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

Example 86 with TikaException

use of org.apache.tika.exception.TikaException in project stanbol by apache.

the class TikaEngine method computeEnhancements.

@Override
public void computeEnhancements(ContentItem ci) throws EngineException {
    MediaTypeAndStream mtas = extractMediaType(ci);
    if (mtas.mediaType == null) {
        //unable to parse and detect content type
        return;
    }
    MediaType plainMediaType = mtas.mediaType.getBaseType();
    if (plainMediaType.equals(MediaType.TEXT_PLAIN)) {
        //we need not to process plain text!
        return;
    }
    final ParseContext context = new ParseContext();
    context.set(Parser.class, parser);
    Set<MediaType> supproted = parser.getSupportedTypes(context);
    if (supproted.contains(plainMediaType)) {
        final InputStream in;
        if (mtas.in == null) {
            in = ci.getStream();
        } else {
            in = mtas.in;
        }
        final Metadata metadata = new Metadata();
        //set the already parsed contentType
        metadata.set(Metadata.CONTENT_TYPE, mtas.mediaType.toString());
        //also explicitly set the charset as contentEncoding
        String charset = mtas.mediaType.getParameters().get("charset");
        if (charset != null) {
            metadata.set(Metadata.CONTENT_ENCODING, charset);
        }
        ContentSink plainTextSink;
        try {
            plainTextSink = ciFactory.createContentSink(TEXT_PLAIN + "; charset=" + UTF8.name());
        } catch (IOException e) {
            //close the input stream
            IOUtils.closeQuietly(in);
            throw new EngineException("Error while initialising Blob for" + "writing the text/plain version of the parsed content", e);
        }
        final Writer plainTextWriter = new OutputStreamWriter(plainTextSink.getOutputStream(), UTF8);
        final ContentHandler textHandler = new //only the Body
        BodyContentHandler(//skip ignoreable
        new PlainTextHandler(plainTextWriter, false, skipLinebreaks));
        final ToXMLContentHandler xhtmlHandler;
        final ContentHandler mainHandler;
        ContentSink xhtmlSink = null;
        try {
            if (!plainMediaType.equals(XHTML)) {
                //do not parse XHTML from XHTML
                try {
                    xhtmlSink = ciFactory.createContentSink(XHTML + "; charset=" + UTF8.name());
                } catch (IOException e) {
                    throw new EngineException("Error while initialising Blob for" + "writing the application/xhtml+xml version of the parsed content", e);
                }
                try {
                    xhtmlHandler = new ToXMLContentHandler(xhtmlSink.getOutputStream(), UTF8.name());
                } catch (UnsupportedEncodingException e) {
                    throw new EngineException("This system does not support the encoding " + UTF8, e);
                }
                mainHandler = new MultiHandler(textHandler, xhtmlHandler);
            } else {
                mainHandler = textHandler;
                xhtmlHandler = null;
                xhtmlSink = null;
            }
            try {
                AccessController.doPrivileged(new PrivilegedExceptionAction<Object>() {

                    public Object run() throws IOException, SAXException, TikaException {
                        /* 
                             * We need to replace the context Classloader with the Bundle ClassLoader
                             * to ensure that Singleton instances of XML frameworks (such as node4j) 
                             * do not leak into the OSGI environment.
                             * 
                             * Most Java XML libs prefer to load implementations by using the 
                             * {@link Thread#getContextClassLoader()}. However OSGI has no control over
                             * this {@link ClassLoader}. Because of that there can be situations where
                             * Interfaces are loaded via the Bundle Classloader and the implementations
                             * are taken from the context Classloader. What can cause 
                             * {@link ClassCastException}, {@link ExceptionInInitializerError}s, ...
                             * 
                             * Setting the context Classloader to the Bundle classloader helps to avoid
                             * those situations.
                             */
                        ClassLoader contextClassLoader = updateContextClassLoader();
                        try {
                            parser.parse(in, mainHandler, metadata, context);
                        } finally {
                            //reset the previous context ClassLoader
                            Thread.currentThread().setContextClassLoader(contextClassLoader);
                        }
                        return null;
                    }
                });
            } catch (PrivilegedActionException pae) {
                Exception e = pae.getException();
                if (e instanceof IOException || e instanceof SAXException || e instanceof TikaException) {
                    throw new EngineException("Unable to convert ContentItem " + ci.getUri() + " with mimeType '" + ci.getMimeType() + "' to " + "plain text!", e);
                } else {
                    //runtime exception
                    throw RuntimeException.class.cast(e);
                }
            }
        } finally {
            //ensure that the writers are closed correctly
            IOUtils.closeQuietly(in);
            IOUtils.closeQuietly(plainTextWriter);
            if (xhtmlSink != null) {
                IOUtils.closeQuietly(xhtmlSink.getOutputStream());
            }
        }
        String random = randomUUID().toString();
        IRI textBlobUri = new IRI("urn:tika:text:" + random);
        ci.addPart(textBlobUri, plainTextSink.getBlob());
        if (xhtmlHandler != null) {
            IRI xhtmlBlobUri = new IRI("urn:tika:xhtml:" + random);
            ci.addPart(xhtmlBlobUri, xhtmlSink.getBlob());
        }
        //add the extracted metadata
        if (log.isInfoEnabled()) {
            for (String name : metadata.names()) {
                log.info("{}: {}", name, Arrays.toString(metadata.getValues(name)));
            }
        }
        ci.getLock().writeLock().lock();
        try {
            Graph graph = ci.getMetadata();
            IRI id = ci.getUri();
            Set<String> mapped = ontologyMappings.apply(graph, id, metadata);
            if (includeUnmappedProperties) {
                Set<String> unmapped = new HashSet<String>(Arrays.asList(metadata.names()));
                unmapped.removeAll(mapped);
                for (String name : unmapped) {
                    if (name.indexOf(':') >= 0 || includeAllUnmappedProperties) {
                        //only mapped
                        IRI prop = new IRI(new StringBuilder(TIKA_URN_PREFIX).append(name).toString());
                        for (String value : metadata.getValues(name)) {
                            //TODO: without the Property for the name we have no datatype
                            //      information ... so we add PlainLiterals for now
                            graph.add(new TripleImpl(id, prop, new PlainLiteralImpl(value)));
                        }
                    }
                }
            }
        } finally {
            ci.getLock().writeLock().unlock();
        }
    }
//else not supported format
}

Example 87 with TikaException

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

the class ExternalEmbedder method embed.

/**
     * Executes the configured external command and passes the given document
     * stream as a simple XHTML document to the given SAX content handler.
     * Metadata is only extracted if {@link #setMetadataCommandArguments(Map)}
     * has been called to set arguments.
     */
public void embed(final Metadata metadata, final InputStream inputStream, final OutputStream outputStream, final ParseContext context) throws IOException, TikaException {
    boolean inputToStdIn = true;
    boolean outputFromStdOut = true;
    boolean hasMetadataCommandArguments = (metadataCommandArguments != null && !metadataCommandArguments.isEmpty());
    boolean serializeMetadataCommandArgumentsToken = false;
    boolean replacedMetadataCommandArgumentsToken = false;
    TikaInputStream tikaInputStream = TikaInputStream.get(inputStream);
    File tempOutputFile = null;
    List<String> commandMetadataSegments = null;
    if (hasMetadataCommandArguments) {
        commandMetadataSegments = getCommandMetadataSegments(metadata);
    }
    // Build our command
    List<String> origCmd = Arrays.asList(command);
    List<String> cmd = new ArrayList<String>();
    for (String commandSegment : origCmd) {
        if (commandSegment.indexOf(ExternalParser.INPUT_FILE_TOKEN) != -1) {
            commandSegment = commandSegment.replace(ExternalParser.INPUT_FILE_TOKEN, tikaInputStream.getFile().toString());
            inputToStdIn = false;
        }
        if (commandSegment.indexOf(ExternalParser.OUTPUT_FILE_TOKEN) != -1) {
            tempOutputFile = tmp.createTemporaryFile();
            commandSegment = commandSegment.replace(ExternalParser.OUTPUT_FILE_TOKEN, tempOutputFile.toString());
            outputFromStdOut = false;
        }
        if (commandSegment.indexOf(METADATA_COMMAND_ARGUMENTS_SERIALIZED_TOKEN) != -1) {
            serializeMetadataCommandArgumentsToken = true;
        }
        if (commandSegment.indexOf(METADATA_COMMAND_ARGUMENTS_TOKEN) != -1) {
            if (hasMetadataCommandArguments) {
                for (String commandMetadataSegment : commandMetadataSegments) {
                    cmd.add(commandMetadataSegment);
                }
            }
            replacedMetadataCommandArgumentsToken = true;
        } else {
            cmd.add(commandSegment);
        }
    }
    if (hasMetadataCommandArguments) {
        if (serializeMetadataCommandArgumentsToken) {
            // Find all metadata tokens and replace with encapsulated metadata
            int i = 0;
            for (String commandSegment : cmd) {
                if (commandSegment.indexOf(METADATA_COMMAND_ARGUMENTS_SERIALIZED_TOKEN) != -1) {
                    commandSegment = commandSegment.replace(METADATA_COMMAND_ARGUMENTS_SERIALIZED_TOKEN, serializeMetadata(commandMetadataSegments));
                    cmd.set(i, commandSegment);
                }
                i++;
            }
        } else if (!replacedMetadataCommandArgumentsToken && !serializeMetadataCommandArgumentsToken) {
            // Tack metadata onto the end of the cmd as arguments
            cmd.addAll(commandMetadataSegments);
        }
    }
    // Execute
    Process process;
    if (cmd.toArray().length == 1) {
        process = Runtime.getRuntime().exec(cmd.toArray(new String[] {})[0]);
    } else {
        process = Runtime.getRuntime().exec(cmd.toArray(new String[] {}));
    }
    ByteArrayOutputStream stdErrOutputStream = new ByteArrayOutputStream();
    try {
        sendStdErrToOutputStream(process, stdErrOutputStream);
        if (inputToStdIn) {
            sendInputStreamToStdIn(inputStream, process);
        } else {
            // We're not writing to std in this case so close
            process.getOutputStream().close();
        }
        if (outputFromStdOut) {
            sendStdOutToOutputStream(process, outputStream);
        } else {
            tmp.dispose();
            try {
                process.waitFor();
            } catch (InterruptedException ignore) {
            }
            // The command is finished, read the output file into the given output stream
            InputStream tempOutputFileInputStream = TikaInputStream.get(tempOutputFile);
            IOUtils.copy(tempOutputFileInputStream, outputStream);
        }
    } finally {
        if (outputFromStdOut) {
            try {
                process.waitFor();
            } catch (InterruptedException ignore) {
            }
        } else {
            try {
                // Clean up temp output files
                tempOutputFile.delete();
            } catch (Exception e) {
            }
        }
        if (!inputToStdIn) {
            // Close input file (and delete if created by up TemporaryResources.createTemporaryFile) 
            IOUtils.closeQuietly(tikaInputStream);
        }
        IOUtils.closeQuietly(outputStream);
        IOUtils.closeQuietly(stdErrOutputStream);
        if (process.exitValue() != 0) {
            throw new TikaException("There was an error executing the command line" + "\nExecutable Command:\n\n" + cmd + "\nExecutable Error:\n\n" + stdErrOutputStream.toString(UTF_8.name()));
        }
    }
}

Example 88 with TikaException

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

the class ParserContainerExtractor method extract.

public void extract(TikaInputStream stream, ContainerExtractor recurseExtractor, EmbeddedResourceHandler handler) throws IOException, TikaException {
    ParseContext context = new ParseContext();
    context.set(Parser.class, new RecursiveParser(recurseExtractor, handler));
    try {
        parser.parse(stream, new DefaultHandler(), new Metadata(), context);
    } catch (SAXException e) {
        throw new TikaException("Unexpected SAX exception", e);
    }
}

Example 89 with TikaException

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

the class ParserDecorator method withFallbacks.

/**
     * Decorates the given parsers into a virtual parser, where they'll
     *  be tried in preference order until one works without error.
     * TODO Is this the right name?
     * TODO Is this the right place to put this? Should it be in CompositeParser? Elsewhere?
     * TODO Should we reset the Metadata if we try another parser?
     * TODO Should we reset the ContentHandler if we try another parser?
     * TODO Should we log/report failures anywhere?
     * @deprecated Do not use until the TODOs are resolved, see TIKA-1509
     */
public static final Parser withFallbacks(final Collection<? extends Parser> parsers, final Set<MediaType> types) {
    Parser parser = EmptyParser.INSTANCE;
    if (!parsers.isEmpty())
        parser = parsers.iterator().next();
    return new ParserDecorator(parser) {

        private static final long serialVersionUID = 1625187131782069683L;

        @Override
        public Set<MediaType> getSupportedTypes(ParseContext context) {
            return types;
        }

        @Override
        public void parse(InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context) throws IOException, SAXException, TikaException {
            // Must have a TikaInputStream, so we can re-use it if parsing fails
            // Need to close internally created tstream to release resources
            TemporaryResources tmp = (TikaInputStream.isTikaInputStream(stream)) ? null : new TemporaryResources();
            try {
                TikaInputStream tstream = TikaInputStream.get(stream, tmp);
                tstream.getFile();
                // Try each parser in turn
                for (Parser p : parsers) {
                    tstream.mark(-1);
                    try {
                        p.parse(tstream, handler, metadata, context);
                        return;
                    } catch (Exception e) {
                    // TODO How to log / record this failure?
                    }
                    // Prepare for the next parser, if present
                    tstream.reset();
                }
            } finally {
                if (tmp != null) {
                    tmp.dispose();
                }
            }
        }

        @Override
        public String getDecorationName() {
            return "With Fallback";
        }
    };
}

Example 90 with TikaException

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

the class CompositeParser method parse.

/**
     * Delegates the call to the matching component parser.
     * <p>
     * Potential {@link RuntimeException}s, {@link IOException}s and
     * {@link SAXException}s unrelated to the given input stream and content
     * handler are automatically wrapped into {@link TikaException}s to better
     * honor the {@link Parser} contract.
     */
public void parse(InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context) throws IOException, SAXException, TikaException {
    Parser parser = getParser(metadata, context);
    TemporaryResources tmp = new TemporaryResources();
    try {
        TikaInputStream taggedStream = TikaInputStream.get(stream, tmp);
        TaggedContentHandler taggedHandler = handler != null ? new TaggedContentHandler(handler) : null;
        if (parser instanceof ParserDecorator) {
            metadata.add("X-Parsed-By", ((ParserDecorator) parser).getWrappedParser().getClass().getName());
        } else {
            metadata.add("X-Parsed-By", parser.getClass().getName());
        }
        try {
            parser.parse(taggedStream, taggedHandler, metadata, context);
        } catch (RuntimeException e) {
            throw new TikaException("Unexpected RuntimeException from " + parser, e);
        } catch (IOException e) {
            taggedStream.throwIfCauseOf(e);
            throw new TikaException("TIKA-198: Illegal IOException from " + parser, e);
        } catch (SAXException e) {
            if (taggedHandler != null)
                taggedHandler.throwIfCauseOf(e);
            throw new TikaException("TIKA-237: Illegal SAXException from " + parser, e);
        }
    } finally {
        tmp.dispose();
    }
}