问题
I have tried these ways of finding the MIME type of a file...
Path source = Paths
.get("C://Users/akash/Desktop/FW Internal release of MSTClient-Server5.02.04_24.msg");
System.out.println(Files.probeContentType(source));
The above code returns null...
And if I use the TIKA API from Apache to get the MIME type then it gives it as text/plain...
But I want the result as application/vnd.ms-outlook
UPDATE
I also used MIME-Util.jar as follows with code...
MimeUtil2 mimeUtil = new MimeUtil2();
mimeUtil.registerMimeDetector("eu.medsea.mimeutil.detector.MagicMimeMimeDetector");
RandomAccessFile file1 = new RandomAccessFile(
"C://Users/akash/Desktop/FW Internal release of MSTClient-Server5.02.04_24.msg",
"r");
System.out.println(file1.length());
byte[] file = new byte[624128];
file1.read(file, 0, 624128);
String mimeType = MimeUtil2.getMostSpecificMimeType(mimeUtil.getMimeTypes(file)).toString();
This gives me output as application/msword
UPDATE:
Tika API is out of scope as it is too large to include in the project...
So how can I find the MIME type?
回答1:
I tried some of the possible ways and using tika gives the result you expected, I don't see the code you used so i cannot double check it.
I tried different ways, not all in the code snippet:
- Java 7
Files.probeContentType(path) URLConnectionmime detection from file name and content type guessing- JDK 6 JAF API
javax.activation.MimetypesFileTypeMap - MimeUtil with all available subclass of
MimeDetectorI found - Apache Tika
- Apache POI scratchpad
Here the test class:
import java.io.BufferedInputStream;
import java.io.File;
import java.io.FileInputStream;
import java.io.InputStream;
import java.net.URLConnection;
import java.util.Collection;
import javax.activation.MimetypesFileTypeMap;
import org.apache.tika.detect.Detector;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.mime.MediaType;
import org.apache.tika.parser.AutoDetectParser;
import eu.medsea.mimeutil.MimeUtil;
public class FindMime {
public static void main(String[] args) {
File file = new File("C:\\Users\\qwerty\\Desktop\\test.msg");
System.out.println("urlConnectionGuess " + urlConnectionGuess(file));
System.out.println("fileContentGuess " + fileContentGuess(file));
MimetypesFileTypeMap mimeTypesMap = new MimetypesFileTypeMap();
System.out.println("mimeTypesMap.getContentType " + mimeTypesMap.getContentType(file));
System.out.println("mimeutils " + mimeutils(file));
System.out.println("tika " + tika(file));
}
private static String mimeutils(File file) {
try {
MimeUtil.registerMimeDetector("eu.medsea.mimeutil.detector.MagicMimeMimeDetector");
MimeUtil.registerMimeDetector("eu.medsea.mimeutil.detector.ExtensionMimeDetector");
// MimeUtil.registerMimeDetector("eu.medsea.mimeutil.detector.OpendesktopMimeDetector");
MimeUtil.registerMimeDetector("eu.medsea.mimeutil.detector.WindowsRegistryMimeDetector");
// MimeUtil.registerMimeDetector("eu.medsea.mimeutil.detector.TextMimeDetector");
InputStream is = new BufferedInputStream(new FileInputStream(file));
Collection<?> mimeTypes = MimeUtil.getMimeTypes(is);
return mimeTypes.toString();
} catch (Exception e) {
// TODO: handle exception
}
return null;
}
private static String tika(File file) {
try {
InputStream is = new BufferedInputStream(new FileInputStream(file));
AutoDetectParser parser = new AutoDetectParser();
Detector detector = parser.getDetector();
Metadata md = new Metadata();
md.add(Metadata.RESOURCE_NAME_KEY, "test.msg");
MediaType mediaType = detector.detect(is, md);
return mediaType.toString();
} catch (Exception e) {
// TODO: handle exception
}
return null;
}
private static String urlConnectionGuess(File file) {
String mimeType = URLConnection.guessContentTypeFromName(file.getName());
return mimeType;
}
private static String fileContentGuess(File file) {
try {
InputStream is = new BufferedInputStream(new FileInputStream(file));
return URLConnection.guessContentTypeFromStream(is);
} catch (Exception e) {
e.printStackTrace();
return null;
}
}
}
and this is the output:
urlConnectionGuess null
fileContentGuess null
mimeTypesMap.getContentType application/octet-stream
mimeutils application/msword,application/x-hwp
tika application/vnd.ms-outlook
Updated I added this method to test other ways with Tika:
private static void tikaMore(File file) {
Tika defaultTika = new Tika();
Tika mimeTika = new Tika(new MimeTypes());
Tika typeTika = new Tika(new TypeDetector());
try {
System.out.println(defaultTika.detect(file));
System.out.println(mimeTika.detect(file));
System.out.println(typeTika.detect(file));
} catch (Exception e) {
// TODO: handle exception
}
}
tested with a msg file without extension:
application/vnd.ms-outlook
application/octet-stream
application/octet-stream
tested with a txt file renamed to msg:
text/plain
text/plain
application/octet-stream
seems that the most simple way by using the empty constructor is the most reliable in this case.
Update you can make your own checker using Apache POI scratchpad, for example this is a simple implementation to get the mime of the message or null if the file is not in the proper format (usually org.apache.poi.poifs.filesystem.NotOLE2FileException: Invalid header signature):
import org.apache.poi.hsmf.MAPIMessage;
public class PoiMsgMime {
public String getMessageMime(String fileName) {
try {
new MAPIMessage(fileName);
return "application/vnd.ms-outlook";
} catch (Exception e) {
return null;
}
}
}
回答2:
Taking a cue from comment of @Duffydake, I tried reading the magic numbers. Agreed that first 8 bytes of header for MS files remains same D0 CF 11 E0 A1 B1 1A E1 ( Interesting to see first four byte which looks lik eDoCFilE) but you can check this link how to understand complete header and find the file type. (e.g in the link finds an excel file but you can use similar byte reading to find the msg file type)
If you can make assumption that no one is going to play around and store, .doc or .xls file as .msg file, then you can just read the first 8 bytes of header and combine it with file extension e.g if(fileExtension.equals(".msg")&&hexHeaderString.equals('D0 CF 11 E0 A1 B1 1A E1'){mimeType=="application/vnd.ms-outlook"}
回答3:
What you could do is to try to convert the file to byte[] and then useMimeMagic (Maven location here) to handle it. Something like that:
byte[] data = FileUtils.toByteArray("file.msg");
MagicMatch match = Magic.getMagicMatch(data);
String mimeType = match.getMimeType();
I'm not really sure that this will work 100%, but to try is not to die :)
回答4:
I had to get another workaround. What I found was that MS documents (doc, docx, xls, xlsx, msg) are compressed files with a different extension. I have not tested every MS File Type as it is outside of current scope
Simply expand the file and:
Docx : open [Content_Types].xml and check if it contains "wordprocessingml"
XlsX : open [Content_Types].xml and check if it contains "spreadsheetml"
doc : check for file "WordDocument"
xls : check for file "Workbook"
msg : check for file "__properties_version1.0"
I am still testing msg to see if there is something better to use, but this file exists in sent and unsent messages, so I assume it is safe to use.
来源：https://stackoverflow.com/questions/31071425/how-to-get-the-mime-type-of-a-msg-file