import sys, os, re
import codecs
sys.path.append(os.environ['HOME'] + os.sep + '.pylib')
from textual import freq, tokenize, byfreq

sys.stdin = codecs.getreader('utf-8')(sys.stdin)
sys.stdout = codecs.getwriter('utf-8')(sys.stdout)

words = []
[words.extend(tokenize(line.strip())) for line in sys.stdin.readlines()]

letterfreq = freq(re.sub('\w+', '', ''.join(words)))

for letter in letterfreq.keys(): print letter


exit()
for word in words:
  for pattern in pattern:
    print pattern
    if set(word).issubset(pattern):
      words.remove(word)
      print word
    print