Examples with RichTextContentHandler org.apache.tika.sax.RichTextContentHandler used on opensource projects

Example 1 with RichTextContentHandler

use of org.apache.tika.sax.RichTextContentHandler in project tika by apache.

the class UnpackerResource method process.

private Map<String, byte[]> process(InputStream is, @Context HttpHeaders httpHeaders, @Context UriInfo info, boolean saveAll) throws Exception {
    Metadata metadata = new Metadata();
    ParseContext pc = new ParseContext();
    Parser parser = TikaResource.createParser();
    if (parser instanceof DigestingParser) {
        //no need to digest for unwrapping
        parser = ((DigestingParser) parser).getWrappedParser();
    }
    TikaResource.fillMetadata(parser, metadata, pc, httpHeaders.getRequestHeaders());
    TikaResource.logRequest(LOG, info, metadata);
    ContentHandler ch;
    ByteArrayOutputStream text = new ByteArrayOutputStream();
    if (saveAll) {
        ch = new BodyContentHandler(new RichTextContentHandler(new OutputStreamWriter(text, UTF_8)));
    } else {
        ch = new DefaultHandler();
    }
    Map<String, byte[]> files = new HashMap<>();
    MutableInt count = new MutableInt();
    pc.set(EmbeddedDocumentExtractor.class, new MyEmbeddedDocumentExtractor(count, files));
    TikaResource.parse(parser, LOG, info.getPath(), is, ch, metadata, pc);
    if (count.intValue() == 0 && !saveAll) {
        throw new WebApplicationException(Response.Status.NO_CONTENT);
    }
    if (saveAll) {
        files.put(TEXT_FILENAME, text.toByteArray());
        ByteArrayOutputStream metaStream = new ByteArrayOutputStream();
        metadataToCsv(metadata, metaStream);
        files.put(META_FILENAME, metaStream.toByteArray());
    }
    return files;
}

Example 2 with RichTextContentHandler

use of org.apache.tika.sax.RichTextContentHandler in project tika by apache.

the class TikaResource method produceText.

public StreamingOutput produceText(final InputStream is, MultivaluedMap<String, String> httpHeaders, final UriInfo info) {
    final Parser parser = createParser();
    final Metadata metadata = new Metadata();
    final ParseContext context = new ParseContext();
    fillMetadata(parser, metadata, context, httpHeaders);
    fillParseContext(context, httpHeaders, parser);
    logRequest(LOG, info, metadata);
    return new StreamingOutput() {

        public void write(OutputStream outputStream) throws IOException, WebApplicationException {
            Writer writer = new OutputStreamWriter(outputStream, UTF_8);
            BodyContentHandler body = new BodyContentHandler(new RichTextContentHandler(writer));
            parse(parser, LOG, info.getPath(), is, body, metadata, context);
        }
    };
}