Oops, forgot to add "mat"

2 files changed, 239 insertions, 0 deletions

diff --git a/mat b/mat
new file mode 100755
index 0000000..ef83d84
--- /dev/null
+++ b/mat
@@ -0,0 +1,161 @@
+#!/usr/bin/env python
+'''
+    Metadata anonymisation toolkit - CLI edition
+'''
+
+import sys
+import xml.sax
+import optparse
+import os
+
+import hachoir_core
+
+from lib import mat
+
+
+def parse():
+    '''
+        Get, and parse options passed to the program
+    '''
+    parser = optparse.OptionParser(usage='%prog [options] files\n\
+The default behaviour is to clean files given in argument')
+    options = optparse.OptionGroup(parser, 'Options')
+    options.add_option('--add2archive', '-a', action='store_true',
+        default=False, help='Add to output archive non-supported filetypes')
+    options.add_option('--backup', '-b', action='store_true', default=False,
+        help='Keep a backup copy')
+    options.add_option('--force', '-f', action='store_true', default=False,
+        help='Don\'t check if files are clean before cleaning')
+    options.add_option('--strict', '-u', action='store_true', default=False,
+            help='Strict cleaning mode : loss can occur')
+
+    info = optparse.OptionGroup(parser, 'Informations')
+    info.add_option('--check', '-c',  action='store_true', default=False,
+        help='Check if a file is free of harmful metadatas')
+    info.add_option('--display', '-d', action='store_true', default=False,
+        help='List all the harmful metadata of a file without removing them')
+    info.add_option('--list', '-l', action='store_true', default=False,
+        help='List all supported fileformat')
+    info.add_option('--version', '-v', action='callback',
+        callback=display_version, help='Display version and exit')
+    parser.add_option_group(options)
+    parser.add_option_group(info)
+
+    values, arguments = parser.parse_args()
+    if not arguments and values.list is False:
+        # if no argument and no files are passed,
+        # print help and exit
+        parser.print_help()
+        sys.exit(0)
+    return values, arguments
+
+
+def display_version(*_):
+    '''
+        Display the program's version, and exit
+    '''
+    print('Metadata Anonymisation Toolkit version %s') % mat.__version__
+    print('Hachoir version %s') % hachoir_core.__version__
+    sys.exit(0)
+
+
+def list_meta(class_file, filename, force):
+    '''
+        Print all the metadata of 'filename' on stdout
+    '''
+    print('[+] File %s :' % filename)
+    if force is False and class_file.is_clean():
+        print('No harmful metadata found')
+    else:
+        meta = class_file.get_meta()
+        print ('Harmful metadata found:')
+        if meta is not None:
+            for key, value in class_file.get_meta().iteritems():
+                print('\t' + key + ' : ' + str(value))
+
+
+def is_clean(class_file, filename, force):
+    '''
+        Say if 'filename' is clean or not
+    '''
+    if class_file.is_clean():
+        print('[+] %s is clean' % filename)
+    else:
+        print('[+] %s is not clean' % filename)
+
+
+def clean_meta(class_file, filename, force):
+    '''
+        Clean the file 'filename'
+    '''
+    print('[+] Cleaning %s' % filename)
+    if force is False and class_file.is_clean():
+        print('%s is already clean' % filename)
+    else:
+        if class_file.remove_all():
+            print('%s cleaned !' % filename)
+        else:
+            print('Unable to clean %s', filename)
+
+def clean_meta_strict(class_file, filename, force):
+    '''
+        Clean the file 'filename', strict way
+    '''
+    print('[+] Cleaning %s' % filename)
+    if force is False and class_file.is_clean():
+        print('%s is already clean' % filename)
+    else:
+        class_file.remove_all_strict()
+        print('%s cleaned' % filename)
+
+
+def list_supported():
+    '''
+        Print all supported fileformat, and exit
+    '''
+    handler = mat.XMLParser()
+    parser = xml.sax.make_parser()
+    parser.setContentHandler(handler)
+    path = os.path.join(mat.get_sharedir(), 'FORMATS')
+    with open(path, 'r') as xmlfile:
+        parser.parse(xmlfile)
+
+    for item in handler.list:
+        print('%s (%s)' % (item['name'], item['extension']))
+        print('\tsupport : ' + item['support'])
+        print('\tmetadata : ' + item['metadata'])
+        print('\tmethod : ' + item['method'])
+        if item['support'] == 'partial':
+            print('\tremaining : ' + item['remaining'])
+        print('\n')
+    sys.exit(0)
+
+
+def main():
+    '''
+        main function : get args, and launch the appropriate function
+    '''
+    args, filenames = parse()
+
+    #func receive the function correponding to the options given as parameters
+    if args.display is True:  # only print metadatas
+        func = list_meta
+    elif args.check is True:  # only check if the file is clean
+        func = is_clean
+    elif args.strict is True:  # destructive anonymisation method
+        func = clean_meta_strict
+    elif args.list is True:  # print the list of all supported format
+        list_supported()
+    else:  # clean the file
+        func = clean_meta
+
+    for filename in filenames:
+        class_file = mat.create_class_file(filename, args.backup,
+            args.add2archive)
+        if class_file is not None:
+            func(class_file, filename, args.force)
+        else:
+            print('Unable to process %s' % filename)
+
+if __name__ == '__main__':
+    main()
diff --git a/mat.1 b/mat.1
new file mode 100644
index 0000000..48ebf9a
--- /dev/null
+++ b/mat.1
@@ -0,0 +1,78 @@
+.TH METADATA "1" "August 2011" "Metadata Anonymisation Toolkit" "User Commands"
+
+
+.SH NAME
+MAT \- Metadata Anonymisation Toolkit
+
+
+.SH SYNOPSIS
+.B mat-cli
+[\fIoptions\fR] \fIfiles\fR
+.TP
+.B mat-gui
+
+
+.SH DESCRIPTION
+The \fBMetadata Anonymisation Toolkit\fR is a lib (with a CLI and a GUI)
+created to anonymise file's \fBmetadata\fR. In essence, metadata answer who,
+what, when, where, why, and how about every face of the data that are being
+documented. They can be a \fBrisk for privacy\fR.
+
+
+.SH OPTIONS
+.TP
+\fB\-h\fR, \fB\-\-help\fR
+show this help message and exit
+.TP
+\fB\-a\fR, \fB\-\-add2archive\fR
+Add to outputed archive non\-supported filetypes
+.TP
+\fB\-b\fR, \fB\-\-backup\fR
+Keep a backup copy
+.TP
+\fB\-c\fR, \fB\-\-check\fR
+Check if a file is free of harmful metadatas
+.TP
+\fB\-d\fR, \fB\-\-display\fR
+List all the harmful meta of a file without removing them
+.TP
+\fB\-f\fR, \fB\-\-force\fR
+Don't check if files are clean before cleaning
+.TP
+\fB\-l\fR, \fB\-\-list\fR
+List all supported fileformat
+.TP
+\fB\-u\fR, \fB\-\-strict\fR
+Remove harmful meta, but loss can occure
+.TP
+\fB\-v\fR, \fB\-\-version\fR
+Display version and exit
+
+
+.SH EXAMPLES
+.TP
+\fBmat-cli \-\-display\fR mydocument.pdf
+Display the mydocument.pdf's harmful metadata
+.TP
+\fBmat-cli \-\-check *.jpg\fR
+Check all the jpg images from the current folder
+
+
+.SH NOTES
+MAT \fBonly process metadata\fR, it does \fBnot\fR handle file data.
+Blame yourself if you are traced back because of the data of your files.
+MAT is not perfect : In most of the cases, a forensic expert with a lot
+of time \fBcan trace back\fR your document.
+If you want absolute privacy, use plain-text.
+
+
+.SH AUTHOR
+Julien (\fBjvoisin\fR) Voisin <pouicpouicpouic@gmail.com>, during the GSoC 2011
+
+
+.SH BUGS
+MAT does not handle watermaking/tattoo for now.
+
+
+.SH "SEE ALSO"
+exiftool, hachoir-metadata