Package org.apache.tika.parser

Examples of org.apache.tika.parser.Parser

173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
     * The from isn't in the usual form.
     * See TIKA-618
     */
    @Test
    public void testUnusualFromAddress() throws Exception {
       Parser parser = new RFC822Parser();
       Metadata metadata = new Metadata();
       InputStream stream = getStream("test-documents/testRFC822_oddfrom");
       ContentHandler handler = mock(DefaultHandler.class);

       parser.parse(stream, handler, metadata, new ParseContext());
       assertEquals("Saved by Windows Internet Explorer 7",
               metadata.get(TikaCoreProperties.CREATOR));
       assertEquals("Air Permit Programs | Air & Radiation | US EPA",
               metadata.get(TikaCoreProperties.TITLE));
       assertEquals("Air Permit Programs | Air & Radiation | US EPA",
View Full Code Here
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
        assertEquals(2, content.split("<\\/body>").length);
    }

    @Test
    public void testOutlookForwarded() throws Exception {
        Parser parser = new AutoDetectParser();
        Metadata metadata = new Metadata();
      
        // Check the HTML version
        StringWriter sw = new StringWriter();
        SAXTransformerFactory factory = (SAXTransformerFactory)
                 SAXTransformerFactory.newInstance();
        TransformerHandler handler = factory.newTransformerHandler();
        handler.getTransformer().setOutputProperty(OutputKeys.METHOD, "xml");
        handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "yes");
        handler.setResult(new StreamResult(sw));

        InputStream stream = OutlookParserTest.class.getResourceAsStream(
               "/test-documents/testMSG_forwarded.msg");
        try {
           parser.parse(stream, handler, metadata, new ParseContext());
        } finally {
           stream.close();
        }
        
        // Make sure we don't have nested docs
View Full Code Here
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
                    "really really really really really really long name ");
        }
        String name = inputBuilder.toString();
        byte[] data = ("From: " + name + "\r\n\r\n").getBytes("US-ASCII");

        Parser parser = new RFC822Parser();
        ContentHandler handler = new DefaultHandler();
        Metadata metadata = new Metadata();
        ParseContext context = new ParseContext();

        try {
            parser.parse(
                    new ByteArrayInputStream(data), handler, metadata, context);
            fail();
        } catch (TikaException expected) {
        }

        MimeConfig config = new MimeConfig();
        config.setMaxHeaderLen(-1);
        config.setMaxLineLen(-1);
        context.set(MimeConfig.class, config);
        parser.parse(
                new ByteArrayInputStream(data), handler, metadata, context);
        assertEquals(name.trim(), metadata.get(TikaCoreProperties.CREATOR));
    }
View Full Code Here
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
        assertEquals(2, content.split("<\\/body>").length);
    }
   
    @Test
    public void testOutlookHTMLfromRTF() throws Exception {
        Parser parser = new AutoDetectParser();
        Metadata metadata = new Metadata();
      
        // Check the HTML version
        StringWriter sw = new StringWriter();
        SAXTransformerFactory factory = (SAXTransformerFactory)
                 SAXTransformerFactory.newInstance();
        TransformerHandler handler = factory.newTransformerHandler();
        handler.getTransformer().setOutputProperty(OutputKeys.METHOD, "xml");
        handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "yes");
        handler.setResult(new StreamResult(sw));

        InputStream stream = OutlookParserTest.class.getResourceAsStream(
                "/test-documents/test-outlook2003.msg");
        try {
           parser.parse(stream, handler, metadata, new ParseContext());
        } finally {
           stream.close();
        }
        
        // As the HTML version should have been processed, ensure
View Full Code Here
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
    /**
     * Test for TIKA-678 - not all headers may be present
     */
    @Test
    public void testSomeMissingHeaders() throws Exception {
       Parser parser = new RFC822Parser();
       Metadata metadata = new Metadata();
       InputStream stream = getStream("test-documents/testRFC822-limitedheaders");
       ContentHandler handler = new BodyContentHandler();

       parser.parse(stream, handler, metadata, new ParseContext());
       assertEquals(true, metadata.isMultiValued(TikaCoreProperties.CREATOR));
       assertEquals("xyz", metadata.getValues(TikaCoreProperties.CREATOR)[0]);
       assertEquals("abc", metadata.getValues(TikaCoreProperties.CREATOR)[1]);
       assertEquals(true, metadata.isMultiValued(Metadata.MESSAGE_FROM));
       assertEquals("xyz", metadata.getValues(Metadata.MESSAGE_FROM)[0]);
View Full Code Here
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
public class MboxParserTest {

    @Test
    public void testSimple() {
        Parser parser = new MboxParser();
        Metadata metadata = new Metadata();
        InputStream stream = getStream("test-documents/simple.mbox");
        ContentHandler handler = mock(DefaultHandler.class);

        try {
            parser.parse(stream, handler, metadata, new ParseContext());
            verify(handler).startDocument();
            verify(handler, times(2)).startElement(eq(XHTMLContentHandler.XHTML), eq("p"), eq("p"), any(Attributes.class));
            verify(handler, times(2)).endElement(XHTMLContentHandler.XHTML, "p", "p");
            verify(handler).characters(new String("Test content 1").toCharArray(), 0, 14);
            verify(handler).characters(new String("Test content 2").toCharArray(), 0, 14);
View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
        }
    }

    @Test
    public void testHeaders() {
        Parser parser = new MboxParser();
        Metadata metadata = new Metadata();
        InputStream stream = getStream("test-documents/headers.mbox");
        ContentHandler handler = mock(DefaultHandler.class);

        try {
            parser.parse(stream, handler, metadata, new ParseContext());

            verify(handler).startDocument();
            verify(handler).startElement(eq(XHTMLContentHandler.XHTML), eq("p"), eq("p"), any(Attributes.class));
            verify(handler).characters(new String("Test content").toCharArray(), 0, 12);
            verify(handler).endDocument();
View Full Code Here
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
        }
    }

    @Test
    public void testMultilineHeader() {
        Parser parser = new MboxParser();
        Metadata metadata = new Metadata();
        InputStream stream = getStream("test-documents/multiline.mbox");
        ContentHandler handler = mock(DefaultHandler.class);

        try {
            parser.parse(stream, handler, metadata, new ParseContext());

            verify(handler).startDocument();
            verify(handler).startElement(eq(XHTMLContentHandler.XHTML), eq("p"), eq("p"), any(Attributes.class));
            verify(handler).characters(new String("Test content").toCharArray(), 0, 12);
            verify(handler).endDocument();
View Full Code Here
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
        }
    }

    @Test
    public void testQuoted() {
        Parser parser = new MboxParser();
        Metadata metadata = new Metadata();
        InputStream stream = getStream("test-documents/quoted.mbox");
        ContentHandler handler = mock(DefaultHandler.class);

        try {
            parser.parse(stream, handler, metadata, new ParseContext());

            verify(handler).startDocument();
            verify(handler).startElement(eq(XHTMLContentHandler.XHTML), eq("p"), eq("p"), any(Attributes.class));
            verify(handler).startElement(eq(XHTMLContentHandler.XHTML), eq("q"), eq("q"), any(Attributes.class));
            verify(handler).endElement(eq(XHTMLContentHandler.XHTML), eq("q"), eq("q"));
View Full Code Here
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
        }
    }

    @Test
    public void testComplex() {
        Parser parser = new MboxParser();
        Metadata metadata = new Metadata();
        InputStream stream = getStream("test-documents/complex.mbox");
        ContentHandler handler = mock(DefaultHandler.class);

        try {
            parser.parse(stream, handler, metadata, new ParseContext());

            // TODO: Remove subject and author in Tika 2.0
            assertEquals("Re: question about when shuffle/sort start working", metadata.get(Metadata.SUBJECT));
            assertEquals("Re: question about when shuffle/sort start working", metadata.get(TikaCoreProperties.TITLE));
            assertEquals("Jothi Padmanabhan <jothipn@yahoo-inc.com>", metadata.get(Metadata.AUTHOR));
View Full Code Here
TOP

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

  • bixo.parser.TikaCallableTest
  • com.baasbox.controllers.File
  • com.cloudera.cdk.morphline.solrcell.SolrCellBuilder$SolrCell
  • com.digitalpebble.behemoth.tika.TikaProcessor
  • com.findwise.hydra.stage.tika.SimpleFetchingTikaStageTest
  • com.ikanow.infinit.e.harvest.extraction.document.file.FileHarvester
  • com.ikanow.infinit.e.harvest.extraction.text.legacy.TextExtractorTika
  • com.nidhinova.tika.server.TikaService
  • com.tamingtext.tika.TikaTest
  • ipcress.Extraction
    • 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.