/fuzzyhashlib

Python hashlib-like wrapper for several fuzzy hash algorithms.

Primary LanguagePythonApache License 2.0Apache-2.0

What if fuzzy hashing algorithms were usable through a common API, the same way that Python's hashlib exposes secure hash and message digest algorithms? This is the question fuzzyhashlib attempts to answer by providing a hashlib-like interface to:

  • ssdeep (via a Python ctypes wrapper)
  • sdbf (via sdhash's Python swig interface)
  • tlsh (via tlsh's Python C interface)

Currently supported for Python 2.7 on 32- and 64-bit Ubuntu (Ubuntu 16.04).

Example usage in iPython is provided below.

In [1]: import fuzzyhashlib

In [2]: fuzzyhashlib.ssdeep("a" * 1024).hexdigest()
Out[2]: '3:tj1:n'

In [3]: fuzzyhashlib.sdhash("a" * 1024).hexdigest()
Out[3]: 'sdbf:03:0::1024:sha1:256:5:7ff:160:1:0:AAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAA==\n'

In [4]: # Comparison example, using subtraction operator.

In [5]: fuzzyhashlib.ssdeep("ab" * 2048) - fuzzyhashlib.ssdeep("ab" * 2048)
Out[5]: 100

In [6]: # Comparison example, using .compare() method.

In [7]: fuzzyhashlib.ssdeep("ab" * 2048).compare(fuzzyhashlib.ssdeep("ab" * 2048))
Out[7]: 100

Version 0.0.8 - Improved documentation and class interfaces:

  • Improve documentation to include examples of how to do comparisons
  • Adds .compare() method to fuzzyhash objects

Version 0.0.7 - Add support for tlsh:

  • Adds support for tlsh

Version 0.0.6 - Documentation and error handling:

  • documentation adds description of behaviour if buf and hash paramters are provided at initialisation
  • sdhash class raises exception if provided buffer is < 512 bytes in size

Version 0.0.5 - Imporoved packaging:

  • (and the versions in-between) fix packaging so libs install correctly

Version 0.0.2 - 32-bit libraries for linux:

  • libraries installed for x86-32 linux too

Version 0.0.1 - Initial version:

  • ssdeep and sdhash wrapped with hashlib-like interfaces
  • libraries installed for x86-64 linux only

The fuzzyhashlib library and package is licensed under APLv2.

The information in this section is also provided under fuzzyhashlib's source tree under 'NOTICE'.

ssdeep was written by Jesse Kornblum of the ManTech International Corporation, with acknowledged contributions by Jason Sherman and testing by the Computer Science Department at the University of Iowa. Compiled libraries built from unmodified ssdeep source, version 2.13, are installed as part of this package. The original source code for this version of ssdeep available from:

http://downloads.sourceforge.net/project/ssdeep/ssdeep-2.13/ssdeep-2.13.tar.gz

ssdeep is open source software and is licensed under GPLv2.

sdhash was written by Vassil Roussev and Candice Quates. Compiled libraries built from unmodified sdhash source, version 3.4, are installed are installed as part of this package. The original source code for this version of sdhash is available from:

http://github.com/sdhash/sdhash

sdhash is open source software and is licensed under APLv2.

tlsh was written by Jonathan Oliver, Chun Cheng, Yanggui Chen, Scott Forman and Jayson Pryde of Trend Micro. Compiled libraries built from unmodified tlsh source, version 3.9.1 (641cb4), are installed are installed as part of this package. The original source code for this version of tlsh is available from:

http://github.com/trendmicro/tlsh.git

tlsh is open source software and is licensed under APLv2 and BSDv3.

Thanks to Michael Dorman whose excellent 'yara-ctypes' provided logical layout for a Python ctypes-based project. Several concepts and functions here have been borrowed accordingly :-)