Examples of org.apache.tika.sax.BodyContentHandler

org.apache.tika.sax.BodyContentHandler
Content handler decorator that only passes everything inside the XHTML <body/> tag to the underlying handler. Note that the <body/> tag itself is not passed on.

    @Test
    public void testProtectedExcelFile() throws Exception {


        Parser parser = new AutoDetectParser();
        Metadata metadata = new Metadata();
        ContentHandler handler = new BodyContentHandler();
        ParseContext context = new ParseContext();


        InputStream input = getTestDocument("protectedFile.xlsx");
        try {
            parser.parse(input, handler, metadata, context);


            assertEquals(
                    "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
                    metadata.get(Metadata.CONTENT_TYPE));


            assertEquals("true", metadata.get(TikaMetadataKeys.PROTECTED));
            
            String content = handler.toString();
            assertTrue(content.contains("Office"));
        } finally {
            input.close();
        }
    }

View Full Code Here

     */
    @Test
    public void testNullHeaders() throws Exception {
        Parser parser = new AutoDetectParser();
        Metadata metadata = new Metadata();
        ContentHandler handler = new BodyContentHandler();
        ParseContext context = new ParseContext();


        InputStream input = getTestDocument("NullHeader.docx");
        try {
            parser.parse(input, handler, metadata, context);
            assertFalse(handler.toString().length()==0);
        } finally {
            input.close();
        }
    }

View Full Code Here

        }
    }


    @Test
    public void testVarious() throws Exception {
        ContentHandler handler = new BodyContentHandler();
        Metadata metadata = new Metadata();


        InputStream stream = OOXMLParserTest.class.getResourceAsStream(
                "/test-documents/testWORD_various.docx");
        try {
            new AutoDetectParser().parse(stream, handler, metadata, new ParseContext());
        } finally {
            stream.close();
        }


        String content = handler.toString();
        //content = content.replaceAll("\\s+"," ");
        assertContains("Footnote appears here", content);
        assertContains("This is a footnote.", content);
        assertContains("This is the header text.", content);
        assertContains("This is the footer text.", content);

View Full Code Here

        assertContains("\uD800\uDF32\uD800\uDF3f\uD800\uDF44\uD800\uDF39\uD800\uDF43\uD800\uDF3A", content);
    }


    @Test
    public void testVariousPPTX() throws Exception {
        ContentHandler handler = new BodyContentHandler();
        Metadata metadata = new Metadata();


        InputStream stream = OOXMLParserTest.class.getResourceAsStream(
                "/test-documents/testPPT_various.pptx");
        try {
            new AutoDetectParser().parse(stream, handler, metadata, new ParseContext());
        } finally {
            stream.close();
        }


        String content = handler.toString();
        //content = content.replaceAll("\\s+"," ");
        assertContains("Footnote appears here", content);
        assertContains("This is a footnote.", content);
        assertContains("This is the header text.", content);
        assertContains("This is the footer text.", content);

View Full Code Here

    @Test
    public void testWordParser() throws Exception {
        InputStream input = WordParserTest.class.getResourceAsStream(
                "/test-documents/testWORD.doc");
        try {
            ContentHandler handler = new BodyContentHandler();
            Metadata metadata = new Metadata();
            new OfficeParser().parse(input, handler, metadata, new ParseContext());


            assertEquals(
                    "application/msword",
                    metadata.get(Metadata.CONTENT_TYPE));
            assertEquals("Sample Word Document", metadata.get(TikaCoreProperties.TITLE));
            assertEquals("Keith Bennett", metadata.get(TikaCoreProperties.CREATOR));
            assertEquals("Keith Bennett", metadata.get(Metadata.AUTHOR));
            assertTrue(handler.toString().contains("Sample Word Document"));
        } finally {
            input.close();
        }
    }

View Full Code Here

        assertContains("\uD800\uDF32\uD800\uDF3f\uD800\uDF44\uD800\uDF39\uD800\uDF43\uD800\uDF3A", content);
    }


    @Test
    public void testMasterFooter() throws Exception {
        ContentHandler handler = new BodyContentHandler();
        Metadata metadata = new Metadata();


        InputStream stream = OOXMLParserTest.class.getResourceAsStream(
                "/test-documents/testPPT_masterFooter.pptx");
        try {
            new AutoDetectParser().parse(stream, handler, metadata, new ParseContext());
        } finally {
            stream.close();
        }


        String content = handler.toString();
        assertContains("Master footer is here", content);
    }

View Full Code Here

    @Test
    public void testWordWithWAV() throws Exception {
        InputStream input = WordParserTest.class.getResourceAsStream(
                "/test-documents/Doc1_ole.doc");
        try {
            ContentHandler handler = new BodyContentHandler();
            Metadata metadata = new Metadata();
            new OfficeParser().parse(input, handler, metadata, new ParseContext());


            assertTrue(handler.toString().contains("MSj00974840000[1].wav"));
        } finally {
            input.close();
        }
    }

View Full Code Here

    }
    */


    @Test
    public void testWordArt() throws Exception {
        ContentHandler handler = new BodyContentHandler();
        Metadata metadata = new Metadata();


        InputStream stream = OOXMLParserTest.class.getResourceAsStream(
                "/test-documents/testWordArt.pptx");
        try {
            new AutoDetectParser().parse(stream, handler, metadata, new ParseContext());
        } finally {
            stream.close();
        }
        String content = handler.toString();
        assertContains("Here is some red word Art", content);
    }

View Full Code Here

       InputStream input = OOXMLParserTest.class.getResourceAsStream(
             "/test-documents/testEXCEL_custom_props.xlsx");
       Metadata metadata = new Metadata();
       
       try {
          ContentHandler handler = new BodyContentHandler(-1);
          ParseContext context = new ParseContext();
          context.set(Locale.class, Locale.US);
          new OOXMLParser().parse(input, handler, metadata, context);
       } finally {
          input.close();

View Full Code Here

    @Test
    public void testWord6Parser() throws Exception {
        InputStream input = WordParserTest.class.getResourceAsStream(
                "/test-documents/testWORD6.doc");
        try {
            ContentHandler handler = new BodyContentHandler();
            Metadata metadata = new Metadata();
            new OfficeParser().parse(input, handler, metadata, new ParseContext());


            assertEquals(
                    "application/msword",
                    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(OfficeOpenXMLCore.SUBJECT));
            assertEquals("Gym class featuring a brown fox and lazy dog", metadata.get(Metadata.SUBJECT));
            assertEquals("Nevin Nollop", metadata.get(TikaCoreProperties.CREATOR));
            assertEquals("Nevin Nollop", metadata.get(Metadata.AUTHOR));
            assertTrue(handler.toString().contains("The quick brown fox jumps over the lazy dog"));
        } finally {
            input.close();
        }
    }

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of org.apache.tika.sax.BodyContentHandler

com.nidhinova.tika.server.TikaService

com.tamingtext.tika.TikaTest

cx.fbn.nevernote.threads.IndexRunner

org.apache.droids.tika.TikaDocumentParser

org.apache.oodt.cas.protocol.imaps.ImapsProtocol

org.apache.stanbol.enhancer.engines.tika.TikaEngine

org.apache.tika.bundle.BundleIT

org.apache.tika.embedder.ExternalEmbedderTest

org.apache.tika.example.ContentHandlerExample

org.apache.tika.extractor.ParsingEmbeddedDocumentExtractor

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.