def test_entity_train(name):
chubot = ChuBotBrain(name, language='vi')
chubot.load_data(
"/media/nvidia/ssd/catkin_ws/src/chu_bot_source/chubot/usingdata/full_train.json"
)
# chubot.load_data("usingdata/vi_nlu_ask_way.json")
meta = chubot.train_nercrf()
def test_train_models():
botname = "greet_en"
nludatafile = "/media/nvidia/ssd/catkin_ws/src/chu_bot_source/chubot/data/nlu_greet.json"
chubot = ChuBotBrain(botname, language='en')
chubot.load_data(nludatafile)
meta = chubot.train()
print(meta)
def test_intent_classification():
nludatafile = "/media/nvidia/ssd/catkin_ws/src/chu_bot_source/chubot/data/nlu_greet.json"
chubot = ChuBotBrain("greet_en", language='en')
chubot.load_data(nludatafile)
meta = chubot.train_intent_classification()
print(meta)
inmessage = "sad but great. want a dog picture"
intent_probs = chubot.predict_intent(inmessage)
print(intent_probs)
def test_nercrf():
nludatafile = "/media/nvidia/ssd/catkin_ws/src/chu_bot_source/chubot/data/nlu_greet.json"
chubot = ChuBotBrain("greet_en", language='en')
chubot.load_data(nludatafile)
meta = chubot.train_nercrf()
print(meta)
inmessage = "a bird and a dog"
tagged_entities = chubot.predict_entity(inmessage)
print(tagged_entities)
def test_intent_train(name):
chubot = ChuBotBrain(name, language='vi')
chubot.load_data("data/train.json")
# chubot.load_data("data/vi_nlu_ask_way.json")
meta = chubot.train_intent_classification()
test_link = "data/test.txt"
with open(test_link,'r',encoding='utf=8') as f:
rows = f.readlines()
intent_list_test = []
intent_list_result = []
data_list_test = []
for row in rows:
parts = row.split(',')
intent_list_test.append(encode_intent(parts[0]))
data_list_test.append(parts[1])
for data in data_list_test:
responses = chubot.predict_intent(data)
(prob, intent) = responses[0]
intent_list_result.append(encode_intent(intent))
print("accuracy",accuracy_score(intent_list_test,intent_list_result))
def retrieve_image_deplicated(action_args, **kwargs):
#TODO make a synonym mapper class
def get_default_synonym(entity, value, entity_ldict):
for ent in entity_ldict:
if entity == ent['entity']:
if value in ent['synonyms']:
return ent['default_value']
nludatafile = "data/nlu_greet.json"
chubot = ChuBotBrain("greet_en", language='en')
chubot.load_data(nludatafile)
entity_synonyms = chubot.entity_synonyms
#handle arguments
arg_entity = action_args["entity"]
arg_value_type = action_args["value"]
active_entities = kwargs["active_entities"]
for ent in active_entities:
if ent['entity'] == arg_entity:
if arg_value_type == "value":
arg_value = ent['value']
elif arg_value_type == "default_value":
arg_value = get_default_synonym(ent['entity'], ent['value'], entity_synonyms)
print(arg_value)
# print('http://shibe.online/api/{}?count=1&urls=true&httpsUrls=true'.format(arg_value))
r = requests.get('http://shibe.online/api/{}?count=1&urls=true&httpsUrls=true'.format(arg_value))
response = r.content.decode()
response = response.replace('["', '')
response = response.replace('"]', '')
return response
def test_entity_train(name):
chubot = ChuBotBrain(name, language='vi')
chubot.load_data("usingdata/full_train.json")
# chubot.load_data("usingdata/vi_nlu_ask_way.json")
meta = chubot.train_nercrf()
def create_model(name):
chubot = ChuBotBrain(name, language='vi')
chubot.load_data("usingdata/full_train.json")
# chubot.load_data("usingdata/vi_nlu_ask_way.json")
meta = chubot.train()