dialogy.types.entity.numerical package

Module contents

Numerical Entity

Provides the entity class for representing numbers in natural language. This entity is obtained via DucklingPlugin.

Plugin Walkthrough

In [1]: from dialogy.plugins import DucklingPlugin

In [2]: duckling_plugin = DucklingPlugin(
   ...:     dest="output.entities",
   ...:     dimensions=["number"],
   ...:     locale="en_IN",
   ...:     timezone="Asia/Kolkata",
   ...: )
   ...: 

In [3]: duckling_plugin.parse("12")
Out[3]: [NumericalEntity(body='12', type='value', parsers=['DucklingPlugin'], score=1.0, alternative_index=0, alternative_indices=[0], latent=False, value=12, origin='value', entity_type='number')]

In [4]: duckling_plugin.parse("twenty two")
Out[4]: [NumericalEntity(body='twenty two', type='value', parsers=['DucklingPlugin'], score=1.0, alternative_index=0, alternative_indices=[0], latent=False, value=22, origin='value', entity_type='number')]

Workflow Integration

In [5]: from dialogy.base import Input
   ...: from dialogy.plugins import DucklingPlugin
   ...: from dialogy.workflow import Workflow
   ...: 

In [6]: duckling_plugin = DucklingPlugin(
   ...:     dest="output.entities",
   ...:     dimensions=["number"],
   ...:     locale="en_IN",
   ...:     timezone="Asia/Kolkata",
   ...: )
   ...: 

In [7]: workflow = Workflow([duckling_plugin])

In [8]: _, output = workflow.run(Input(utterances=["seventy two", "4 fruits"]))

In [9]: output
Out[9]: 
{'intents': [],
 'entities': [{'range': {'start': 0, 'end': 11},
   'body': 'seventy two',
   'type': 'value',
   'parsers': ['DucklingPlugin'],
   'score': 0.5,
   'alternative_index': 0,
   'value': 72,
   'entity_type': 'number'},
  {'range': {'start': 0, 'end': 1},
   'body': '4',
   'type': 'value',
   'parsers': ['DucklingPlugin'],
   'score': 0.5,
   'alternative_index': 1,
   'value': 4,
   'entity_type': 'number'}],
 'original_intent': {}}

CASTING

We have noticed that there is a need to cast numerical values to time. The as_time method can be used to do this. The default assumption is to use the number as day but we can also replace the month and hour.

In [1]: import pytz
   ...: from datetime import datetime
   ...: from dialogy.workflow import Workflow
   ...: from dialogy.base import Input
   ...: from dialogy.types import Intent
   ...: from dialogy.plugins import DucklingPlugin
   ...: from dialogy.utils import dt2timestamp
   ...: 

In [2]: duckling_plugin = DucklingPlugin(
   ...:     dest="output.entities",
   ...:     dimensions=["number"],
   ...:     locale="en_IN",
   ...:     timezone="Asia/Kolkata",
   ...: )
   ...: 

In [3]: reference_dt = datetime(2022, 1, 1, 12, 0, 0, tzinfo=pytz.timezone("Asia/Kolkata"))
   ...: reference_time = dt2timestamp(reference_dt)
   ...: 

In [4]: entities = duckling_plugin.parse("12")

In [5]: [entity.as_time(reference_time, "Asia/Kolkata") for entity in entities]
Out[5]: [TimeEntity(body='12', type='value', parsers=[], score=1.0, alternative_index=0, alternative_indices=None, latent=False, value='2022-01-12T11:37:00+05:30', entity_type='date', origin='value', dim='time', grain='day')]

class NumericalEntity(body, dim=None, parsers=NOTHING, score=None, alternative_index=None, alternative_indices=None, latent=False, values=NOTHING, value=None, origin='value', entity_type='number', *, range, type='value')

Bases: dialogy.types.entity.base_entity.BaseEntity

Numerical Entity Type

Use this type for handling all entities that can be parsed to obtain: - numbers - date - time - datetime

Attributes:

• dim dimension of the entity from duckling parser

• type is the type of the entity which can have values in [“value”, “interval”]

as_time(reference_unix_ts, timezone, replace='day')

Converts a duration entity to a time entity.

Return type

TimeEntity

entity_type: Optional[str]

classmethod from_duckling(d, alternative_index, **kwargs)

Return type

NumericalEntity