Package org.apache.tika

Examples of org.apache.tika.Tika

27
28
29
30
31
32
33
34
35
36
37
public class ClassParserTest extends TestCase {

    public void testClassParsing() throws Exception {
        String path = "/test-documents/AutoDetectParser.class";
        Metadata metadata = new Metadata();
        String content = new Tika().parseToString(
                ClassParserTest.class.getResourceAsStream(path), metadata);

        assertEquals("AutoDetectParser", metadata.get(Metadata.TITLE));
        assertEquals(
                "AutoDetectParser.class",
View Full Code Here
85
86
87
88
89
90
91
92
93
94
95
    }

    public void XtestParseUTF8() throws IOException, SAXException, TikaException {
        String path = "/test-documents/testXHTML_utf8.html";
        Metadata metadata = new Metadata();
        String content = new Tika().parseToString(
                HtmlParserTest.class.getResourceAsStream(path), metadata);

        assertTrue("Did not contain expected text:"
                + "Title : Tilte with UTF-8 chars öäå", content
                .contains("Title : Tilte with UTF-8 chars öäå"));
View Full Code Here
103
104
105
106
107
108
109
110
111
112
113
    }

    public void testXhtmlParsing() throws Exception {
        String path = "/test-documents/testXHTML.html";
        Metadata metadata = new Metadata();
        String content = new Tika().parseToString(
                HtmlParserTest.class.getResourceAsStream(path), metadata);

        assertEquals("application/xhtml+xml", metadata.get(Metadata.CONTENT_TYPE));
        assertEquals("XHTML test document", metadata.get(Metadata.TITLE));
View Full Code Here
130
131
132
133
134
135
136
137
138
     * Test case for TIKA-210
     * @see <a href="https://issues.apache.org/jira/browse/TIKA-210">TIKA-210</a>
     */
    public void testCharactersDirectlyUnderBodyElement() throws Exception {
        String test = "<html><body>test</body></html>";
        String content = new Tika().parseToString(
                new ByteArrayInputStream(test.getBytes("UTF-8")));
        assertEquals("test", content);
    }
View Full Code Here
210
211
212
213
214
215
216
217
218
219
220
     * @see <a href="https://issues.apache.org/jira/browse/TIKA-268">TIKA-268</a>
     */
    public void testWhitespaceBetweenTableCells() throws Exception {
        String test =
            "<html><body><table><tr><td>a</td><td>b</td></table></body></html>";
        String content = new Tika().parseToString(
                new ByteArrayInputStream(test.getBytes("UTF-8")));
        assertTrue(content.contains("a"));
        assertTrue(content.contains("b"));
        assertFalse(content.contains("ab"));
    }
View Full Code Here
24
25
26
27
28
29
30
31
32
33
34
public class AudioParserTest extends TestCase {

    public void testWAV() throws Exception {
        String path = "/test-documents/testWAV.wav";
        Metadata metadata = new Metadata();
        String content = new Tika().parseToString(
                AudioParserTest.class.getResourceAsStream(path), metadata);

        assertEquals("audio/x-wav", metadata.get(Metadata.CONTENT_TYPE));
        assertEquals("44100.0", metadata.get("samplerate"));
        assertEquals("2", metadata.get("channels"));
View Full Code Here
39
40
41
42
43
44
45
46
47
48
49
    }

    public void testAIFF() throws Exception {
        String path = "/test-documents/testAIFF.aif";
        Metadata metadata = new Metadata();
        String content = new Tika().parseToString(
                AudioParserTest.class.getResourceAsStream(path), metadata);

        assertEquals("audio/x-aiff", metadata.get(Metadata.CONTENT_TYPE));
        assertEquals("44100.0", metadata.get("samplerate"));
        assertEquals("2", metadata.get("channels"));
View Full Code Here
54
55
56
57
58
59
60
61
62
63
64
    }

    public void testAU() throws Exception {
        String path = "/test-documents/testAU.au";
        Metadata metadata = new Metadata();
        String content = new Tika().parseToString(
                AudioParserTest.class.getResourceAsStream(path), metadata);

        assertEquals("audio/basic", metadata.get(Metadata.CONTENT_TYPE));
        assertEquals("44100.0", metadata.get("samplerate"));
        assertEquals("2", metadata.get("channels"));
View Full Code Here
24
25
26
27
28
29
30
31
32
33
34
public class MidiParserTest extends TestCase {

    public void testMID() throws Exception {
        String path = "/test-documents/testMID.mid";
        Metadata metadata = new Metadata();
        String content = new Tika().parseToString(
                MidiParserTest.class.getResourceAsStream(path), metadata);

        assertEquals("audio/midi", metadata.get(Metadata.CONTENT_TYPE));
        assertEquals("2", metadata.get("tracks"));
        assertEquals("0", metadata.get("patches"));
View Full Code Here
97
98
99
100
101
102
103
104
105
106
107
    }

    public void XtestParseUTF8() throws IOException, SAXException, TikaException {
        String path = "/test-documents/testXHTML_utf8.html";
        Metadata metadata = new Metadata();
        String content = new Tika().parseToString(
                HtmlParserTest.class.getResourceAsStream(path), metadata);

        assertTrue("Did not contain expected text:"
                + "Title : Tilte with UTF-8 chars öäå", content
                .contains("Title : Tilte with UTF-8 chars öäå"));
View Full Code Here
TOP

Related Classes of org.apache.tika.Tika

  • ch.entwine.weblounge.kernel.site.SiteServlet
  • com.digitalpebble.storm.crawler.bolt.ParserBolt
  • com.gentics.cr.file.ResolvableFileBean
  • com.gitblit.servlet.RawServlet
  • com.ikanow.infinit.e.harvest.extraction.document.file.FileHarvester
  • com.ikanow.infinit.e.harvest.extraction.text.boilerpipe.TextExtractorBoilerpipe
  • com.ikanow.infinit.e.harvest.extraction.text.legacy.TextExtractorTika
  • com.javajing.webjar.FileUploadServlet
  • com.trsst.Command
  • com.trsst.ui.AppServlet
    • 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.