Examples with EncryptedDocumentException org.apache.tika.exception.EncryptedDocumentException used on opensource projects

Example 1 with EncryptedDocumentException

use of org.apache.tika.exception.EncryptedDocumentException in project tika by apache.

the class SXWPFExtractorTest method testEncrypted.

@Test
public void testEncrypted() throws Exception {
    Map<String, String> tests = new HashMap<String, String>();
    tests.put("testWORD_protected_passtika.docx", "This is an encrypted Word 2007 File");
    Parser parser = new AutoDetectParser();
    Metadata m = new Metadata();
    PasswordProvider passwordProvider = new PasswordProvider() {

        @Override
        public String getPassword(Metadata metadata) {
            return "tika";
        }
    };
    OfficeParserConfig opc = new OfficeParserConfig();
    opc.setUseSAXDocxExtractor(true);
    ParseContext passwordContext = new ParseContext();
    passwordContext.set(org.apache.tika.parser.PasswordProvider.class, passwordProvider);
    passwordContext.set(OfficeParserConfig.class, opc);
    for (Map.Entry<String, String> e : tests.entrySet()) {
        assertContains(e.getValue(), getXML(e.getKey(), passwordContext).xml);
    }
    //now try with no password
    for (Map.Entry<String, String> e : tests.entrySet()) {
        boolean exc = false;
        try {
            getXML(e.getKey(), parseContext);
        } catch (EncryptedDocumentException ex) {
            exc = true;
        }
        assertTrue(exc);
    }
}

Example 2 with EncryptedDocumentException

use of org.apache.tika.exception.EncryptedDocumentException in project tika by apache.

the class OOXMLParserTest method testEncrypted.

@Test
public void testEncrypted() throws Exception {
    Map<String, String> tests = new HashMap<String, String>();
    tests.put("testWORD_protected_passtika.docx", "This is an encrypted Word 2007 File");
    tests.put("testPPT_protected_passtika.pptx", "This is an encrypted PowerPoint 2007 slide.");
    tests.put("testEXCEL_protected_passtika.xlsx", "This is an Encrypted Excel spreadsheet.");
    Parser parser = new AutoDetectParser();
    Metadata m = new Metadata();
    PasswordProvider passwordProvider = new PasswordProvider() {

        @Override
        public String getPassword(Metadata metadata) {
            return "tika";
        }
    };
    ParseContext passwordContext = new ParseContext();
    passwordContext.set(org.apache.tika.parser.PasswordProvider.class, passwordProvider);
    for (Map.Entry<String, String> e : tests.entrySet()) {
        try (InputStream is = getTestDocument(e.getKey())) {
            ContentHandler handler = new BodyContentHandler();
            parser.parse(is, handler, m, passwordContext);
            assertContains(e.getValue(), handler.toString());
        }
    }
    ParseContext context = new ParseContext();
    //now try with no password
    for (Map.Entry<String, String> e : tests.entrySet()) {
        boolean exc = false;
        try (InputStream is = getTestDocument(e.getKey())) {
            ContentHandler handler = new BodyContentHandler();
            parser.parse(is, handler, m, context);
        } catch (EncryptedDocumentException ex) {
            exc = true;
        }
        assertTrue(exc);
    }
}

Example 3 with EncryptedDocumentException

use of org.apache.tika.exception.EncryptedDocumentException in project tika by apache.

the class ExcelParserTest method testExcelParserPassword.

@Test
public void testExcelParserPassword() throws Exception {
    try (InputStream input = ExcelParserTest.class.getResourceAsStream("/test-documents/testEXCEL_protected_passtika.xls")) {
        Metadata metadata = new Metadata();
        ContentHandler handler = new BodyContentHandler();
        ParseContext context = new ParseContext();
        context.set(Locale.class, Locale.US);
        new OfficeParser().parse(input, handler, metadata, context);
        fail("Document is encrypted, shouldn't parse");
    } catch (EncryptedDocumentException e) {
    // Good
    }
    // Try again, this time with the password
    try (InputStream input = ExcelParserTest.class.getResourceAsStream("/test-documents/testEXCEL_protected_passtika.xls")) {
        Metadata metadata = new Metadata();
        ContentHandler handler = new BodyContentHandler();
        ParseContext context = new ParseContext();
        context.set(Locale.class, Locale.US);
        context.set(PasswordProvider.class, new PasswordProvider() {

            @Override
            public String getPassword(Metadata metadata) {
                return "tika";
            }
        });
        new OfficeParser().parse(input, handler, metadata, context);
        assertEquals("application/vnd.ms-excel", metadata.get(Metadata.CONTENT_TYPE));
        assertEquals(null, metadata.get(TikaCoreProperties.TITLE));
        assertEquals("Antoni", metadata.get(TikaCoreProperties.CREATOR));
        assertEquals("2011-11-25T09:52:48Z", metadata.get(TikaCoreProperties.CREATED));
        String content = handler.toString();
        assertContains("This is an Encrypted Excel spreadsheet", content);
        assertNotContained("9.0", content);
    }
}

Example 4 with EncryptedDocumentException

use of org.apache.tika.exception.EncryptedDocumentException in project tika by apache.

the class JackcessParser method parse.

@Override
public void parse(InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context) throws IOException, SAXException, TikaException {
    TikaInputStream tis = TikaInputStream.get(stream);
    Database db = null;
    XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
    xhtml.startDocument();
    String password = null;
    PasswordProvider passwordProvider = context.get(PasswordProvider.class);
    if (passwordProvider != null) {
        password = passwordProvider.getPassword(metadata);
    }
    try {
        if (password == null) {
            //do this to ensure encryption/wrong password exception vs. more generic
            //"need right codec" error message.
            db = new DatabaseBuilder(tis.getFile()).setCodecProvider(new CryptCodecProvider()).setReadOnly(true).open();
        } else {
            db = new DatabaseBuilder(tis.getFile()).setCodecProvider(new CryptCodecProvider(password)).setReadOnly(true).open();
        }
        //just in case
        db.setLinkResolver(IGNORE_LINK_RESOLVER);
        JackcessExtractor ex = new JackcessExtractor(metadata, context, locale);
        ex.parse(db, xhtml);
    } catch (IllegalStateException e) {
        if (e.getMessage() != null && e.getMessage().contains("Incorrect password")) {
            throw new EncryptedDocumentException(e);
        }
        throw e;
    } finally {
        if (db != null) {
            try {
                db.close();
            } catch (IOException e) {
            //swallow = silent close
            }
        }
    }
    xhtml.endDocument();
}

Example 5 with EncryptedDocumentException

use of org.apache.tika.exception.EncryptedDocumentException in project tika by apache.

the class RarParser method parse.

@Override
public void parse(InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context) throws IOException, SAXException, TikaException {
    XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
    xhtml.startDocument();
    EmbeddedDocumentExtractor extractor = EmbeddedDocumentUtil.getEmbeddedDocumentExtractor(context);
    Archive rar = null;
    try (TemporaryResources tmp = new TemporaryResources()) {
        TikaInputStream tis = TikaInputStream.get(stream, tmp);
        rar = new Archive(tis.getFile());
        if (rar.isEncrypted()) {
            throw new EncryptedDocumentException();
        }
        //Without this BodyContentHandler does not work
        xhtml.element("div", " ");
        FileHeader header = rar.nextFileHeader();
        while (header != null && !Thread.currentThread().isInterrupted()) {
            if (!header.isDirectory()) {
                try (InputStream subFile = rar.getInputStream(header)) {
                    Metadata entrydata = PackageParser.handleEntryMetadata("".equals(header.getFileNameW()) ? header.getFileNameString() : header.getFileNameW(), header.getCTime(), header.getMTime(), header.getFullUnpackSize(), xhtml);
                    if (extractor.shouldParseEmbedded(entrydata)) {
                        extractor.parseEmbedded(subFile, handler, entrydata, true);
                    }
                }
            }
            header = rar.nextFileHeader();
        }
    } catch (RarException e) {
        throw new TikaException("RarParser Exception", e);
    } finally {
        if (rar != null)
            rar.close();
    }
    xhtml.endDocument();
}