Dead-simple (and somewhat stupid) NTFS data-recovery program. Works with Python 2.7 or Python 3.x, no dependencies. It can be used to recover deleted files or files off of damaged drives, although recovery quality will depend on how badly the file data has been damaged or overwritten.
Warning: to avoid data loss, please use this on a clean clone of the disk (e.g. by using ddrescue), rather than on the original disk. This program will read large swaths of the disk (specifically the entire Master File Table), which may stress an already damaged disk. Making a clean copy ensures that you can rerun the program as many times as you want without further data loss.
A friend recently had an NTFS drive crash on him, and I happened to have learned about NTFS literally the day before (thanks MMA/TWCTF 2016!). So I put that to good use by writing this NTFS data recovery tool.
You may specify a path to a partition image file (previously created using e.g. ddrescue), or a raw disk path to read directly from the physical disk. Note that the latter should be used with extreme caution if the disk has failed, as it may stress an already-damaged disk to the point of failure.
Specifying disk paths is OS-specific:
On Windows disk paths should be specified using the device path:
\\.\Harddisk*Partition*
For example, \\.\Harddisk0Partition1 for the first partition on the first drive (note that Harddisk is 0-indexed while Partition is 1-indexed).
The program diskpart may be used to view the disk and partition numbers - use list disk, followed by select disk N, followed by list partition.
On Linux, disk paths should be specified using /dev paths, which depends on the device type. fdisk -l, parted -l or lsblk can show you which device path to use.
On macOS, disk paths should be specified using /dev/diskNsM paths. diskutil list will show you all partitions and their corresponding disk paths.
First, make a backup of your MFT:
python ntfsrecover.py /dev/diskX --save-mft mft
This will also print out the full paths to every single file on your disk. (This will be verbose as hell, but it's very useful!). Next, you can use --pattern in conjunction with --mft to selectively recover files. (--mft saves the program from having to read the MFT again; only file data will need to be read).
python ntfsrecover.py /dev/diskX --mft mft --pattern '*.jpg' --outdir recovered
You can specify --pattern multiple times to recover multiple different kinds of files in one run. It will match either the full path or the filename; thus, you can do things like --pattern '*/My Documents/*'.