count_filetypes.py

import os
from math import log

######################################
# http://stackoverflow.com/a/10171475
unit_list = zip(['bytes', 'kB', 'MB', 'GB', 'TB', 'PB'], [0, 0, 1, 2, 2, 2])
def sizeof_fmt(num):
    """Human friendly file size"""
    if num > 1:
        exponent = min(int(log(num, 1024)), len(unit_list) - 1)
        quotient = float(num) / 1024**exponent
        unit, num_decimals = unit_list[exponent]
        format_string = '{:.%sf} {}' % (num_decimals)
        return format_string.format(quotient, unit)
    if num == 0:
        return '0 bytes'
    if num == 1:
        return '1 byte'
######################################

def searchExt(directory,searchExt):
    
    # Check for "." in extension input and add if necessary
    if not searchExt.startswith("."):
        searchExt= "."+searchExt
        
    # Create dictionary with format {ext:[path,path,...]}
    pathsDict={}
    for path, dirs, files in os.walk(directory):
        for f in files: pathsDict[os.path.splitext(f)[1]]=[] #instantiate with empyt list
        for f in files: pathsDict[os.path.splitext(f)[1]].append(os.path.join(path,f)) # load list

    # Print results
    print "found",len(pathsDict.keys()),"filetypes" # entries in dictionary
    for ext in sorted(pathsDict.keys()): # sort alphabetically
        print len(pathsDict[ext]),"\t",ext # length of pathlist and each ext

    # Print paths for the selected extension
    if searchExt in pathsDict.keys(): # avoid error if file type not in dictionary
        print "found",len(pathsDict[searchExt]),searchExt+"s" # length of pathlist
        for filepath in pathsDict[searchExt]: # for each path in the list
            print sizeof_fmt(os.stat(filepath).st_size),"\t",filepath # filesize, path

if __name__ == '__main__':
    directory = r"C:\Temp"
    searchExt(directory,"")
    while 1==1: # repeat forever
        ext = raw_input("Enter extension or ctrl-c to quit: ")
        searchExt(directory,ext)