directory-recursive file suffix frequency histograms

example.txt

First example: no args, lists all suffixes, use defaults for: directory (current working directory), sort order (alphabetical), graph character ('-'), graph width (50 chars). 
Second example: -n (sort by file count), -t 20 (list only the top 20 suffixes), --width=35 (scale graph width to 35 chars), --imatch="^\w{3,4}$" (case-insensitive regexp -- only print suffixes that are between 3 and for chars in length), -g \* (specify a new graph character). unflagged arguments are assumed to be directories to be scanned.

See the source for all options.

OTUS-ASIO:face fish$ alias sufs='/wherever/you/put/getfilesuffixes.py'
OTUS-ASIO:face fish$ sufs

>>>			/this/current/directory (total 47846)
===			                    	  21	  -
===			                 css	  60	  -
===			                 gif	 728	  -
===			                  gz	   1	  -
===			                 htm	   1	  -
===			                html	  56	  -
===			                 icc	 282	  -
===			                 ico	   1	  -
===			                jpeg	  76	  -
===			                 jpg	 42181	  ---------------------------------------------------
===			                 JPG	 686	  -
===			                  js	 146	  -
===			                less	   1	  -
===			                  md	   3	  -
===			                 pdf	   4	  -
===			                 php	   4	  -
===			                 png	 3331	  ----
===			                 psd	   7	  -
===			                 svg	  27	  -
===			            svn-base	  93	  -
===			                 swf	   2	  -
===			                 tif	  84	  -
===			              tmproj	   1	  -
===			                 txt	   7	  -
===			                 zip	  20	  -

OTUS-ASIO:face fish$ sufs -n -t 20 --width=35 --imatch="^\w{3,4}$" -g \* ~/Screenshots/

>>>			/Users/fish/Screenshots (total 28088)
===			                 jpg	 23768	  ************************************
===			                 JPG	 2005	  ***
===			                 png	 1462	  ***
===			                 dwg	 249	  *
===			                html	 163	  *
===			                 pdf	 102	  *
===			                 doc	  44	  *
===			                 mpg	  30	  *
===			                 DWG	  24	  *
===			                 xls	  17	  *
===			                 gif	  13	  *
===			                 mp3	  10	  *
===			                 wmv	   8	  *
===			                 psd	   7	  *
===			                 rtf	   6	  *
===			                 txt	   2	  *
===			                 PCT	   2	  *
===			                 PDF	   2	  *
===			                 aif	   2	  *
===			                 inv	   1	  *

getfilesuffixes.py

#!/usr/bin/env python
# encoding: utf-8
"""
getfilesuffixes.py

Created by FI$H 2000 on 2010-10-15.
Copyright (c) 2010 OST, LLC. All rights reserved.
"""

import sys, os, getopt, re, string


help_message = '''
Prints a list of all the file suffixes found in each DIR, with counts.
Defaults to the current directory wth no args.

$ %s DIR [DIR DIR etc ...]
''' % os.path.basename(__file__)


dirs = dict()
skips = ['DS_Store','hgignore']
onlies = []


class Usage(Exception):
	def __init__(self, msg):
		self.msg = msg


def getmesomesuffixes(herp):
	for wdir, wsubdirs, wfiles in os.walk(herp):
		for wfile in wfiles:
			if wfile.rfind('.') > -1:
				suf = wfile.rsplit('.').pop()
				dirs[herp][suf] = dirs[herp].get(suf, 0) + 1


#printsuf = lambda ssuf, ssufcount, smaxcount: print "===\t\t\t%20s\t %3s\t  %s" % (ssuf, ssufcount, "-" * (int(float(float(ssufcount) / float(smaxcount)) * 50) + 1))
def printsuf(ssuf, ssufcount, smaxcount, gchar="-", width=50):
	print "===\t\t\t%20s\t %3s\t  %s" % (ssuf, ssufcount, gchar * (int(float(float(ssufcount) / float(smaxcount)) * width) + 1))


def main(argv=None):
	
	numerical = False
	top = 0
	match = None
	exclude = None
	graphchar = "-"
	width = 50
	
	if argv is None:
		argv = sys.argv
	try:
		try:
			opts, args = getopt.getopt(argv[1:], "hnt:s:o:M:X:m:x:g:w:", [
				"help",
				"numerical-sort",
				"top=",
				"skip=",
				"only=",
				"match=",
				"exclude=",
				"imatch=",
				"iexclude=",
				"graph-character=",
				"width=",
			])
		except getopt.error, msg:
			raise Usage(msg)
		
		for option, value in opts:
			if option in ("-h", "--help"):
				raise Usage(help_message)
			if option in ("-n", "--numerical"):
				numerical = True
			if option in ("-t", "--top"):
				top = int(value)
			if option in ("-s", "--skip"):
				skips.extend(str(value).split(','))
			if option in ("-o", "--only"):
				onlies.extend(str(value).split(','))
			if option in ("-M", "--match"):
				match = re.compile(str(value))
			if option in ("-X", "--exclude"):
				exclude = re.compile(str(value))
			if option in ("-m", "--imatch"):
				match = re.compile(str(value), re.IGNORECASE)
			if option in ("-x", "--iexclude"):
				exclude = re.compile(str(value), re.IGNORECASE)
			if option in ("-g", "--graph-character"):
				graphchar = str(value)[0]
			if option in ("-w", "--width"):
				ww = int(value)
				if ww > 0:
					width = ww
		
		if len(args) == 0:
			args.append(os.getcwd())
		
		for durr in [os.path.abspath(arg) for arg in args]:
			if os.path.isdir(durr):
				dirs[durr] = dict()
		
		for k, v in dirs.items():
			getmesomesuffixes(k)
		
		print ""
		
		for k, v in dirs.items():
			sufs = v.items()
			
			if not numerical:
				sufs.sort(lambda s,t: cmp(s[0].lower(), t[0]))
			else:
				sufs.sort(lambda a,b: a[1] > b[1] and -1 or 1)
			
			maxcount = reduce(lambda fs, ns: fs > ns and fs or ns, map(lambda t: t[1], sufs), 1)
			#mincount = reduce(lambda fs, ns: fs < ns and fs or ns, map(lambda t: t[1], sufs), 1)
			total = reduce(lambda fs, ns: fs + ns, map(lambda t: t[1], sufs), 0)
			print ">>>\t\t\t%s (total %s)" % (k, total)
			
			i = 0
			for suf, sufcount in sufs:
				
				if len(onlies) > 0:
					try:
						onlies.index(suf)
					except ValueError:
						pass
					else:
						if not match and not exclude:
							printsuf(suf, sufcount, maxcount, graphchar, width)
							i = i + 1
						elif match and exclude:
							if not exclude.search(suf):
								if match.search(suf):
									printsuf(suf, sufcount, maxcount, graphchar, width)
									i = i + 1
						elif match and not exclude:
							if match.search(suf):
								printsuf(suf, sufcount, maxcount, graphchar, width)
								i = i + 1
						elif not match and exclude:
							if not exclude.search(suf):
								printsuf(suf, sufcount, maxcount, graphchar, width)
								i = i + 1
					if i == top:
						break
				else:
					try:
						skips.index(suf)
					except ValueError:
						if not match and not exclude:
							printsuf(suf, sufcount, maxcount, graphchar, width)
							i = i + 1
						elif match and exclude:
							if not exclude.search(suf):
								if match.search(suf):
									printsuf(suf, sufcount, maxcount, graphchar, width)
									i = i + 1
						elif match and not exclude:
							if match.search(suf):
								printsuf(suf, sufcount, maxcount, graphchar, width)
								i = i + 1
						elif not match and exclude:
							if not exclude.search(suf):
								printsuf(suf, sufcount, maxcount, graphchar, width)
								i = i + 1
					if i == top:
						break
			print ""
		
		
	except Usage, err:
		print >> sys.stderr, sys.argv[0].split("/")[-1] + ": " + str(err.msg)
		print >> sys.stderr, "\t for help use --help"
		return 2


if __name__ == "__main__":
	sys.exit(main())