Apache Tika-不正确的MIME（内容类型）检测

Question

I'm trying to detect a file content type passed to a web service into the SOAP envelop. This file can be indicated in two ways :

• from its url,
• from its contain (base64 compressed data).

At this point, I'm able to translate this file into a stream buffer. But, all my tries to get its content type failed. The content type is detected if the file extension is indicated otherwise the content is always detected as "plain/text".

Bellow is my class code :

class MetadataAnalyser {

private InputStream _is;

private File _file;

private void initializeAttributes() {

    _is = null;
    _file= null;

}


private void createTemporaryFile(byte[] pData) {

    FileOutputStream fos = null;
    try {
        _file = File.createTempFile(
                UUID.randomUUID().toString().replace("-", ""),
                null,
                new File("C:\\Users\\Florent\\Documents\\NetBeansProjects\\ServiceEdition\\tmp"));
    } catch (IOException e) {
        e.printStackTrace();
    }
    try {
        fos = new FileOutputStream(_file);
    } catch (FileNotFoundException e) {
        e.printStackTrace();
    }
    try {
        fos.write(pData);
    } catch (IOException e) {
        e.printStackTrace();
    }
    try {
        fos.close();
    } catch (IOException e) {
        e.printStackTrace();
    }

    _file.deleteOnExit();

}

public MetadataAnalyser(byte[] pData) {

    initializeAttributes();
    _is = new ByteArrayInputStream(pData);
    createTemporaryFile(pData);

}

public MetadataAnalyser(InputStream pIs) {

    initializeAttributes();
    _is = pIs;
    _file = null;

}

public MetadataAnalyser(File pFile) {

    initializeAttributes();
    try {
        _file = pFile;
        _is = new FileInputStream(_file);
    } catch (FileNotFoundException e) {
        e.printStackTrace();
    } catch (Exception e) {
        e.printStackTrace();
    }

}

public MetadataAnalyser(String pFile) {

    initializeAttributes();
    try {
        _file = new File(pFile);
        if (_file.exists()) {
            _is = new FileInputStream(_file);
        }
    } catch (FileNotFoundException e) {
        e.printStackTrace();
    } catch (Exception e) {
        e.printStackTrace();
    }

}

public String getContentType() {

    AutoDetectParser parser = null;
    Metadata metadata = null;
    InputStream is = null;
    String mimeType = null;

    parser = new AutoDetectParser();
    parser.setParsers(new HashMap<MediaType, Parser>());
    metadata = new Metadata();
    if(_file != null) {
        metadata.add(TikaMetadataKeys.RESOURCE_NAME_KEY, _file.getName());
    }
    try {
        is = new FileInputStream(_file);
        parser.parse(is, new DefaultHandler(), metadata, new ParseContext());
        mimeType = metadata.get(HttpHeaders.CONTENT_TYPE);
    } catch (IOException e) {
        e.printStackTrace();
    } catch (SAXException e) {
        e.printStackTrace();
    } catch (TikaException e) {
        e.printStackTrace();
    } finally {
        return mimeType;
    }

}

}

So, how to detect the MIME type even if the file extension is unknown ?

Answer 1

我认为您无法检测到没有扩展名的mime类型，您需要知道是哪个系统正在写入文件，以及预期将存在哪种文件，并因此需要设置MIME类型（我想您在您的回复中使用它）。

Answer 2

您需要确保内容在发送到Tika之前已解码，并且不需要，绝对不需要扩展，检测是通过此处所述的众所周知的mime魔术过程进行的： https ： //tika.apache.org/1.1/detection .html