class Rasa_NLU(object):
def __init__(self):
self.data_training = "testData.json"
self.Bot = Bot()
self.training_nlu()
def training_nlu(self):
training_data = load_data(self.data_training)
config = load("config_spacy.yml")
trainer = Trainer(config)
trainer.train(training_data)
model_directory = trainer.persist("projects")
self.interpreter = Interpreter.load(model_directory)
def parse_message(self, message):
parsed_message = self.interpreter.parse(message)
return parsed_message
def check_entities(self, dic):
for k, v in dic.items():
if k == 'genres':
result = self.Bot.choose_film_by_genres(v)
return result
def find_respond_from_bot(self, message):
parse_mes = self.parse_message(message)
#print(parse_mes)
if parse_mes['intent']['name'] == 'greet':
result = self.Bot.hello_message()
return result
if parse_mes['intent']['name'] == 'popular_film':
result = self.Bot.popular_film()
return result
if parse_mes['intent']['name'] == 'goodbye':
result = self.Bot.goodbye_message()
return result
if parse_mes['intent']['name'] == 'affirm':
result = self.Bot.aff_message()
return result
if parse_mes['intent']['name'] == 'new_search':
result = self.Bot.new_search()
return result
if parse_mes['intent']['name'] == 'thanks':
result = self.Bot.thanks_message()
return result
if parse_mes['intent']['name'] == 'film_search':
params = {}
if len(parse_mes['entities']) > 0:
for e in parse_mes['entities']:
params[e["entity"]] = str(e['value'])
result = self.check_entities(params)
return result
if not "intent" in parse_mes or parse_mes['intent'] is None:
result = self.Bot.bad_message()
return result
return self.Bot.bad_message()
