class Entity(RDFBase):
def __init__(self, id, label, types, offset=None, confidence=0.0):
# type: (str, str, List[str], Optional[slice], float) -> None
"""
Construct Entity Object
Parameters
----------
id: str
URI of entity
label: str
Label of entity
types: List[str]
List of types for this entity
offset: Optional[slice]
Indeces of substring where this entity was mentioned
confidence: float
Confidence value that this entity was mentioned
"""
super(Entity, self).__init__(id, label, offset, confidence)
self._types = [t for t in types if t != '' and t is not None]
self._types = list(dict.fromkeys(self._types))
@property
def types(self):
# type: () -> List[str]
return self._types
@property
def types_names(self):
# type: () -> str
return ' or '.join(
[t for t in self._types if t.lower() not in NOT_TO_MENTION_TYPES])
def add_types(self, types):
# type: (List[str]) -> ()
fixed_types = [t for t in types if t != '' and t is not None]
self._types.extend(fixed_types)
def casefold(self, format='triple'):
# type (str) -> ()
"""
Format the labels to match triples or natural language
Parameters
----------
format
Returns
-------
"""
if format == 'triple':
# Label
self._label = Literal(self.label.lower().replace(" ", "-"))
# Types
self._types = [t.lower().replace(" ", "_") for t in self.types]
elif format == 'natural':
# Label
self._label = self.label.lower().replace("_", " ")
self._label = self._label.capitalize() if is_proper_noun(
self.types) else self._label
# Types
self._types = [t.lower().replace("_", " ") for t in self.types]
