from deprecated.sphinx import deprecated, versionchanged, versionadded
import tempfile
from mowl.datasets import Dataset
from mowl.owlapi import OWLAPIAdapter
from mowl.evaluation import RankingEvaluator
from mowl.error import messages as msg
from java.util import HashSet
[docs]
@versionchanged(version="0.1.0", reason="Parameter ``model_filepath`` added in the base class for \
all models. Optional parameter that will use temporary files in case it is not set.")
class Model():
"""Abstract model class.
:param dataset: Dataset object.
:type dataset: mowl.datasets.base.Dataset
:param model_filepath: Path for saving the model. Defaults to a temporary file path.
:type model_filepath: str, optional
"""
def __init__(self, dataset, model_filepath=None):
if not isinstance(dataset, Dataset):
raise TypeError("Parameter dataset must be a mOWL Dataset.")
if model_filepath is not None and not isinstance(model_filepath, str):
raise TypeError("Optional parameter model_filepath must be of type str.")
self.dataset = dataset
self._model_filepath = model_filepath
self._testing_set = None
self._evaluator = None
self._evaluation_model = None
self._metrics = None
[docs]
def train(self, *args, **kwargs):
'''Abstract method for training the model. This method must be implemented in children classes
'''
raise NotImplementedError("Method train is not implemented.")
[docs]
def evaluate(self, *args, **kwargs):
if self._evaluator is None:
raise AttributeError(msg.EVALUATOR_NOT_SET)
self._metrics = self._evaluator.evaluate(self.evaluation_model,
*args, **kwargs)
[docs]
def eval_fn(self, *args, **kwargs):
raise NotImplementedError("Method eval_fn is not implemented.")
[docs]
@versionadded(version="0.2.0", reason="Axiom scoring method added to the base class.")
def score(self, axiom):
"""
Returns the score of the given axiom.
:param axiom: The axiom to score.
:type axiom: :class:`org.semanticweb.owlapi.model.OWLAxiom`
"""
raise NotImplementedError("Method score is not implemented.")
@property
def model_filepath(self):
"""Path for saving the model.
:rtype: str
"""
if self._model_filepath is None:
model_filepath = tempfile.NamedTemporaryFile()
self._model_filepath = model_filepath.name
return self._model_filepath
@property
def class_index_dict(self):
"""Dictionary with class names as keys and class indexes as values.
:rtype: dict
"""
class_to_id = {v: k for k, v in enumerate(self.dataset.classes.as_str)}
return class_to_id
@property
def individual_index_dict(self):
"""Dictionary with individual names as keys and indexes as values.
:rtype: dict
"""
individual_to_id = {v: k for k, v in enumerate(self.dataset.individuals.as_str)}
return individual_to_id
@property
def object_property_index_dict(self):
"""Dictionary with object property names as keys and object property indexes as values.
:rtype: dict
"""
object_property_to_id = {v: k for k, v in enumerate(self.dataset.object_properties.as_str)}
return object_property_to_id
@versionadded(version="0.2.0")
@property
def class_embeddings(self):
"""
Returns a dictionary with class names as keys and class embeddings as values.
:rtype: dict
"""
raise NotImplementedError()
@versionadded(version="0.2.0")
@property
def object_property_embeddings(self):
"""
Returns a dictionary with object property names as keys and object property embeddings as values.
:rtype: dict
"""
raise NotImplementedError()
@versionadded(version="0.2.0")
@property
def individual_embeddings(self):
"""
Returns a dictionary with individual names as keys and individual embeddings as values.
:rtype: dict
"""
raise NotImplementedError()
@versionadded(version="1.0.0")
@property
def evaluation_model(self):
"""Returns the evaluation model. In models relying on Word2Vec embeddings, this method calls an auxiliary evaluation model for scoring. Methods using KGEs or Geometric Embeddings would return the model itself."""
raise NotImplementedError("Method evaluation_model must be implemented in a subclass.")
@versionadded(version="1.0.0")
@property
def metrics(self):
if self._metrics is None:
raise AttributeError("Model has not been evaluated yet.")
else:
return self._metrics
[docs]
@versionadded(version="0.2.0")
def add_axioms(self, *axioms):
"""
This method adds axioms to the dataset contained in the model and reorders the embedding information for each entity accordingly. New entites are initalized with random embedding.
:param axioms: Axioms to be added to the dataset.
:type axioms: org.semanticweb.owlapi.model.OWLAxiom
"""
raise NotImplementedError()
[docs]
@versionadded(version="0.2.0")
def from_pretrained(self, file_name):
"""
This method loads a pretrained model from a file.
:param file_name: Path to the pretrained model file.
:type file_name: str
"""
raise NotImplementedError()
[docs]
@versionadded(version="1.0.0")
def set_evaluator(self, evaluator, *args, **kwargs):
"""
This method sets the evaluator for the model.
:param evaluator: Evaluator object.
:type evaluator: mowl.evaluation.base.Evaluator
"""
if isinstance(evaluator, RankingEvaluator):
self._evaluator = evaluator
else:
self._evaluator = evaluator(self.dataset, *args, **kwargs)