def find_wh(text):
""" Given text, returns WH type """
text_s = text.lower().strip()
words = utils.tokenizer(text_s)
wh_categories = set([])
for wh_type, keywords in WH_TYPES.iteritems():
temp_categories = [wh_type for keyword in keywords if keyword in words]
wh_categories = wh_categories.union(set(temp_categories))
category_weights = dict((category, 100) for category in wh_categories)
if category_weights:
return max(category_weights, key=category_weights.get)
return None
def classify(text):
text = utils.replace_slack_mentions(text)
text_s = text.lower().strip()
# Correct all the spelling mistakes
words = utils.tokenizer(text_s)
words = [corrector.correct(word) for word in words]
# find exact matches of keywords
first_categories = set([])
for category, keywords in category_map.iteritems():
temp_categories = set([category for k in keywords if k in words])
first_categories = first_categories.union(temp_categories)
category_weights = dict((category, 100) for category in first_categories)
if category_weights:
return max(category_weights, key=category_weights.get)
return None