# -*- coding: utf-8 -*-
"""A wrapper which combines an interaction function with NodePiece entity representations."""
import logging
from typing import Any, Callable, ClassVar, List, Mapping
import torch
from class_resolver import Hint, HintOrType, OptionalKwargs
from ..nbase import ERModel
from ...constants import DEFAULT_EMBEDDING_HPO_EMBEDDING_DIM_RANGE
from ...nn import NodePieceRepresentation, SubsetRepresentation, representation_resolver
from ...nn.modules import DistMultInteraction, Interaction
from ...nn.node_piece import RelationTokenizer, Tokenizer, tokenizer_resolver
from ...regularizers import Regularizer
from ...triples.triples_factory import CoreTriplesFactory
from ...typing import Constrainer, Initializer, Normalizer, OneOrSequence
from ...utils import upgrade_to_sequence
__all__ = [
"NodePiece",
]
logger = logging.getLogger(__name__)
[docs]class NodePiece(ERModel):
"""A wrapper which combines an interaction function with NodePiece entity representations from [galkin2021]_.
This model uses the :class:`pykeen.nn.NodePieceRepresentation` instead of a typical
:class:`pykeen.nn.representation.Embedding` to more efficiently store representations.
---
citation:
author: Galkin
year: 2021
link: https://arxiv.org/abs/2106.12144
github: https://github.com/migalkin/NodePiece
"""
hpo_default: ClassVar[Mapping[str, Any]] = dict(
embedding_dim=DEFAULT_EMBEDDING_HPO_EMBEDDING_DIM_RANGE,
)
def __init__(
self,
*,
triples_factory: CoreTriplesFactory,
num_tokens: OneOrSequence[int] = 2,
tokenizers: OneOrSequence[HintOrType[Tokenizer]] = None,
tokenizers_kwargs: OneOrSequence[OptionalKwargs] = None,
embedding_dim: int = 64,
interaction: HintOrType[Interaction] = DistMultInteraction,
aggregation: Hint[Callable[[torch.Tensor, int], torch.Tensor]] = None,
entity_initializer: Hint[Initializer] = None,
entity_normalizer: Hint[Normalizer] = None,
entity_constrainer: Hint[Constrainer] = None,
entity_regularizer: Hint[Regularizer] = None,
relation_initializer: Hint[Initializer] = None,
relation_normalizer: Hint[Normalizer] = None,
relation_constrainer: Hint[Constrainer] = None,
relation_regularizer: Hint[Regularizer] = None,
**kwargs,
) -> None:
"""
Initialize the model.
:param triples_factory:
the triples factory. Must have create_inverse_triples set to True.
:param num_tokens:
the number of relations to use to represent each entity, cf.
:class:`pykeen.nn.NodePieceRepresentation`.
:param tokenizers:
the tokenizer to use, cf. `pykeen.nn.node_piece.tokenizer_resolver`.
:param tokenizers_kwargs:
additional keyword-based parameters passed to the tokenizer upon construction.
:param embedding_dim:
the embedding dimension. Only used if embedding_specification is not given.
:param interaction:
the interaction module, or a hint for it.
:param aggregation:
aggregation of multiple token representations to a single entity representation. By default,
this uses :func:`torch.mean`. If a string is provided, the module assumes that this refers to a top-level
torch function, e.g. "mean" for :func:`torch.mean`, or "sum" for func:`torch.sum`. An aggregation can
also have trainable parameters, .e.g., ``MLP(mean(MLP(tokens)))`` (cf. DeepSets from [zaheer2017]_). In
this case, the module has to be created outside of this component.
Moreover, we support providing "mlp" as a shortcut to use the MLP aggregation version from [galkin2021]_.
We could also have aggregations which result in differently shapes output, e.g. a concatenation of all
token embeddings resulting in shape ``(num_tokens * d,)``. In this case, `shape` must be provided.
The aggregation takes two arguments: the (batched) tensor of token representations, in shape
``(*, num_tokens, *dt)``, and the index along which to aggregate.
:param entity_initializer:
a hint for initializing anchor embeddings
:param entity_normalizer:
a hint for normalizing anchor embeddings
:param entity_constrainer:
a hint for constraining anchor embeddings
:param entity_regularizer:
a hint for regularizing anchor embeddings
:param relation_initializer:
a hint for initializing relation embeddings
:param relation_normalizer:
a hint for normalizing relation embeddings
:param relation_constrainer:
a hint for constraining relation embeddings
:param relation_regularizer:
a hint for regularizing relation embeddings
:param kwargs:
additional keyword-based arguments passed to :meth:`ERModel.__init__`
:raises ValueError:
if the triples factory does not create inverse triples
"""
if not triples_factory.create_inverse_triples:
raise ValueError(
"The provided triples factory does not create inverse triples. However, for the node piece "
"representations inverse relation representations are required.",
)
# always create representations for normal and inverse relations and padding
relation_representations = representation_resolver.make(
query=None,
max_id=2 * triples_factory.real_num_relations + 1,
shape=embedding_dim,
initializer=relation_initializer,
normalizer=relation_normalizer,
constrainer=relation_constrainer,
regularizer=relation_regularizer,
)
# normalize embedding specification
anchor_kwargs = dict(
shape=embedding_dim,
initializer=entity_initializer,
normalizer=entity_normalizer,
constrainer=entity_constrainer,
regularizer=entity_regularizer,
)
# prepare token representations & kwargs
token_representations = []
token_representations_kwargs: List[OptionalKwargs] = []
for tokenizer in upgrade_to_sequence(tokenizers):
if tokenizer_resolver.lookup(tokenizer) is RelationTokenizer:
token_representations.append(relation_representations)
token_representations_kwargs.append(None)
else:
token_representations.append(None) # Embedding
token_representations_kwargs.append(anchor_kwargs)
super().__init__(
triples_factory=triples_factory,
interaction=interaction,
entity_representations=NodePieceRepresentation,
entity_representations_kwargs=dict(
triples_factory=triples_factory,
token_representations=token_representations,
token_representations_kwargs=token_representations_kwargs,
tokenizers=tokenizers,
tokenizers_kwargs=tokenizers_kwargs,
aggregation=aggregation,
num_tokens=num_tokens,
),
relation_representations=SubsetRepresentation,
relation_representations_kwargs=dict( # hide padding relation
# max_id=triples_factory.num_relations, # will get added by ERModel
base=relation_representations,
),
**kwargs,
)