Make the last function also a generator, so that the explanation flows better
This commit is contained in:
@@ -31,15 +31,18 @@ def non_stop_words(filename):
|
|||||||
yield w
|
yield w
|
||||||
|
|
||||||
def count_and_sort(filename):
|
def count_and_sort(filename):
|
||||||
freqs = {}
|
freqs, i = {}, 1
|
||||||
for w in non_stop_words(filename):
|
for w in non_stop_words(filename):
|
||||||
freqs[w] = 1 if w not in freqs else freqs[w]+1
|
freqs[w] = 1 if w not in freqs else freqs[w]+1
|
||||||
return sorted(freqs.iteritems(), key=operator.itemgetter(1), reverse=True)
|
if i % 5000 == 0:
|
||||||
|
yield sorted(freqs.iteritems(), key=operator.itemgetter(1), reverse=True)
|
||||||
|
i = i+1
|
||||||
|
yield sorted(freqs.iteritems(), key=operator.itemgetter(1), reverse=True)
|
||||||
#
|
#
|
||||||
# The main function
|
# The main function
|
||||||
#
|
#
|
||||||
word_freqs = count_and_sort(sys.argv[1])
|
for word_freqs in count_and_sort(sys.argv[1]):
|
||||||
for (w, c) in word_freqs[0:25]:
|
print "-----------------------------"
|
||||||
print w, ' - ', c
|
for (w, c) in word_freqs[0:25]:
|
||||||
|
print w, ' - ', c
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user