import org.apache.tika.parser.pdf.PDFParser; //导入依赖的package包/类
@Override
public String fromFile(File file) {
	String resultText = "";

	try {
		FileInputStream inputstream = new FileInputStream(file);
		BodyContentHandler handler = new BodyContentHandler(-1);
		Metadata metadata = new Metadata();
		ParseContext pcontext = new ParseContext();

		PDFParserConfig config = new PDFParserConfig();
		config.setSortByPosition(true);

		PDFParser pdfparser = new PDFParser();
		pdfparser.setPDFParserConfig(config);
		
		System.out.println("Parsing PDF to TEXT...");
		
		pdfparser.parse(inputstream, handler, metadata, pcontext);
		resultText = handler.toString();
		
		System.out.println("Parsing complete");
	} catch (Exception ex) {
		throw new RuntimeException(ex);
	}

	return resultText;
}
 

import org.apache.tika.parser.pdf.PDFParser; //导入依赖的package包/类
private static String parse(final InputStream input) throws TikaException, SAXException, IOException {
    final Parser parser = new PDFParser();
    final ContentHandler handler = new BodyContentHandler();
    final Metadata metadata = new Metadata();
    final ParseContext parseContext = new ParseContext();

    parser.parse(input, handler, metadata, parseContext);

    return handler.toString();
}
 

import org.apache.tika.parser.pdf.PDFParser; //导入依赖的package包/类
@POST
@Path("import")
@Consumes(MediaType.MULTIPART_FORM_DATA)
@Produces(MediaType.APPLICATION_JSON)
public Response importData(@DefaultValue("true") @FormDataParam("enabled") boolean enabled,
		@FormDataParam("file") InputStream inputStream,
		@FormDataParam("file") FormDataContentDisposition fileDetail, @Context UriInfo uriInfo)
		throws JSONException, IOException, SAXException, TikaException {

	BodyContentHandler handler = new BodyContentHandler(-1);
	Metadata metadata = new Metadata();

	ParseContext pcontext = new ParseContext();

	// parsing the document using PDF parser
	PDFParser pdfparser = new PDFParser();
	pdfparser.parse(inputStream, handler, metadata, pcontext);


	String sentences[] = turOpenNLPConnector.sentenceDetect(handler.toString());

	TurData turData = new TurData();

	turData.setName(fileDetail.getFileName());
	turData.setType(FilenameUtils.getExtension(fileDetail.getFileName()));
	this.turDataRepository.save(turData);

	for (String sentence : sentences) {
		TurDataGroupSentence turDataGroupSentence = new TurDataGroupSentence();
		turDataGroupSentence.setTurData(turData);
		turDataGroupSentence.setSentence(sentence);
		this.turDataGroupSentenceRepository.save(turDataGroupSentence);
	}
	
	JSONObject jsonTraining = new JSONObject();
	jsonTraining.put("sentences", sentences);
	return Response.status(200).entity(jsonTraining.toString()).build();
}
 

import org.apache.tika.parser.pdf.PDFParser; //导入依赖的package包/类
@Override
protected Parser getParser() {
   return new PDFParser();
}
 

import org.apache.tika.parser.pdf.PDFParser; //导入依赖的package包/类
@Override
protected Parser getParser() 
{
   return new PDFParser();
}
 

import org.apache.tika.parser.pdf.PDFParser; //导入依赖的package包/类
@POST
@Path("import")
@Consumes(MediaType.MULTIPART_FORM_DATA)
@Produces(MediaType.APPLICATION_JSON)
public TurDataGroupData importData(@PathParam("dataGroupId") int dataGroupId,
		@DefaultValue("true") @FormDataParam("enabled") boolean enabled,
		@FormDataParam("file") InputStream inputStream,
		@FormDataParam("file") FormDataContentDisposition fileDetail, @Context UriInfo uriInfo)
		throws JSONException, IOException, SAXException, TikaException {

	TurDataGroup turDataGroup = this.turDataGroupRepository.getOne(dataGroupId);
	BodyContentHandler handler = new BodyContentHandler(-1);
	Metadata metadata = new Metadata();

	ParseContext pcontext = new ParseContext();

	// parsing the document using PDF parser
	PDFParser pdfparser = new PDFParser();
	pdfparser.parse(inputStream, handler, metadata, pcontext);

	String sentences[] = turOpenNLPConnector.sentenceDetect(handler.toString());

	TurData turData = new TurData();

	turData.setName(fileDetail.getFileName());
	turData.setType(FilenameUtils.getExtension(fileDetail.getFileName()));
	this.turDataRepository.save(turData);

	for (String sentence : sentences) {
		TurDataGroupSentence turDataGroupSentence = new TurDataGroupSentence();
		turDataGroupSentence.setTurData(turData);
		turDataGroupSentence.setSentence(sentence);
		turDataGroupSentence.setTurDataGroup(turDataGroup);
		turDataGroupSentenceRepository.save(turDataGroupSentence);
	}

	TurDataGroupData turDataGroupData = new TurDataGroupData();

	turDataGroupData.setTurData(turData);
	turDataGroupData.setTurDataGroup(turDataGroup);
	turDataGroupDataRepository.save(turDataGroupData);

	return turDataGroupData;
}