def test_ambiguous_data(self, caplog, classifier_class, training_data,
component_config, **kwargs):
json_data = {
"rasa_nlu_data": {
"common_examples": [
{
"text": "good",
"intent": "affirm",
"entities": []
},
{
"text": "good morning",
"intent": "greet",
"entities": []
},
{
"text": "see you",
"intent": "goodbye",
"entities": []
},
{
"text": "nice to see you",
"intent": "greet",
"entities": []
},
]
}
}
rasa_reader = RasaReader()
data = rasa_reader.read_from_json(json_data)
with pytest.warns(UserWarning) as record:
self._train_classifier(classifier_class, data, component_config,
**kwargs)
assert len(record) == 2
def test_identical_data(self, caplog, classifier_class, training_data,
component_config, **kwargs):
json_data = {
"rasa_nlu_data": {
"common_examples": [
{
"text": "good",
"intent": "affirm",
"entities": []
},
{
"text": "good",
"intent": "goodbye",
"entities": []
},
]
}
}
rasa_reader = RasaReader()
data = rasa_reader.read_from_json(json_data)
with pytest.warns(UserWarning) as record:
self._train_classifier(classifier_class, data, component_config,
**kwargs)
assert len(record) == 1
assert ("Remove (one of) the duplicates from the training data."
in record[0].message.args[0])
def test_identical_data(self, caplog, classifier_class, training_data,
component_config, **kwargs):
json_data = {
"rasa_nlu_data": {
"common_examples": [
{
"text": "good",
"intent": "affirm",
"entities": []
},
{
"text": "good",
"intent": "goodbye",
"entities": []
},
]
}
}
rasa_reader = RasaReader()
data = rasa_reader.read_from_json(json_data)
with caplog.at_level(
logging.DEBUG,
logger="rasa.nlu.classifiers.keyword__intent_classifer"):
self._train_classifier(classifier_class, data, component_config,
**kwargs)
assert len(caplog.records) == 1
def test_dump_trainable_entities(entity_extractor: Optional[Text],
expected_output: Text):
training_data_json = {
"rasa_nlu_data": {
"common_examples": [{
"text":
"test",
"intent":
"greet",
"entities": [{
"start": 0,
"end": 4,
"value": "random",
"entity": "word"
}],
}]
}
}
if entity_extractor is not None:
training_data_json["rasa_nlu_data"]["common_examples"][0]["entities"][
0]["extractor"] = entity_extractor
training_data_object = RasaReader().read_from_json(training_data_json)
md_dump = MarkdownWriter().dumps(training_data_object)
assert md_dump.splitlines()[1] == expected_output
def test_valid_data(
self, caplog, classifier_class, training_data, component_config, **kwargs
):
json_data = {
"rasa_nlu_data": {
"common_examples": [
{"text": "good", "intent": "affirm", "entities": []},
{"text": "bye", "intent": "goodbye", "entities": []},
{"text": "see ya", "intent": "goodbye", "entities": []},
{"text": "yes", "intent": "affirm", "entities": []},
{"text": "ciao", "intent": "goodbye", "entities": []},
]
}
}
rasa_reader = RasaReader()
data = rasa_reader.read_from_json(json_data)
with pytest.warns(None) as record:
self._train_classifier(classifier_class, data, component_config, **kwargs)
assert len(record) == 0