Example 56 with Metadata
use of org.apache.tika.metadata.Metadata in project tika by apache.
the class AudioParserTest method testAU.
@Test
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"));
assertEquals("16", metadata.get("bits"));
assertEquals("PCM_SIGNED", metadata.get("encoding"));
assertEquals("", content);
}Example 57 with Metadata
use of org.apache.tika.metadata.Metadata in project tika by apache.
the class MidiParserTest method testMID.
@Test
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"));
assertEquals("PPQ", metadata.get("divisionType"));
assertContains("Untitled", content);
}Example 58 with Metadata
use of org.apache.tika.metadata.Metadata in project tika by apache.
the class DWGParserTest method testParser.
@SuppressWarnings("deprecation")
private void testParser(InputStream input) throws Exception {
try {
Metadata metadata = new Metadata();
ContentHandler handler = new BodyContentHandler();
new DWGParser().parse(input, handler, metadata);
assertEquals("image/vnd.dwg", metadata.get(Metadata.CONTENT_TYPE));
assertEquals("The quick brown fox jumps over the lazy dog", metadata.get(TikaCoreProperties.TITLE));
assertEquals("Gym class featuring a brown fox and lazy dog", metadata.get(TikaCoreProperties.DESCRIPTION));
assertEquals("Gym class featuring a brown fox and lazy dog", metadata.get(Metadata.SUBJECT));
assertEquals("Nevin Nollop", metadata.get(TikaCoreProperties.CREATOR));
assertEquals("Pangram, fox, dog", metadata.get(TikaCoreProperties.KEYWORDS));
assertEquals("Lorem ipsum", metadata.get(TikaCoreProperties.COMMENTS).substring(0, 11));
assertEquals("http://www.alfresco.com", metadata.get(TikaCoreProperties.RELATION));
// Check some of the old style metadata too
assertEquals("The quick brown fox jumps over the lazy dog", metadata.get(Metadata.TITLE));
assertEquals("Gym class featuring a brown fox and lazy dog", metadata.get(Metadata.SUBJECT));
String content = handler.toString();
assertContains("The quick brown fox jumps over the lazy dog", content);
assertContains("Gym class", content);
assertContains("www.alfresco.com", content);
} finally {
input.close();
}
}
Also used :
BodyContentHandler(org.apache.tika.sax.BodyContentHandler)
Metadata(org.apache.tika.metadata.Metadata)
BodyContentHandler(org.apache.tika.sax.BodyContentHandler)
ContentHandler(org.xml.sax.ContentHandler)
Example 59 with Metadata
use of org.apache.tika.metadata.Metadata in project tika by apache.
the class EnviHeaderParserTest method testParseGlobalMetadata.
@Test
public void testParseGlobalMetadata() throws Exception {
if (System.getProperty("java.version").startsWith("1.5")) {
return;
}
Parser parser = new EnviHeaderParser();
ToXMLContentHandler handler = new ToXMLContentHandler();
Metadata metadata = new Metadata();
try (InputStream stream = EnviHeaderParser.class.getResourceAsStream("/test-documents/envi_test_header.hdr")) {
assertNotNull("Test ENVI file not found", stream);
parser.parse(stream, handler, metadata, new ParseContext());
}
// Check content of test file
String content = handler.toString();
assertContains("<body><p>ENVI</p>", content);
assertContains("<p>samples = 2400</p>", content);
assertContains("<p>lines = 2400</p>", content);
assertContains("<p>map info = {Sinusoidal, 1.5000, 1.5000, -10007091.3643, 5559289.2856, 4.6331271653e+02, 4.6331271653e+02, , units=Meters}</p>", content);
assertContains("content=\"application/envi.hdr\"", content);
assertContains("projection info = {16, 6371007.2, 0.000000, 0.0, 0.0, Sinusoidal, units=Meters}", content);
}
Also used :
ToXMLContentHandler(org.apache.tika.sax.ToXMLContentHandler)
InputStream(java.io.InputStream)
Metadata(org.apache.tika.metadata.Metadata)
ParseContext(org.apache.tika.parser.ParseContext)
Parser(org.apache.tika.parser.Parser)
Test(org.junit.Test)
Example 60 with Metadata
use of org.apache.tika.metadata.Metadata in project tika by apache.
the class DWGParserTest method testDWG2010CustomPropertiesParser.
@Test
public void testDWG2010CustomPropertiesParser() throws Exception {
// Check that standard parsing works
InputStream testInput = DWGParserTest.class.getResourceAsStream("/test-documents/testDWG2010_custom_props.dwg");
testParser(testInput);
// Check that custom properties with alternate padding work
try (InputStream input = DWGParserTest.class.getResourceAsStream("/test-documents/testDWG2010_custom_props.dwg")) {
Metadata metadata = new Metadata();
ContentHandler handler = new BodyContentHandler();
new DWGParser().parse(input, handler, metadata, null);
assertEquals("valueforcustomprop1", metadata.get("customprop1"));
assertEquals("valueforcustomprop2", metadata.get("customprop2"));
}
}
Also used :
BodyContentHandler(org.apache.tika.sax.BodyContentHandler)
InputStream(java.io.InputStream)
Metadata(org.apache.tika.metadata.Metadata)
BodyContentHandler(org.apache.tika.sax.BodyContentHandler)
ContentHandler(org.xml.sax.ContentHandler)
Test(org.junit.Test)
Aggregations
Metadata (org.apache.tika.metadata.Metadata)651
Test (org.junit.Test)467
InputStream (java.io.InputStream)320
ParseContext (org.apache.tika.parser.ParseContext)283
BodyContentHandler (org.apache.tika.sax.BodyContentHandler)269
TikaTest (org.apache.tika.TikaTest)257
ContentHandler (org.xml.sax.ContentHandler)229
AutoDetectParser (org.apache.tika.parser.AutoDetectParser)154
ByteArrayInputStream (java.io.ByteArrayInputStream)143
Parser (org.apache.tika.parser.Parser)136
TikaInputStream (org.apache.tika.io.TikaInputStream)133
IOException (java.io.IOException)66
DefaultHandler (org.xml.sax.helpers.DefaultHandler)59
TikaException (org.apache.tika.exception.TikaException)48
ExcelParserTest (org.apache.tika.parser.microsoft.ExcelParserTest)36
WordParserTest (org.apache.tika.parser.microsoft.WordParserTest)36
StringWriter (java.io.StringWriter)33
Tika (org.apache.tika.Tika)29
MediaType (org.apache.tika.mime.MediaType)29
SAXException (org.xml.sax.SAXException)29
