#!/usr/bin/env python2 -W ignore::FutureWarning
# put this in your path, rename to just 'Occur'
# Occur import <- call with argument to count one word
# or call no arguments to list frequencies of all words
# -- Uses the default Tokenizer settings
#    which obliterate all punctuation
#    and make everything lower case.
import os
import subprocess
import sys
from keras.preprocessing.text import Tokenizer

winid = os.environ['winid']
px = subprocess.Popen(
  ("9p read acme/%s/body" % winid).split(),
  stdout=subprocess.PIPE)
body, err = px.communicate()

T = Tokenizer()
T.fit_on_texts([body])

#print(T.word_counts)

if len(sys.argv) > 1:
  print(T.word_counts[sys.argv[1]])

else:
  freqs = [(v,k) for k,v in T.word_counts.items()]
  freqs.sort()
  for pair in freqs:
    print("%05d %s" % pair)