2020-09-23 11:14:34 +03:00
|
|
|
from typing import List, Tuple, Iterable, Union, Iterator
|
2020-06-26 20:34:12 +03:00
|
|
|
import warnings
|
2020-09-23 11:14:34 +03:00
|
|
|
|
2020-06-26 20:34:12 +03:00
|
|
|
from ..errors import Errors, Warnings
|
2020-09-23 11:14:34 +03:00
|
|
|
from ..tokens import Span, Doc
|
2020-06-26 20:34:12 +03:00
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def iob_to_biluo(tags: Iterable[str]) -> List[str]:
|
2020-06-26 20:34:12 +03:00
|
|
|
out = []
|
|
|
|
tags = list(tags)
|
|
|
|
while tags:
|
|
|
|
out.extend(_consume_os(tags))
|
|
|
|
out.extend(_consume_ent(tags))
|
|
|
|
return out
|
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def biluo_to_iob(tags: Iterable[str]) -> List[str]:
|
2020-06-26 20:34:12 +03:00
|
|
|
out = []
|
|
|
|
for tag in tags:
|
|
|
|
if tag is None:
|
|
|
|
out.append(tag)
|
|
|
|
else:
|
|
|
|
tag = tag.replace("U-", "B-", 1).replace("L-", "I-", 1)
|
|
|
|
out.append(tag)
|
|
|
|
return out
|
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def _consume_os(tags: List[str]) -> Iterator[str]:
|
2020-06-26 20:34:12 +03:00
|
|
|
while tags and tags[0] == "O":
|
|
|
|
yield tags.pop(0)
|
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def _consume_ent(tags: List[str]) -> List[str]:
|
2020-06-26 20:34:12 +03:00
|
|
|
if not tags:
|
|
|
|
return []
|
|
|
|
tag = tags.pop(0)
|
|
|
|
target_in = "I" + tag[1:]
|
|
|
|
target_last = "L" + tag[1:]
|
|
|
|
length = 1
|
|
|
|
while tags and tags[0] in {target_in, target_last}:
|
|
|
|
length += 1
|
|
|
|
tags.pop(0)
|
|
|
|
label = tag[2:]
|
|
|
|
if length == 1:
|
|
|
|
if len(label) == 0:
|
|
|
|
raise ValueError(Errors.E177.format(tag=tag))
|
|
|
|
return ["U-" + label]
|
|
|
|
else:
|
|
|
|
start = "B-" + label
|
|
|
|
end = "L-" + label
|
|
|
|
middle = [f"I-{label}" for _ in range(1, length - 1)]
|
|
|
|
return [start] + middle + [end]
|
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def doc_to_biluo_tags(doc: Doc, missing: str = "O"):
|
2020-09-22 12:50:19 +03:00
|
|
|
return offsets_to_biluo_tags(
|
2020-06-26 20:34:12 +03:00
|
|
|
doc,
|
|
|
|
[(ent.start_char, ent.end_char, ent.label_) for ent in doc.ents],
|
|
|
|
missing=missing,
|
|
|
|
)
|
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def offsets_to_biluo_tags(
|
|
|
|
doc: Doc, entities: Iterable[Tuple[int, int, Union[str, int]]], missing: str = "O"
|
|
|
|
) -> List[str]:
|
2020-06-26 20:34:12 +03:00
|
|
|
"""Encode labelled spans into per-token tags, using the
|
|
|
|
Begin/In/Last/Unit/Out scheme (BILUO).
|
|
|
|
|
|
|
|
doc (Doc): The document that the entity offsets refer to. The output tags
|
|
|
|
will refer to the token boundaries within the document.
|
|
|
|
entities (iterable): A sequence of `(start, end, label)` triples. `start`
|
|
|
|
and `end` should be character-offset integers denoting the slice into
|
|
|
|
the original string.
|
|
|
|
RETURNS (list): A list of unicode strings, describing the tags. Each tag
|
|
|
|
string will be of the form either "", "O" or "{action}-{label}", where
|
2020-09-23 11:14:34 +03:00
|
|
|
action is one of "B", "I", "L", "U". The missing label is used where the
|
2020-06-26 20:34:12 +03:00
|
|
|
entity offsets don't align with the tokenization in the `Doc` object.
|
|
|
|
The training algorithm will view these as missing values. "O" denotes a
|
|
|
|
non-entity token. "B" denotes the beginning of a multi-token entity,
|
|
|
|
"I" the inside of an entity of three or more tokens, and "L" the end
|
|
|
|
of an entity of two or more tokens. "U" denotes a single-token entity.
|
|
|
|
|
|
|
|
EXAMPLE:
|
|
|
|
>>> text = 'I like London.'
|
|
|
|
>>> entities = [(len('I like '), len('I like London'), 'LOC')]
|
|
|
|
>>> doc = nlp.tokenizer(text)
|
2020-09-22 12:50:19 +03:00
|
|
|
>>> tags = offsets_to_biluo_tags(doc, entities)
|
2020-06-26 20:34:12 +03:00
|
|
|
>>> assert tags == ["O", "O", 'U-LOC', "O"]
|
|
|
|
"""
|
|
|
|
# Ensure no overlapping entity labels exist
|
|
|
|
tokens_in_ents = {}
|
|
|
|
starts = {token.idx: token.i for token in doc}
|
|
|
|
ends = {token.idx + len(token): token.i for token in doc}
|
|
|
|
biluo = ["-" for _ in doc]
|
|
|
|
# Handle entity cases
|
|
|
|
for start_char, end_char, label in entities:
|
|
|
|
if not label:
|
2020-07-04 17:25:34 +03:00
|
|
|
for s in starts: # account for many-to-one
|
2020-06-26 20:34:12 +03:00
|
|
|
if s >= start_char and s < end_char:
|
|
|
|
biluo[starts[s]] = "O"
|
|
|
|
else:
|
|
|
|
for token_index in range(start_char, end_char):
|
|
|
|
if token_index in tokens_in_ents.keys():
|
|
|
|
raise ValueError(
|
|
|
|
Errors.E103.format(
|
|
|
|
span1=(
|
|
|
|
tokens_in_ents[token_index][0],
|
|
|
|
tokens_in_ents[token_index][1],
|
|
|
|
tokens_in_ents[token_index][2],
|
|
|
|
),
|
|
|
|
span2=(start_char, end_char, label),
|
|
|
|
)
|
|
|
|
)
|
|
|
|
tokens_in_ents[token_index] = (start_char, end_char, label)
|
|
|
|
start_token = starts.get(start_char)
|
|
|
|
end_token = ends.get(end_char)
|
|
|
|
# Only interested if the tokenization is correct
|
|
|
|
if start_token is not None and end_token is not None:
|
|
|
|
if start_token == end_token:
|
|
|
|
biluo[start_token] = f"U-{label}"
|
|
|
|
else:
|
|
|
|
biluo[start_token] = f"B-{label}"
|
|
|
|
for i in range(start_token + 1, end_token):
|
|
|
|
biluo[i] = f"I-{label}"
|
|
|
|
biluo[end_token] = f"L-{label}"
|
|
|
|
# Now distinguish the O cases from ones where we miss the tokenization
|
|
|
|
entity_chars = set()
|
|
|
|
for start_char, end_char, label in entities:
|
|
|
|
for i in range(start_char, end_char):
|
|
|
|
entity_chars.add(i)
|
|
|
|
for token in doc:
|
|
|
|
for i in range(token.idx, token.idx + len(token)):
|
|
|
|
if i in entity_chars:
|
|
|
|
break
|
|
|
|
else:
|
|
|
|
biluo[token.i] = missing
|
|
|
|
if "-" in biluo and missing != "-":
|
|
|
|
ent_str = str(entities)
|
|
|
|
warnings.warn(
|
|
|
|
Warnings.W030.format(
|
|
|
|
text=doc.text[:50] + "..." if len(doc.text) > 50 else doc.text,
|
|
|
|
entities=ent_str[:50] + "..." if len(ent_str) > 50 else ent_str,
|
|
|
|
)
|
|
|
|
)
|
|
|
|
return biluo
|
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def biluo_tags_to_spans(doc: Doc, tags: Iterable[str]) -> List[Span]:
|
2020-06-26 20:34:12 +03:00
|
|
|
"""Encode per-token tags following the BILUO scheme into Span object, e.g.
|
|
|
|
to overwrite the doc.ents.
|
|
|
|
|
|
|
|
doc (Doc): The document that the BILUO tags refer to.
|
|
|
|
entities (iterable): A sequence of BILUO tags with each tag describing one
|
2020-09-21 16:54:05 +03:00
|
|
|
token. Each tag string will be of the form of either "", "O" or
|
2020-06-26 20:34:12 +03:00
|
|
|
"{action}-{label}", where action is one of "B", "I", "L", "U".
|
2020-09-21 16:54:05 +03:00
|
|
|
RETURNS (list): A sequence of Span objects. Each token with a missing IOB
|
|
|
|
tag is returned as a Span with an empty label.
|
2020-06-26 20:34:12 +03:00
|
|
|
"""
|
|
|
|
token_offsets = tags_to_entities(tags)
|
|
|
|
spans = []
|
|
|
|
for label, start_idx, end_idx in token_offsets:
|
|
|
|
span = Span(doc, start_idx, end_idx + 1, label=label)
|
|
|
|
spans.append(span)
|
|
|
|
return spans
|
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def biluo_tags_to_offsets(
|
|
|
|
doc: Doc, tags: Iterable[str]
|
|
|
|
) -> List[Tuple[int, int, Union[str, int]]]:
|
2020-06-26 20:34:12 +03:00
|
|
|
"""Encode per-token tags following the BILUO scheme into entity offsets.
|
|
|
|
|
|
|
|
doc (Doc): The document that the BILUO tags refer to.
|
|
|
|
entities (iterable): A sequence of BILUO tags with each tag describing one
|
|
|
|
token. Each tags string will be of the form of either "", "O" or
|
|
|
|
"{action}-{label}", where action is one of "B", "I", "L", "U".
|
|
|
|
RETURNS (list): A sequence of `(start, end, label)` triples. `start` and
|
|
|
|
`end` will be character-offset integers denoting the slice into the
|
|
|
|
original string.
|
|
|
|
"""
|
2020-09-22 12:50:19 +03:00
|
|
|
spans = biluo_tags_to_spans(doc, tags)
|
2020-06-26 20:34:12 +03:00
|
|
|
return [(span.start_char, span.end_char, span.label_) for span in spans]
|
|
|
|
|
|
|
|
|
2020-09-23 11:14:34 +03:00
|
|
|
def tags_to_entities(tags: Iterable[str]) -> List[Tuple[str, int, int]]:
|
|
|
|
"""Note that the end index returned by this function is inclusive.
|
2020-06-26 20:34:12 +03:00
|
|
|
To use it for Span creation, increment the end by 1."""
|
|
|
|
entities = []
|
|
|
|
start = None
|
|
|
|
for i, tag in enumerate(tags):
|
2020-09-17 22:10:41 +03:00
|
|
|
if tag is None or tag.startswith("-"):
|
2020-06-26 20:34:12 +03:00
|
|
|
# TODO: We shouldn't be getting these malformed inputs. Fix this.
|
|
|
|
if start is not None:
|
|
|
|
start = None
|
|
|
|
else:
|
|
|
|
entities.append(("", i, i))
|
2020-09-17 22:10:41 +03:00
|
|
|
elif tag.startswith("O"):
|
|
|
|
pass
|
2020-06-26 20:34:12 +03:00
|
|
|
elif tag.startswith("I"):
|
|
|
|
if start is None:
|
2020-09-08 23:44:25 +03:00
|
|
|
raise ValueError(Errors.E067.format(start="I", tags=tags[: i + 1]))
|
2020-09-17 22:10:41 +03:00
|
|
|
elif tag.startswith("U"):
|
2020-06-26 20:34:12 +03:00
|
|
|
entities.append((tag[2:], i, i))
|
|
|
|
elif tag.startswith("B"):
|
|
|
|
start = i
|
|
|
|
elif tag.startswith("L"):
|
2020-09-08 23:44:25 +03:00
|
|
|
if start is None:
|
|
|
|
raise ValueError(Errors.E067.format(start="L", tags=tags[: i + 1]))
|
2020-06-26 20:34:12 +03:00
|
|
|
entities.append((tag[2:], start, i))
|
|
|
|
start = None
|
|
|
|
else:
|
|
|
|
raise ValueError(Errors.E068.format(tag=tag))
|
|
|
|
return entities
|
2020-09-23 11:14:34 +03:00
|
|
|
|
|
|
|
|
|
|
|
# Fallbacks to make backwards-compat easier
|
|
|
|
offsets_from_biluo_tags = biluo_tags_to_offsets
|
|
|
|
spans_from_biluo_tags = biluo_tags_to_spans
|
|
|
|
biluo_tags_from_offsets = offsets_to_biluo_tags
|