • 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.