mstoeck3/autopsy-flatpak
1
0
Fork 0
mirror of https://github.com/overcuriousity/autopsy-flatpak.git synced 2025-07-17 18:17:43 +00:00
Code Issues Packages Projects Releases Wiki Activity

Implemented the second pass algorithm to improve mime type accuracy

Browse Source
This commit is contained in:
U-BASIS\dsmyda 2019-05-07 12:54:31 -04:00
parent 616ac03da4
commit aaa22f80d8
1 changed files with 33 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
Core/src/org/sleuthkit/autopsy/modules/filetypeid/FileTypeDetector.java
View File

@ -109,7 +109,9 @@ public class FileTypeDetector {
* Tika, and Autopsy file type definitions take precendence over Tika. * Tika, and Autopsy file type definitions take precendence over Tika.
* *
* @throws FileTypeDetectorInitException If an initialization error occurs, * @throws FileTypeDetectorInitException If an initialization error occurs,
* e.g., user-defined file type definitions exist but cannot be loaded. * e.g., user-defined file type
* definitions exist but cannot be
* loaded.
*/ */
public FileTypeDetector() throws FileTypeDetectorInitException { public FileTypeDetector() throws FileTypeDetectorInitException {
try { try {
@ -223,7 +225,7 @@ public class FileTypeDetector {
ReadContentInputStream stream = new ReadContentInputStream(file); ReadContentInputStream stream = new ReadContentInputStream(file);
try (TikaInputStream tikaInputStream = TikaInputStream.get(stream)) { try (TikaInputStream tikaInputStream = TikaInputStream.get(stream)) {
String tikaType = tika.detect(tikaInputStream, file.getName()); String tikaType = tika.detect(tikaInputStream);
/* /*
* Remove the Tika suffix from the MIME type name. * Remove the Tika suffix from the MIME type name.
@ -234,6 +236,23 @@ public class FileTypeDetector {
*/ */
mimeType = removeOptionalParameter(mimeType); mimeType = removeOptionalParameter(mimeType);
/*
* If Tika recognizes the file signature, then use the file
* name to refine the type. In short, this is to exclude the
* mime types that are determined solely by file extension.
* More details in JIRA-4871.
*/
if (!mimeType.equals(MimeTypes.OCTET_STREAM)) {
ReadContentInputStream secondPassStream = new ReadContentInputStream(file);
try (TikaInputStream secondPassTikaStream = TikaInputStream.get(secondPassStream)) {
tikaType = tika.detect(secondPassTikaStream, file.getName());
mimeType = tikaType.replace("tika-", ""); //NON-NLS
mimeType = removeOptionalParameter(mimeType);
} catch (Exception ex) {
throw ex;
}
}
/** /**
* We cannot trust Tika's audio/mpeg mimetype. Lets verify the * We cannot trust Tika's audio/mpeg mimetype. Lets verify the
* first two bytes and confirm it is not 0xffff. Details in * first two bytes and confirm it is not 0xffff. Details in
@ -275,6 +294,7 @@ public class FileTypeDetector {
* first 4 bits. * first 4 bits.
* *
* @param x byte * @param x byte
*
* @return Flag indicating the byte if 0xFF * @return Flag indicating the byte if 0xFF
*/ */
private boolean byteIs0xFF(byte x) { private boolean byteIs0xFF(byte x) {
@ -287,6 +307,7 @@ public class FileTypeDetector {
* @param file Abstract file to read * @param file Abstract file to read
* @param offset Offset to begin reading * @param offset Offset to begin reading
* @param n Number of bytes to read * @param n Number of bytes to read
*
* @return Byte array of size n * @return Byte array of size n
* *
* @throws TskCoreException * @throws TskCoreException