file_processor.py

#!/usr/bin/env python
#DiabloHorn - https://diablohorn.com
import sys
import os
import csv
import argparse
import shutil

try:
    import magic
except:
    print "sudo pip install python-magic"
    sys.exit()
        
def file_ident(treeroot):
    EXCLUDEDEXTENSIONS = ('.json','.html')
    for root, dirs, files in os.walk(treeroot):
        for f in files:
            if f.endswith(EXCLUDEDEXTENSIONS):
                continue
            fullpath = os.path.join(root,f) 
            if os.path.isfile(fullpath) and not os.path.islink(fullpath):
                 csvstdout = csv.writer(sys.stdout, quoting=csv.QUOTE_ALL)
                 csvstdout.writerow([os.path.abspath(fullpath),magic.from_file(fullpath)])

def file_move(csvfile,keywords,targetdir):
    with open(csvfile,'r') as f:
        reader = csv.reader(f)
        for row in reader:
            if any(keyword in row[1] for keyword in keywords):
                print 'moving {0} to {1}'.format(row[0], targetdir)
                try:
                    shutil.move(row[0],targetdir)
                except shutil.Error, e:
                    print e
                    if 'already exists' in str(e):
                        os.remove(row[0])     
                        print '**removed**'                   
                    continue

        
                 
if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='split file in a PE aware manner')
    parser.add_argument('input',type=str,help='target directory or csv file to process')
    parser.add_argument('--identify',action='store_true',help='identify the file types and write CSV output to stdout')
    parser.add_argument('--move',action='store_true',help='move files with keywords to targetdir')
    parser.add_argument('--keywords',type=str,nargs='*',help='keywords to search for')
    parser.add_argument('--targetdir',help='target directory to store files')
    myargs = parser.parse_args()
    
    if myargs.identify:
        file_ident(myargs.input)
        
    if myargs.move and myargs.targetdir and myargs.keywords:
        file_move(myargs.input, myargs.keywords, myargs.targetdir)