def fit(self, dataset, force_retrain=True):
"""Fits the NLU engine
Args:
dataset (dict): A valid Snips dataset
force_retrain (bool, optional): If *False*, will not retrain intent
parsers when they are already fitted. Default to *True*.
Returns:
The same object, trained.
"""
dataset = validate_and_format_dataset(dataset)
if self.config is None:
language = dataset[LANGUAGE]
default_config = DEFAULT_CONFIGS.get(language)
if default_config is not None:
self.config = self.config_type.from_dict(default_config)
else:
self.config = self.config_type()
self.load_resources_if_needed(dataset[LANGUAGE])
self.fit_builtin_entity_parser_if_needed(dataset)
self.fit_custom_entity_parser_if_needed(dataset)
parsers = []
for parser_config in self.config.intent_parsers_configs:
# Re-use existing parsers to allow pre-training
recycled_parser = None
for parser in self.intent_parsers:
if parser.unit_name == parser_config.unit_name:
recycled_parser = parser
break
if recycled_parser is None:
recycled_parser = IntentParser.from_config(
parser_config,
builtin_entity_parser=self.builtin_entity_parser,
custom_entity_parser=self.custom_entity_parser,
resources=self.resources,
random_state=self.random_state,
)
if force_retrain or not recycled_parser.fitted:
recycled_parser.fit(dataset, force_retrain)
parsers.append(recycled_parser)
self.intent_parsers = parsers
self.dataset_metadata = _get_dataset_metadata(dataset)
return self
def test_default_configs_should_work(self):
# Given
dataset = deepcopy(WEATHER_DATASET)
for language in get_all_languages():
# When
config = DEFAULT_CONFIGS.get(language)
self.assertIsNotNone(config,
"Missing default config for '%s'" % language)
dataset[LANGUAGE] = language
engine = SnipsNLUEngine(config).fit(dataset)
result = engine.parse("Please give me the weather in Paris")
# Then
intent_name = result[RES_INTENT][RES_INTENT_NAME]
self.assertEqual("SearchWeatherForecast", intent_name)
def fit(self, dataset, force_retrain=True):
"""Fit the NLU engine
Args:
dataset (dict): A valid Snips dataset
force_retrain (bool, optional): If *False*, will not retrain intent
parsers when they are already fitted. Default to *True*.
Returns:
The same object, trained.
"""
logger.info("Fitting NLU engine...")
dataset = validate_and_format_dataset(dataset)
self._dataset_metadata = _get_dataset_metadata(dataset)
if self.config is None:
language = self._dataset_metadata["language_code"]
default_config = DEFAULT_CONFIGS.get(language)
if default_config is not None:
self.config = self.config_type.from_dict(default_config)
else:
self.config = self.config_type()
self.fit_builtin_entity_parser_if_needed(dataset)
self.fit_custom_entity_parser_if_needed(dataset)
parsers = []
for parser_config in self.config.intent_parsers_configs:
# Re-use existing parsers to allow pre-training
recycled_parser = None
for parser in self.intent_parsers:
if parser.unit_name == parser_config.unit_name:
recycled_parser = parser
break
if recycled_parser is None:
recycled_parser = build_processing_unit(parser_config)
recycled_parser.builtin_entity_parser = self.builtin_entity_parser
recycled_parser.custom_entity_parser = self.custom_entity_parser
if force_retrain or not recycled_parser.fitted:
recycled_parser.fit(dataset, force_retrain)
parsers.append(recycled_parser)
self.intent_parsers = parsers
return self
def test_default_configs_should_work(self):
# Given
dataset_stream = io.StringIO("""
---
type: intent
name: TurnLightOn
utterances:
- turn on the lights
- please switch on the light
- switch the light on
- can you turn the light on ?
- I need you to turn on the lights
---
type: intent
name: GetWeather
utterances:
- what is the weather today
- What's the weather in tokyo today?
- Can you tell me the weather please ?
- what is the weather forecast for this weekend""")
dataset = Dataset.from_yaml_files("en", [dataset_stream]).json
for language in get_all_languages():
# When
config = DEFAULT_CONFIGS.get(language)
self.assertIsNotNone(config, "Missing default config for '%s'"
% language)
dataset[LANGUAGE] = language
shared = self.get_shared_data(dataset)
engine = SnipsNLUEngine(config, **shared).fit(dataset)
result = engine.parse("Please give me the weather in Paris")
# Then
self.assertIsNotNone(result[RES_INTENT])
intent_name = result[RES_INTENT][RES_INTENT_NAME]
self.assertEqual("GetWeather", intent_name)