Java Tika类代码示例

import org.apache.tika.Tika; //导入依赖的package包/类
/**
 * Creates a new {@link RepositoryLocationChooser} that allows to specify a media or MIME type
 * for the given data source.
 *
 * @param source
 *            the data source
 * @param initialDestination
 *            the initial location (optional)
 */
public BinaryImportDestinationChooser(BinaryDataSource source, String initialDestination) {
	super(null, null, initialDestination, true, false, true, true, Colors.WHITE);

	// Use generic mime type as default and try to guess more specific.
	String type = "application/octet-stream";
	Tika tika = new Tika();
	try {
		type = tika.detect(source.getLocation());
	} catch (IOException e) {
		// ignore
	}

	JLabel mediaTypelabel = new ResourceLabel("repository_chooser.mime_type");
	mediaType = new JTextArea(type);

	GridBagConstraints c = new GridBagConstraints();
	c.insets = new Insets(ButtonDialog.GAP, 0, 0, ButtonDialog.GAP);
	c.gridwidth = GridBagConstraints.RELATIVE;
	add(mediaTypelabel, c);

	c.insets = new Insets(ButtonDialog.GAP, 0, 0, 0);
	c.gridwidth = GridBagConstraints.REMAINDER;
	c.fill = GridBagConstraints.HORIZONTAL;
	add(mediaType, c);
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
@Before
public void setUp() throws TikaException, IOException, SAXException {
    VelocityEngine engine = new VelocityEngine();
    engine.setProperty(RuntimeConstants.RESOURCE_LOADER, "classpath");
    engine.setProperty("classpath.resource.loader.class", ClasspathResourceLoader.class.getName());
    engine.init();

    Templater templater = new Templater();
    templater.setEngine(engine);

    exporter = new HtmlExporter();
    exporter.setTemplater(templater);

    TikaProvider provider = new TikaProvider();
    Tika tika = provider.tika();

    transformer = new TikaTransformer();
    transformer.setTika(tika);
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
@Before
public void setUp() throws TikaException, IOException, SAXException {
    VelocityEngine engine = new VelocityEngine();
    engine.setProperty(RuntimeConstants.RESOURCE_LOADER, "classpath");
    engine.setProperty("classpath.resource.loader.class", ClasspathResourceLoader.class.getName());
    engine.init();

    Templater templater = new Templater();
    templater.setEngine(engine);

    exporter = new PdfExporter();
    exporter.setTemplater(templater);

    TikaProvider provider = new TikaProvider();
    Tika tika = provider.tika();

    transformer = new TikaTransformer();
    transformer.setTika(tika);
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
/**
 * 利用 Tika 分析 Mime Type
 * 因为 Tika 要解析 File 、 URL 数据流，所以解析需要一定时间。不要用解析扩展名的方法，无法动态判断，不准。
 * <p>
 * Parses the given file and returns the extracted text content.
 *
 * @param file
 * @return
 */
public static String detect(File file) throws Exception {

    //文件不存在
    if (!file.exists()) {
        throw new Exception("exception ! " + file.getAbsoluteFile() + " not existes.");
    }
    Tika t = new Tika();
    return t.detect(file);

}
 

import org.apache.tika.Tika; //导入依赖的package包/类
@Test
@DirtiesContext
public void failedMimetypeDetectionShouldWork() throws Exception {
    final Reflect controllerReflect = Reflect.on(this.controller);
    // Much more evil isn't possible, i guess... DirtiesContext!!!!
    Tika tika = controllerReflect.field("tika").get();
    tika = spy(tika);
    when(tika.detect(any(InputStream.class), any(String.class))).thenThrow(IOException.class);
    controllerReflect.set("tika", tika);

    final MockMultipartFile multipartFile = new MockMultipartFile("assetData", "asset.png", null, this.getClass().getResourceAsStream("/eu/euregjug/site/assets/asset.png"));
    when(this.gridFsTemplate.findOne(any(Query.class))).thenReturn(null);

    mvc
            .perform(
                    fileUpload("/api/assets")
                    .file(multipartFile)
            )
            .andExpect(status().isCreated())
            .andExpect(content().string("asset.png"));

    verify(this.gridFsTemplate).findOne(any(Query.class));
    verify(this.gridFsTemplate).store(any(InputStream.class), eq("asset.png"), isNull(String.class));
    verifyNoMoreInteractions(this.gridFsTemplate);
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
/**
  * return the mime type of a file, dont check extension
  * @param barr
  * @return mime type of the file
  * @throws IOException 
  */
 public static String getMimeType(byte[] barr, String defaultValue) {
     
 	//String mt = getMimeType(new ByteArrayInputStream(barr), null);
 	//if(!StringUtil.isEmpty(mt,true)) return mt;
 	
 	PrintStream out = System.out;
     try {
     	Tika tika = new Tika();
     	return tika.detect(barr);
     } 
     catch(Throwable t) {
     	ExceptionUtil.rethrowIfNecessary(t);
return defaultValue;
     }
 }
 

import org.apache.tika.Tika; //导入依赖的package包/类
@PostConstruct
public void initialise() {
    // find all kiwi-module.properties and check whether they contain a baseurl property to map module web
    // resources to a certain path prefix; if yes, store the prefix and jar URL in the map for lookup and
    // resource resolving by the filter
    this.resourceMap = new HashMap<>();

    for(String module : moduleService.listModules()) {
        ModuleConfiguration config = moduleService.getModuleConfiguration(module);
        if(config.getConfiguration().containsKey("baseurl")) {
            String path = config.getConfiguration().getString("baseurl");
            resourceMap.put(path.startsWith("/")?path:"/"+path,moduleService.getModuleJar(module).toString());
        }
    }

    this.tika = new Tika();
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
public DownloadedFiles(int thread) {
  File file = new File("./download_cache" + thread);
  Collection<File> list = FileUtils.listFiles(file, null, false);
  if (!list.isEmpty()) {
    try {
      File download = list.iterator().next();
      byte[] bytes = FileUtils.readFileToByteArray(download);
      content = Base64.encodeBase64String(bytes);
      filename = download.getName();
      mimeType = new Tika().detect(bytes, filename);
      int index = filename.lastIndexOf(".");
      if (index > -1 && index < filename.length()) {
        extension = filename.substring(index + 1).toLowerCase();
        filename = filename.substring(0, index);
      }
    } catch (Throwable t) {
      Log.exception(t);
    } finally {
      for (File cur : list) {
        FileUtils.deleteQuietly(cur);
      }
    }
  }
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
private Media makeMedia(String filename) throws FileTikaException {
    File file = new File(filename);

    if (!isBeneathParent(root, file)) {
        throw new FileTikaException(file + " is not a child of " + root);
    }

    Tika tika = new Tika();
    Metadata metadata = new Metadata();
    try {
        logger.trace("Reading metadata for {}", filename);
        InputStream in = TikaInputStream.get(file, metadata);
        Reader reader = tika.parse(in, metadata);
        logger.trace("Metadata read to {}", metadata);
        in.close();
        reader.close();
        return new FileTikaMediaImpl(file, metadata);
    } catch (Exception ex) {
        throw new FileTikaException(ex);
    }
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
	FlowFile flowFile = session.get();
	if (flowFile == null) {
		flowFile = session.create();
	}
	try {
		flowFile.getAttributes();
		flowFile = session.putAttribute(flowFile, "mime.type", "application/json");
		flowFile = session.write(flowFile, new StreamCallback() {
			@Override
			public void process(InputStream inputStream, OutputStream outputStream) throws IOException {
				Tika tika = new Tika();
				String text = "";
				try {
					text = tika.parseToString(inputStream);
				} catch (TikaException e) {
					getLogger().error("Apache Tika failed to parse input " + e.getLocalizedMessage());
					e.printStackTrace();
				}
				// TODO: wrap in JSON???
				outputStream.write(text.getBytes());
			}
		});
		session.transfer(flowFile, REL_SUCCESS);
		session.commit();
	} catch (final Throwable t) {
		getLogger().error("Unable to process ExtractTextProcessor file " + t.getLocalizedMessage());
		getLogger().error("{} failed to process due to {}; rolling back session", new Object[] { this, t });
		throw t;
	}
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
public void open(File file) {

		Tika tika = new Tika();
		tika.setMaxStringLength(999999);
		StringWriter sw = new StringWriter();
		PrintWriter pw = new PrintWriter(sw);
		
		fileName.setText(file.getName());
		int dot = file.getName().lastIndexOf('.');
		String saveName = file.getName().substring(0, dot) + ".txt";
		try {
			//long start = System.currentTimeMillis();
			String text = tika.parseToString(file);
			//long end = System.currentTimeMillis();
			pw.println(text);
			//mimeType.setText(tika.detect(file) + " (" + NumberFormat.getNumberInstance().format(end-start) + "ms)");
		} catch(Exception ex){
			ex.printStackTrace(pw);
		}
		
		pw.flush();

		plain.setText(sw.toString());
		saveFile(sw.toString(), saveName);
		plain.setCaretPosition(0);
		return ;
	}
 

import org.apache.tika.Tika; //导入依赖的package包/类
@Before
public void setUp() throws IOException, TikaException, SAXException {
    MockitoAnnotations.initMocks(this);

    createDirectories(Paths.get("fileTestFiles"));

    docxExporter = new DocxExporter();

    xlsxExporter = new XlsxExporter();

    VelocityEngine engine = new VelocityEngine();
    engine.setProperty(RuntimeConstants.RESOURCE_LOADER, "classpath");
    engine.setProperty("classpath.resource.loader.class", ClasspathResourceLoader.class.getName());
    engine.init();

    templater = new Templater();
    templater.setEngine(engine);

    pdfExporter = new PdfExporter();
    pdfExporter.setTemplater(templater);

    ObjectMapperProducer objectMapperProducer = new ObjectMapperProducer();
    mapper = objectMapperProducer.objectMapper(false, false);

    TikaProvider provider = new TikaProvider();
    Tika tika = provider.tika();

    transformer = new TikaTransformer();
    transformer.setTika(tika);
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
@Before
public void setUp() throws TikaException, IOException, SAXException {
    TikaProvider provider = new TikaProvider();
    Tika tika = provider.tika();

    transformer = new TikaTransformer();
    transformer.setTika(tika);
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
public String getMimeType() {
    try {
        Tika tika = new Tika();
        return tika.detect(Paths.get(getFilePath()));
    } catch (IOException e) {
        throw new RuntimeException(e);
    }
}
 

import org.apache.tika.Tika; //导入依赖的package包/类
@Override
public AldermanAttendance parse(Path targetFile) {
	try {
		Tika tika = new Tika();
		String content = tika.parseToString(targetFile.toFile());
		AldermanAttendance aldermanAttendance = ParserUtils.fromText(content);
		return aldermanAttendance;
	} catch (IOException | TikaException e) {
		e.printStackTrace();
		return null;
	}
}