def test_entity_type(payload) -> None:
"""
Evaluate a set of cases from a file.
"""
body = payload["input"]
mock_entity_json = payload["mock_entity_json"]
expected = payload.get("expected")
exception = payload.get("exception")
duckling_plugin = DucklingPlugin(
dest="output.entities",
dimensions=["people", "time", "date", "duration"],
locale="en_IN",
timezone="Asia/Kolkata",
)
request_callback = request_builder(mock_entity_json)
httpretty.register_uri(httpretty.POST,
"http://0.0.0.0:8000/parse",
body=request_callback)
workflow = Workflow([duckling_plugin])
if expected:
_, output = workflow.run(Input(utterances=body))
entities = output["entities"]
for i, entity in enumerate(entities):
assert entity["entity_type"] == expected[i]["entity_type"]
elif exception:
with pytest.raises(EXCEPTIONS[exception]):
workflow.run(Input(utterances=body))
def test_plugin_no_transform():
mock_entity_json = [{
"body": "today",
"start": 0,
"value": {
"values": [{
"value": "2021-09-14T00:00:00.000+05:30",
"grain": "day",
"type": "value",
}],
"value":
"2021-09-14T00:00:00.000+05:30",
"grain":
"day",
"type":
"value",
},
"end": 5,
"dim": "time",
"latent": False,
}]
request_callback = request_builder(mock_entity_json, response_code=200)
httpretty.register_uri(httpretty.POST,
"http://0.0.0.0:8000/parse",
body=request_callback)
df = pd.DataFrame(
[
{
"data": ["today"],
"reftime": "2021-09-14T00:00:00.000+05:30",
},
{
"data": ["no"],
"reftime": "2021-09-14T00:00:00.000+05:30",
},
],
columns=["data", "reftime"],
)
duckling_plugin = DucklingPlugin(
locale="en_IN",
dimensions=["time"],
timezone="Asia/Kolkata",
dest="output.entities",
threshold=0.2,
timeout=0.01,
use_transform=False,
input_column="data",
output_column="entities",
)
df_ = duckling_plugin.transform(df)
assert "entities" not in df_.columns
def test_plugin_working_cases(payload) -> None:
"""
An end-to-end example showing how to use `DucklingPlugin` with a `Workflow`.
"""
body = payload["input"]
mock_entity_json = payload["mock_entity_json"]
expected_types = payload.get("expected")
exception = payload.get("exception")
duckling_args = payload.get("duckling")
response_code = payload.get("response_code", 200)
locale = payload.get("locale")
reference_time = payload.get("reference_time")
use_latent = payload.get("use_latent")
duckling_plugin = DucklingPlugin(dest="output.entities", **duckling_args)
request_callback = request_builder(mock_entity_json,
response_code=response_code)
httpretty.register_uri(httpretty.POST,
"http://0.0.0.0:8000/parse",
body=request_callback)
workflow = Workflow([duckling_plugin])
if isinstance(reference_time, str):
reference_time = make_unix_ts("Asia/Kolkata")(reference_time)
if expected_types is not None:
input_ = Input(
utterances=body,
locale=locale,
reference_time=reference_time,
latent_entities=use_latent,
)
_, output = workflow.run(input_)
if not output["entities"]:
assert output["entities"] == []
for i, entity in enumerate(output["entities"]):
expected_entity_type = expected_types[i]["entity_type"]
assert entity["entity_type"] == expected_entity_type
else:
with pytest.raises(EXCEPTIONS[exception]):
input_ = Input(
utterances=body,
locale=locale,
reference_time=reference_time,
latent_entities=use_latent,
)
workflow.run(input_)
def test_plugin_transform_existing_entity():
mock_entity_json = [{
"body": "today",
"start": 0,
"value": {
"values": [{
"value": "2021-09-14T00:00:00.000+05:30",
"grain": "day",
"type": "value",
}],
"value":
"2021-09-14T00:00:00.000+05:30",
"grain":
"day",
"type":
"value",
},
"end": 5,
"dim": "time",
"latent": False,
}]
request_callback = request_builder(mock_entity_json, response_code=200)
httpretty.register_uri(httpretty.POST,
"http://0.0.0.0:8000/parse",
body=request_callback)
df = pd.DataFrame(
[
{
"data": '["today"]',
"reftime": "2021-09-14T00:00:00.000+05:30",
},
{
"data":
'["no"]',
"reftime":
"2021-09-14T00:00:00.000+05:30",
"entities": [
KeywordEntity(
range={
"start": 0,
"end": 0
},
value="apple",
entity_type="fruits",
body="apple",
)
],
},
{
"data":
'["no"]',
"reftime":
pd.NA,
"entities": [
KeywordEntity(
range={
"start": 0,
"end": 0
},
value="apple",
entity_type="fruits",
body="apple",
)
],
},
],
columns=["data", "reftime", "entities"],
)
duckling_plugin = DucklingPlugin(
locale="en_IN",
dimensions=["time"],
timezone="Asia/Kolkata",
dest="output.entities",
threshold=0.2,
timeout=0.01,
use_transform=True,
input_column="data",
output_column="entities",
)
df_ = duckling_plugin.transform(df)
today = TimeEntity(
entity_type="date",
body="today",
parsers=["DucklingPlugin"],
range={
"start": 0,
"end": 5
},
score=1.0,
alternative_index=0,
latent=False,
value="2021-09-14T00:00:00.000+05:30",
origin="value",
grain="day",
)
parsed_entity = df_["entities"].iloc[0][0]
assert parsed_entity.value == today.value
assert parsed_entity.type == today.type
assert df_["entities"].iloc[1][0].value == "apple"