def __unicode__(self):
c = len(self.pizza_set.all())
return u"%s pizza%s to %s @ %s" % (
num2eng.num2eng(c),
"s" if c > 1 else "", # Plural
self.client,
self.create_date.strftime("%H:%M"),
)
def normalize_post(all_facts):
all_normalized_facts = {}
for number, facts in all_facts.items():
if facts:
all_normalized_facts[number] = []
for fact in facts:
try:
text = fact['text']
# ignore element if it contains : or \n as these indicate element is complex
if text.find(':') >= 0 or text.find('\n') >= 0:
continue
#if len(text) < 20 or len(text) > 150:
# continue
# capitalize beginning of sentence
text = capitalize_head(text)
# add ending period if necessary
if text[-1] != '.':
text = text + '.'
# see if the number itself appears in the fact
# TODO: should use regexp for this matching to handle word boundary
text_lc = text.lower()
word_number_lc = num2eng.num2eng(int(number)).lower()
if text_lc.find(number) >= 0 or (word_number_lc and text_lc.find(word_number_lc) >= 0):
fact['self'] = True
else:
fact['self'] = False
fact['manual'] = False
fact['text'] = text
all_normalized_facts[number].append(fact)
except:
print 'Error post-parsing [{0}: {1}]'.format(number, fact['text'])
traceback.print_exc(file=sys.stdout)
return all_normalized_facts
def normalize_number(all_facts):
all_normalized_facts = {}
for number, facts in all_facts.items():
all_normalized_facts[number] = []
for fact in facts:
try:
text = fact['text']
# normalize sentence start
# TODO: only keep if it does not contain number further in the sentence
# TODO: handle "In ..., # is ..."
word_number = num2eng.num2eng(int(number))
word_number = capitalize_head(word_number)
text = capitalize_head(text)
words_tags = get_words_tags(text)
word_number_len = len(get_words_tags(word_number))
if words_tags[0][1] =='DET':
offset = 0
elif word_number and text.startswith(word_number) and words_tags[word_number_len][1] =='V':
offset = word_number_len + 1
elif (text.startswith(number) or words_tags[0][1] == 'PRO') and words_tags[1][1] == 'V':
offset = 2
else:
continue
regexp = r'^'
for i in xrange(0, offset):
regexp += r'.*?\s'
text = re.sub(regexp, '', text)
fact['pos'] = words_tags[offset][1]
fact['text'] = capitalize_head(text)
all_normalized_facts[number].append(fact)
except:
print 'Error parsing number [{0}: {1}]'.format(number, fact['text'])
#traceback.print_exc(file=sys.stdout)
return all_normalized_facts
__author__ = 'flaviocaetano'
from datetime import datetime
from num2eng import num2eng
if __name__ == '__main__':
t0 = datetime.now()
result = 0
for i in range(1000):
word = num2eng(i+1)
print word
result += len(word.replace(' ', ''))
t1 = datetime.now() - t0
print 'result: %s (%ss)' % (result, t1)
