def classify(text):
if not text:
return None
batch, hashes = {}, {}
if type(text) is str:
text = [text]
text_prob = text
# Compute hashes / keys
# Process the overrides first
for t in text:
p_hash = hashlib.sha224(t).hexdigest()
hashes[p_hash] = t
over = overrides(t)
if over:
batch[p_hash] = over
text_prob.remove(t)
# Call the Maxent classifier on the remaining items
if text_prob:
maxent = Pycla('419classifier.ser.gz')
result = maxent.classify(text_prob)
failed = []
for elem in result:
if result[elem][1] > 0.4:
batch[elem] = result[elem][0]
else:
failed.append(hashes[elem])
# Attempt fallback on low probability classifications
# Return None for unsuccessful.
for t in failed:
p_hash = hashlib.sha224(t).hexdigest()
batch[p_hash] = fallback(t)
return batch
def hasPQ(text):
if not text:
return None
batch = {}
if type(text) is str:
text = [text]
maxent = Pycla('PQclassifier.ser.gz')
result = maxent.classify(text)
for elem in result:
if result[elem][0] == 'info':
batch[elem] = True
else:
batch[elem] = False
return batch
