spaCy/spacy/pipeline/span_ruler.py

570 lines
20 KiB
Python
Raw Normal View History

Add SpanRuler component (#9880) * Add SpanRuler component Add a `SpanRuler` component similar to `EntityRuler` that saves a list of matched spans to `Doc.spans[spans_key]`. The matches from the token and phrase matchers are deduplicated and sorted before assignment but are not otherwise filtered. * Update spacy/pipeline/span_ruler.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Fix cast * Add self.key property * Use number of patterns as length * Remove patterns kwarg from init * Update spacy/tests/pipeline/test_span_ruler.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Add options for spans filter and setting to ents * Add `spans_filter` option as a registered function' * Make `spans_key` optional and if `None`, set to `doc.ents` instead of `doc.spans[spans_key]`. * Update and generalize tests * Add test for setting doc.ents, fix key property type * Fix typing * Allow independent doc.spans and doc.ents * If `spans_key` is set, set `doc.spans` with `spans_filter`. * If `annotate_ents` is set, set `doc.ents` with `ents_fitler`. * Use `util.filter_spans` by default as `ents_filter`. * Use a custom warning if the filter does not work for `doc.ents`. * Enable use of SpanC.id in Span * Support id in SpanRuler as Span.id * Update types * `id` can only be provided as string (already by `PatternType` definition) * Update all uses of Span.id/ent_id in Doc * Rename Span id kwarg to span_id * Update types and docs * Add ents filter to mimic EntityRuler overwrite_ents * Refactor `ents_filter` to take `entities, spans` args for more filtering options * Give registered filters more descriptive names * Allow registered `filter_spans` filter (`spacy.first_longest_spans_filter.v1`) to take any number of `Iterable[Span]` objects as args so it can be used for spans filter or ents filter * Implement future entity ruler as span ruler Implement a compatible `entity_ruler` as `future_entity_ruler` using `SpanRuler` as the underlying component: * Add `sort_key` and `sort_reverse` to allow the sorting behavior to be customized. (Necessary for the same sorting/filtering as in `EntityRuler`.) * Implement `overwrite_overlapping_ents_filter` and `preserve_existing_ents_filter` to support `EntityRuler.overwrite_ents` settings. * Add `remove_by_id` to support `EntityRuler.remove` functionality. * Refactor `entity_ruler` tests to parametrize all tests to test both `entity_ruler` and `future_entity_ruler` * Implement `SpanRuler.token_patterns` and `SpanRuler.phrase_patterns` properties. Additional changes: * Move all config settings to top-level attributes to avoid duplicating settings in the config vs. `span_ruler/cfg`. (Also avoids a lot of casting.) * Format * Fix filter make method name * Refactor to use same error for removing by label or ID * Also provide existing spans to spans filter * Support ids property * Remove token_patterns and phrase_patterns * Update docstrings * Add span ruler docs * Fix types * Apply suggestions from code review Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Move sorting into filters * Check for all tokens in seen tokens in entity ruler filters * Remove registered sort key * Set Token.ent_id in a backwards-compatible way in Doc.set_ents * Remove sort options from API docs * Update docstrings * Rename entity ruler filters * Fix and parameterize scoring * Add id to Span API docs * Fix typo in API docs * Include explicit labeled=True for scorer Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2022-06-02 14:12:53 +03:00
from typing import Optional, Union, List, Dict, Tuple, Iterable, Any, Callable
from typing import Sequence, Set, cast
import warnings
from functools import partial
from pathlib import Path
import srsly
from .pipe import Pipe
from ..training import Example
from ..language import Language
from ..errors import Errors, Warnings
from ..util import ensure_path, SimpleFrozenList, registry
from ..tokens import Doc, Span
from ..scorer import Scorer
from ..matcher import Matcher, PhraseMatcher
from .. import util
PatternType = Dict[str, Union[str, List[Dict[str, Any]]]]
DEFAULT_SPANS_KEY = "ruler"
@Language.factory(
"future_entity_ruler",
assigns=["doc.ents"],
default_config={
"phrase_matcher_attr": None,
"validate": False,
"overwrite_ents": False,
"scorer": {"@scorers": "spacy.entity_ruler_scorer.v1"},
"ent_id_sep": "__unused__",
},
default_score_weights={
"ents_f": 1.0,
"ents_p": 0.0,
"ents_r": 0.0,
"ents_per_type": None,
},
)
def make_entity_ruler(
nlp: Language,
name: str,
phrase_matcher_attr: Optional[Union[int, str]],
validate: bool,
overwrite_ents: bool,
scorer: Optional[Callable],
ent_id_sep: str,
):
if overwrite_ents:
ents_filter = prioritize_new_ents_filter
else:
ents_filter = prioritize_existing_ents_filter
return SpanRuler(
nlp,
name,
spans_key=None,
spans_filter=None,
annotate_ents=True,
ents_filter=ents_filter,
phrase_matcher_attr=phrase_matcher_attr,
validate=validate,
overwrite=False,
scorer=scorer,
)
@Language.factory(
"span_ruler",
assigns=["doc.spans"],
default_config={
"spans_key": DEFAULT_SPANS_KEY,
"spans_filter": None,
"annotate_ents": False,
"ents_filter": {"@misc": "spacy.first_longest_spans_filter.v1"},
"phrase_matcher_attr": None,
"validate": False,
"overwrite": True,
"scorer": {
"@scorers": "spacy.overlapping_labeled_spans_scorer.v1",
"spans_key": DEFAULT_SPANS_KEY,
},
},
default_score_weights={
f"spans_{DEFAULT_SPANS_KEY}_f": 1.0,
f"spans_{DEFAULT_SPANS_KEY}_p": 0.0,
f"spans_{DEFAULT_SPANS_KEY}_r": 0.0,
f"spans_{DEFAULT_SPANS_KEY}_per_type": None,
},
)
def make_span_ruler(
nlp: Language,
name: str,
spans_key: Optional[str],
spans_filter: Optional[Callable[[Iterable[Span], Iterable[Span]], Iterable[Span]]],
annotate_ents: bool,
ents_filter: Callable[[Iterable[Span], Iterable[Span]], Iterable[Span]],
phrase_matcher_attr: Optional[Union[int, str]],
validate: bool,
overwrite: bool,
scorer: Optional[Callable],
):
return SpanRuler(
nlp,
name,
spans_key=spans_key,
spans_filter=spans_filter,
annotate_ents=annotate_ents,
ents_filter=ents_filter,
phrase_matcher_attr=phrase_matcher_attr,
validate=validate,
overwrite=overwrite,
scorer=scorer,
)
def prioritize_new_ents_filter(
entities: Iterable[Span], spans: Iterable[Span]
) -> List[Span]:
"""Merge entities and spans into one list without overlaps by allowing
spans to overwrite any entities that they overlap with. Intended to
replicate the overwrite_ents=True behavior from the EntityRuler.
entities (Iterable[Span]): The entities, already filtered for overlaps.
spans (Iterable[Span]): The spans to merge, may contain overlaps.
RETURNS (List[Span]): Filtered list of non-overlapping spans.
"""
get_sort_key = lambda span: (span.end - span.start, -span.start)
spans = sorted(spans, key=get_sort_key, reverse=True)
entities = list(entities)
new_entities = []
seen_tokens: Set[int] = set()
for span in spans:
start = span.start
end = span.end
if all(token.i not in seen_tokens for token in span):
new_entities.append(span)
entities = [e for e in entities if not (e.start < end and e.end > start)]
seen_tokens.update(range(start, end))
return entities + new_entities
@registry.misc("spacy.prioritize_new_ents_filter.v1")
def make_prioritize_new_ents_filter():
return prioritize_new_ents_filter
def prioritize_existing_ents_filter(
entities: Iterable[Span], spans: Iterable[Span]
) -> List[Span]:
"""Merge entities and spans into one list without overlaps by prioritizing
existing entities. Intended to replicate the overwrite_ents=False behavior
from the EntityRuler.
entities (Iterable[Span]): The entities, already filtered for overlaps.
spans (Iterable[Span]): The spans to merge, may contain overlaps.
RETURNS (List[Span]): Filtered list of non-overlapping spans.
"""
get_sort_key = lambda span: (span.end - span.start, -span.start)
spans = sorted(spans, key=get_sort_key, reverse=True)
entities = list(entities)
new_entities = []
seen_tokens: Set[int] = set()
seen_tokens.update(*(range(ent.start, ent.end) for ent in entities))
for span in spans:
start = span.start
end = span.end
if all(token.i not in seen_tokens for token in span):
new_entities.append(span)
seen_tokens.update(range(start, end))
return entities + new_entities
@registry.misc("spacy.prioritize_existing_ents_filter.v1")
def make_preverse_existing_ents_filter():
return prioritize_existing_ents_filter
def overlapping_labeled_spans_score(
examples: Iterable[Example], *, spans_key=DEFAULT_SPANS_KEY, **kwargs
) -> Dict[str, Any]:
kwargs = dict(kwargs)
attr_prefix = f"spans_"
kwargs.setdefault("attr", f"{attr_prefix}{spans_key}")
kwargs.setdefault("allow_overlap", True)
kwargs.setdefault("labeled", True)
kwargs.setdefault(
"getter", lambda doc, key: doc.spans.get(key[len(attr_prefix) :], [])
)
kwargs.setdefault("has_annotation", lambda doc: spans_key in doc.spans)
return Scorer.score_spans(examples, **kwargs)
@registry.scorers("spacy.overlapping_labeled_spans_scorer.v1")
def make_overlapping_labeled_spans_scorer(spans_key: str = DEFAULT_SPANS_KEY):
return partial(overlapping_labeled_spans_score, spans_key=spans_key)
class SpanRuler(Pipe):
"""The SpanRuler lets you add spans to the `Doc.spans` using token-based
rules or exact phrase matches.
DOCS: https://spacy.io/api/spanruler
USAGE: https://spacy.io/usage/rule-based-matching#spanruler
"""
def __init__(
self,
nlp: Language,
name: str = "span_ruler",
*,
spans_key: Optional[str] = DEFAULT_SPANS_KEY,
spans_filter: Optional[
Callable[[Iterable[Span], Iterable[Span]], Iterable[Span]]
] = None,
annotate_ents: bool = False,
ents_filter: Callable[
[Iterable[Span], Iterable[Span]], Iterable[Span]
] = util.filter_chain_spans,
phrase_matcher_attr: Optional[Union[int, str]] = None,
validate: bool = False,
overwrite: bool = False,
scorer: Optional[Callable] = partial(
overlapping_labeled_spans_score, spans_key=DEFAULT_SPANS_KEY
),
) -> None:
"""Initialize the span ruler. If patterns are supplied here, they
need to be a list of dictionaries with a `"label"` and `"pattern"`
key. A pattern can either be a token pattern (list) or a phrase pattern
(string). For example: `{'label': 'ORG', 'pattern': 'Apple'}`.
nlp (Language): The shared nlp object to pass the vocab to the matchers
and process phrase patterns.
name (str): Instance name of the current pipeline component. Typically
passed in automatically from the factory when the component is
added. Used to disable the current span ruler while creating
phrase patterns with the nlp object.
spans_key (Optional[str]): The spans key to save the spans under. If
`None`, no spans are saved. Defaults to "ruler".
spans_filter (Optional[Callable[[Iterable[Span], Iterable[Span]], List[Span]]):
The optional method to filter spans before they are assigned to
doc.spans. Defaults to `None`.
annotate_ents (bool): Whether to save spans to doc.ents. Defaults to
`False`.
ents_filter (Callable[[Iterable[Span], Iterable[Span]], List[Span]]):
The method to filter spans before they are assigned to doc.ents.
Defaults to `util.filter_chain_spans`.
phrase_matcher_attr (Optional[Union[int, str]]): Token attribute to
match on, passed to the internal PhraseMatcher as `attr`. Defaults
to `None`.
validate (bool): Whether patterns should be validated, passed to
Matcher and PhraseMatcher as `validate`.
overwrite (bool): Whether to remove any existing spans under this spans
key if `spans_key` is set, and/or to remove any ents under `doc.ents` if
`annotate_ents` is set. Defaults to `True`.
scorer (Optional[Callable]): The scoring method. Defaults to
spacy.pipeline.span_ruler.overlapping_labeled_spans_score.
DOCS: https://spacy.io/api/spanruler#init
"""
self.nlp = nlp
self.name = name
self.spans_key = spans_key
self.annotate_ents = annotate_ents
self.phrase_matcher_attr = phrase_matcher_attr
self.validate = validate
self.overwrite = overwrite
self.spans_filter = spans_filter
self.ents_filter = ents_filter
self.scorer = scorer
self._match_label_id_map: Dict[int, Dict[str, str]] = {}
self.clear()
def __len__(self) -> int:
"""The number of all labels added to the span ruler."""
return len(self._patterns)
def __contains__(self, label: str) -> bool:
"""Whether a label is present in the patterns."""
for label_id in self._match_label_id_map.values():
if label_id["label"] == label:
return True
return False
@property
def key(self) -> Optional[str]:
"""Key of the doc.spans dict to save the spans under."""
return self.spans_key
def __call__(self, doc: Doc) -> Doc:
"""Find matches in document and add them as entities.
doc (Doc): The Doc object in the pipeline.
RETURNS (Doc): The Doc with added entities, if available.
DOCS: https://spacy.io/api/spanruler#call
"""
error_handler = self.get_error_handler()
try:
matches = self.match(doc)
self.set_annotations(doc, matches)
return doc
except Exception as e:
return error_handler(self.name, self, [doc], e)
def match(self, doc: Doc):
self._require_patterns()
with warnings.catch_warnings():
warnings.filterwarnings("ignore", message="\\[W036")
matches = cast(
List[Tuple[int, int, int]],
list(self.matcher(doc)) + list(self.phrase_matcher(doc)),
)
deduplicated_matches = set(
Span(
doc,
start,
end,
label=self._match_label_id_map[m_id]["label"],
span_id=self._match_label_id_map[m_id]["id"],
)
for m_id, start, end in matches
if start != end
)
return sorted(list(deduplicated_matches))
def set_annotations(self, doc, matches):
"""Modify the document in place"""
# set doc.spans if spans_key is set
if self.key:
spans = []
if self.key in doc.spans and not self.overwrite:
spans = doc.spans[self.key]
spans.extend(
self.spans_filter(spans, matches) if self.spans_filter else matches
)
doc.spans[self.key] = spans
# set doc.ents if annotate_ents is set
if self.annotate_ents:
spans = []
if not self.overwrite:
spans = list(doc.ents)
spans = self.ents_filter(spans, matches)
try:
doc.ents = sorted(spans)
except ValueError:
raise ValueError(Errors.E854)
@property
def labels(self) -> Tuple[str, ...]:
"""All labels present in the match patterns.
RETURNS (set): The string labels.
DOCS: https://spacy.io/api/spanruler#labels
"""
return tuple(sorted(set([cast(str, p["label"]) for p in self._patterns])))
@property
def ids(self) -> Tuple[str, ...]:
"""All IDs present in the match patterns.
RETURNS (set): The string IDs.
DOCS: https://spacy.io/api/spanruler#ids
"""
return tuple(
sorted(set([cast(str, p.get("id")) for p in self._patterns]) - set([None]))
)
def initialize(
self,
get_examples: Callable[[], Iterable[Example]],
*,
nlp: Optional[Language] = None,
patterns: Optional[Sequence[PatternType]] = None,
):
"""Initialize the pipe for training.
get_examples (Callable[[], Iterable[Example]]): Function that
returns a representative sample of gold-standard Example objects.
nlp (Language): The current nlp object the component is part of.
patterns (Optional[Iterable[PatternType]]): The list of patterns.
DOCS: https://spacy.io/api/spanruler#initialize
"""
self.clear()
if patterns:
self.add_patterns(patterns) # type: ignore[arg-type]
@property
def patterns(self) -> List[PatternType]:
"""Get all patterns that were added to the span ruler.
RETURNS (list): The original patterns, one dictionary per pattern.
DOCS: https://spacy.io/api/spanruler#patterns
"""
return self._patterns
def add_patterns(self, patterns: List[PatternType]) -> None:
"""Add patterns to the span ruler. A pattern can either be a token
pattern (list of dicts) or a phrase pattern (string). For example:
{'label': 'ORG', 'pattern': 'Apple'}
{'label': 'ORG', 'pattern': 'Apple', 'id': 'apple'}
{'label': 'GPE', 'pattern': [{'lower': 'san'}, {'lower': 'francisco'}]}
patterns (list): The patterns to add.
DOCS: https://spacy.io/api/spanruler#add_patterns
"""
# disable the nlp components after this one in case they haven't been
# initialized / deserialized yet
try:
current_index = -1
for i, (name, pipe) in enumerate(self.nlp.pipeline):
if self == pipe:
current_index = i
break
subsequent_pipes = [pipe for pipe in self.nlp.pipe_names[current_index:]]
except ValueError:
subsequent_pipes = []
with self.nlp.select_pipes(disable=subsequent_pipes):
phrase_pattern_labels = []
phrase_pattern_texts = []
for entry in patterns:
p_label = cast(str, entry["label"])
p_id = cast(str, entry.get("id", ""))
label = repr((p_label, p_id))
self._match_label_id_map[self.nlp.vocab.strings.as_int(label)] = {
"label": p_label,
"id": p_id,
}
if isinstance(entry["pattern"], str):
phrase_pattern_labels.append(label)
phrase_pattern_texts.append(entry["pattern"])
elif isinstance(entry["pattern"], list):
self.matcher.add(label, [entry["pattern"]])
else:
raise ValueError(Errors.E097.format(pattern=entry["pattern"]))
self._patterns.append(entry)
for label, pattern in zip(
phrase_pattern_labels,
self.nlp.pipe(phrase_pattern_texts),
):
self.phrase_matcher.add(label, [pattern])
def clear(self) -> None:
"""Reset all patterns.
RETURNS: None
DOCS: https://spacy.io/api/spanruler#clear
"""
self._patterns: List[PatternType] = []
self.matcher: Matcher = Matcher(self.nlp.vocab, validate=self.validate)
self.phrase_matcher: PhraseMatcher = PhraseMatcher(
self.nlp.vocab,
attr=self.phrase_matcher_attr,
validate=self.validate,
)
def remove(self, label: str) -> None:
"""Remove a pattern by its label.
label (str): Label of the pattern to be removed.
RETURNS: None
DOCS: https://spacy.io/api/spanruler#remove
"""
if label not in self:
raise ValueError(
Errors.E1024.format(attr_type="label", label=label, component=self.name)
)
self._patterns = [p for p in self._patterns if p["label"] != label]
for m_label in self._match_label_id_map:
if self._match_label_id_map[m_label]["label"] == label:
m_label_str = self.nlp.vocab.strings.as_string(m_label)
if m_label_str in self.phrase_matcher:
self.phrase_matcher.remove(m_label_str)
if m_label_str in self.matcher:
self.matcher.remove(m_label_str)
def remove_by_id(self, pattern_id: str) -> None:
"""Remove a pattern by its pattern ID.
pattern_id (str): ID of the pattern to be removed.
RETURNS: None
DOCS: https://spacy.io/api/spanruler#remove_by_id
"""
orig_len = len(self)
self._patterns = [p for p in self._patterns if p.get("id") != pattern_id]
if orig_len == len(self):
raise ValueError(
Errors.E1024.format(
attr_type="ID", label=pattern_id, component=self.name
)
)
for m_label in self._match_label_id_map:
if self._match_label_id_map[m_label]["id"] == pattern_id:
m_label_str = self.nlp.vocab.strings.as_string(m_label)
if m_label_str in self.phrase_matcher:
self.phrase_matcher.remove(m_label_str)
if m_label_str in self.matcher:
self.matcher.remove(m_label_str)
def _require_patterns(self) -> None:
"""Raise a warning if this component has no patterns defined."""
if len(self) == 0:
warnings.warn(Warnings.W036.format(name=self.name))
def from_bytes(
self, bytes_data: bytes, *, exclude: Iterable[str] = SimpleFrozenList()
) -> "SpanRuler":
"""Load the span ruler from a bytestring.
bytes_data (bytes): The bytestring to load.
RETURNS (SpanRuler): The loaded span ruler.
DOCS: https://spacy.io/api/spanruler#from_bytes
"""
self.clear()
deserializers = {
"patterns": lambda b: self.add_patterns(srsly.json_loads(b)),
}
util.from_bytes(bytes_data, deserializers, exclude)
return self
def to_bytes(self, *, exclude: Iterable[str] = SimpleFrozenList()) -> bytes:
"""Serialize the span ruler to a bytestring.
RETURNS (bytes): The serialized patterns.
DOCS: https://spacy.io/api/spanruler#to_bytes
"""
serializers = {
"patterns": lambda: srsly.json_dumps(self.patterns),
}
return util.to_bytes(serializers, exclude)
def from_disk(
self, path: Union[str, Path], *, exclude: Iterable[str] = SimpleFrozenList()
) -> "SpanRuler":
"""Load the span ruler from a directory.
path (Union[str, Path]): A path to a directory.
RETURNS (SpanRuler): The loaded span ruler.
DOCS: https://spacy.io/api/spanruler#from_disk
"""
self.clear()
path = ensure_path(path)
deserializers = {
"patterns": lambda p: self.add_patterns(srsly.read_jsonl(p)),
}
util.from_disk(path, deserializers, {})
return self
def to_disk(
self, path: Union[str, Path], *, exclude: Iterable[str] = SimpleFrozenList()
) -> None:
"""Save the span ruler patterns to a directory.
path (Union[str, Path]): A path to a directory.
DOCS: https://spacy.io/api/spanruler#to_disk
"""
path = ensure_path(path)
serializers = {
"patterns": lambda p: srsly.write_jsonl(p, self.patterns),
}
util.to_disk(path, serializers, {})