spaCy/spacy/pipeline/attributeruler.py

331 lines
13 KiB
Python
Raw Normal View History

2020-10-10 19:55:07 +03:00
from typing import List, Dict, Union, Iterable, Any, Optional, Callable
2020-10-04 14:54:05 +03:00
from typing import Tuple
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
import srsly
from pathlib import Path
from .pipe import Pipe
from ..errors import Errors
2020-10-04 14:54:05 +03:00
from ..training import validate_examples, Example
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
from ..language import Language
from ..matcher import Matcher
from ..scorer import Scorer
from ..symbols import IDS, TAG, POS, MORPH, LEMMA
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
from ..tokens import Doc, Span
from ..tokens._retokenize import normalize_token_attrs, set_token_attrs
from ..vocab import Vocab
from ..util import SimpleFrozenList
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
from .. import util
MatcherPatternType = List[Dict[Union[int, str], Any]]
AttributeRulerPatternType = Dict[str, Union[MatcherPatternType, Dict, int]]
2020-10-04 14:08:21 +03:00
TagMapType = Dict[str, Dict[Union[int, str], Union[int, str]]]
MorphRulesType = Dict[str, Dict[str, Dict[Union[int, str], Union[int, str]]]]
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
2020-10-04 14:54:05 +03:00
@Language.factory("attribute_ruler", default_config={"validate": False})
def make_attribute_ruler(nlp: Language, name: str, validate: bool):
return AttributeRuler(nlp.vocab, name, validate=validate)
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
class AttributeRuler(Pipe):
"""Set token-level attributes for tokens matched by Matcher patterns.
Additionally supports importing patterns from tag maps and morph rules.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""
def __init__(
2020-10-04 14:54:05 +03:00
self, vocab: Vocab, name: str = "attribute_ruler", *, validate: bool = False
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
) -> None:
2020-10-04 14:08:21 +03:00
"""Create the AttributeRuler. After creation, you can add patterns
with the `.initialize()` or `.add_patterns()` methods, or load patterns
with `.from_bytes()` or `.from_disk()`. Loading patterns will remove
any patterns you've added previously.
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
vocab (Vocab): The vocab.
name (str): The pipe name. Defaults to "attribute_ruler".
2020-10-04 14:54:05 +03:00
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
RETURNS (AttributeRuler): The AttributeRuler component.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#init
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""
self.name = name
self.vocab = vocab
2020-08-07 15:43:55 +03:00
self.matcher = Matcher(self.vocab, validate=validate)
self.validate = validate
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
self.attrs = []
2020-08-05 17:00:59 +03:00
self._attrs_unnormed = [] # store for reference
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
self.indices = []
def clear(self) -> None:
"""Reset all patterns."""
self.matcher = Matcher(self.vocab, validate=self.validate)
self.attrs = []
self._attrs_unnormed = []
self.indices = []
2020-10-04 14:08:21 +03:00
def initialize(
self,
2020-10-04 14:54:05 +03:00
get_examples: Optional[Callable[[], Iterable[Example]]],
2020-10-04 14:08:21 +03:00
*,
nlp: Optional[Language] = None,
patterns: Optional[Iterable[AttributeRulerPatternType]] = None,
2020-10-04 14:54:05 +03:00
tag_map: Optional[TagMapType] = None,
morph_rules: Optional[MorphRulesType] = None,
) -> None:
2020-10-04 14:08:21 +03:00
"""Initialize the attribute ruler by adding zero or more patterns.
2020-10-04 14:54:05 +03:00
2020-10-04 14:08:21 +03:00
Rules can be specified as a sequence of dicts using the `patterns`
keyword argument. You can also provide rules using the "tag map" or
"morph rules" formats supported by spaCy prior to v3.
"""
self.clear()
2020-10-04 14:08:21 +03:00
if patterns:
self.add_patterns(patterns)
if tag_map:
self.load_from_tag_map(tag_map)
if morph_rules:
self.load_from_morph_rules(morph_rules)
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
def __call__(self, doc: Doc) -> Doc:
"""Apply the AttributeRuler to a Doc and set all attribute exceptions.
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
doc (Doc): The document to process.
RETURNS (Doc): The processed Doc.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#call
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""
2020-09-26 01:23:09 +03:00
matches = self.matcher(doc, allow_missing=True)
# Sort by the attribute ID, so that later rules have precendence
matches = [
2020-09-29 22:39:28 +03:00
(int(self.vocab.strings[m_id]), m_id, s, e) for m_id, s, e in matches
2020-09-26 01:23:09 +03:00
]
matches.sort()
for attr_id, match_id, start, end in matches:
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
span = Span(doc, start, end, label=match_id)
2020-09-26 01:19:53 +03:00
attrs = self.attrs[attr_id]
index = self.indices[attr_id]
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
try:
2020-09-26 01:23:09 +03:00
# The index can be negative, which makes it annoying to do
# the boundscheck. Let Span do it instead.
2020-09-29 22:39:28 +03:00
token = span[index] # noqa: F841
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
except IndexError:
2020-09-26 01:23:09 +03:00
# The original exception is just our conditional logic, so we
# raise from.
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
raise ValueError(
Errors.E1001.format(
patterns=self.matcher.get(span.label),
span=[t.text for t in span],
index=index,
)
2020-09-29 22:39:28 +03:00
) from None
2020-09-26 01:23:09 +03:00
set_token_attrs(span[index], attrs)
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
return doc
def load_from_tag_map(
self, tag_map: Dict[str, Dict[Union[int, str], Union[int, str]]]
) -> None:
"""Load attribute ruler patterns from a tag map.
tag_map (dict): The tag map that maps fine-grained tags to
coarse-grained tags and morphological features.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#load_from_morph_rules
"""
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
for tag, attrs in tag_map.items():
pattern = [{"TAG": tag}]
attrs, morph_attrs = _split_morph_attrs(attrs)
if "MORPH" not in attrs:
morph = self.vocab.morphology.add(morph_attrs)
attrs["MORPH"] = self.vocab.strings[morph]
else:
morph = self.vocab.morphology.add(attrs["MORPH"])
attrs["MORPH"] = self.vocab.strings[morph]
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
self.add([pattern], attrs)
def load_from_morph_rules(
self, morph_rules: Dict[str, Dict[str, Dict[Union[int, str], Union[int, str]]]]
) -> None:
"""Load attribute ruler patterns from morph rules.
morph_rules (dict): The morph rules that map token text and
fine-grained tags to coarse-grained tags, lemmas and morphological
features.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#load_from_morph_rules
"""
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
for tag in morph_rules:
for word in morph_rules[tag]:
pattern = [{"ORTH": word, "TAG": tag}]
attrs = morph_rules[tag][word]
attrs, morph_attrs = _split_morph_attrs(attrs)
if "MORPH" in attrs:
morph = self.vocab.morphology.add(attrs["MORPH"])
attrs["MORPH"] = self.vocab.strings[morph]
elif morph_attrs:
morph = self.vocab.morphology.add(morph_attrs)
attrs["MORPH"] = self.vocab.strings[morph]
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
self.add([pattern], attrs)
def add(
self, patterns: Iterable[MatcherPatternType], attrs: Dict, index: int = 0
) -> None:
"""Add Matcher patterns for tokens that should be modified with the
provided attributes. The token at the specified index within the
matched span will be assigned the attributes.
patterns (Iterable[List[Dict]]): A list of Matcher patterns.
attrs (Dict): The attributes to assign to the target token in the
matched span.
index (int): The index of the token in the matched span to modify. May
be negative to index from the end of the span. Defaults to 0.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#add
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""
2020-09-26 01:19:53 +03:00
# We need to make a string here, because otherwise the ID we pass back
# will be interpreted as the hash of a string, rather than an ordinal.
key = str(len(self.attrs))
2020-10-10 19:55:07 +03:00
self.matcher.add(self.vocab.strings.add(key), patterns)
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
self._attrs_unnormed.append(attrs)
attrs = normalize_token_attrs(self.vocab, attrs)
self.attrs.append(attrs)
self.indices.append(index)
2020-10-04 14:54:05 +03:00
def add_patterns(self, patterns: Iterable[AttributeRulerPatternType]) -> None:
"""Add patterns from a list of pattern dicts with the keys as the
arguments to AttributeRuler.add.
2020-10-04 14:54:05 +03:00
patterns (Iterable[dict]): A list of pattern dicts with the keys
as the arguments to AttributeRuler.add (patterns/attrs/index) to
add as patterns.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#add_patterns
"""
2020-10-04 14:54:05 +03:00
for p in patterns:
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
self.add(**p)
@property
def patterns(self) -> List[AttributeRulerPatternType]:
"""All the added patterns."""
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
all_patterns = []
for i in range(len(self.attrs)):
p = {}
p["patterns"] = self.matcher.get(str(i))[1]
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
p["attrs"] = self._attrs_unnormed[i]
p["index"] = self.indices[i]
all_patterns.append(p)
return all_patterns
2020-10-04 14:54:05 +03:00
def score(self, examples: Iterable[Example], **kwargs) -> Dict[str, Any]:
"""Score a batch of examples.
examples (Iterable[Example]): The examples to score.
RETURNS (Dict[str, Any]): The scores, produced by
Scorer.score_token_attr for the attributes "tag", "pos", "morph"
and "lemma" for the target token attributes.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/tagger#score
"""
validate_examples(examples, "AttributeRuler.score")
results = {}
attrs = set()
for token_attrs in self.attrs:
attrs.update(token_attrs)
for attr in attrs:
if attr == TAG:
results.update(Scorer.score_token_attr(examples, "tag", **kwargs))
elif attr == POS:
results.update(Scorer.score_token_attr(examples, "pos", **kwargs))
elif attr == MORPH:
results.update(Scorer.score_token_attr(examples, "morph", **kwargs))
elif attr == LEMMA:
results.update(Scorer.score_token_attr(examples, "lemma", **kwargs))
return results
def to_bytes(self, exclude: Iterable[str] = SimpleFrozenList()) -> bytes:
"""Serialize the AttributeRuler to a bytestring.
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
exclude (Iterable[str]): String names of serialization fields to exclude.
RETURNS (bytes): The serialized object.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#to_bytes
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""
serialize = {}
2020-10-10 19:55:07 +03:00
serialize["vocab"] = self.vocab.to_bytes
serialize["patterns"] = lambda: srsly.msgpack_dumps(self.patterns)
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
return util.to_bytes(serialize, exclude)
def from_bytes(
self, bytes_data: bytes, exclude: Iterable[str] = SimpleFrozenList()
2020-10-04 14:54:05 +03:00
) -> "AttributeRuler":
"""Load the AttributeRuler from a bytestring.
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
bytes_data (bytes): The data to load.
exclude (Iterable[str]): String names of serialization fields to exclude.
returns (AttributeRuler): The loaded object.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#from_bytes
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""
2020-08-29 14:01:10 +03:00
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
def load_patterns(b):
self.add_patterns(srsly.msgpack_loads(b))
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
deserialize = {
2020-10-10 19:55:07 +03:00
"vocab": lambda b: self.vocab.from_bytes(b),
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"patterns": load_patterns,
}
util.from_bytes(bytes_data, deserialize, exclude)
return self
def to_disk(
self, path: Union[Path, str], exclude: Iterable[str] = SimpleFrozenList()
) -> None:
"""Serialize the AttributeRuler to disk.
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
path (Union[Path, str]): A path to a directory.
exclude (Iterable[str]): String names of serialization fields to exclude.
2020-10-04 14:54:05 +03:00
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#to_disk
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""
serialize = {
2020-10-10 19:55:07 +03:00
"vocab": lambda p: self.vocab.to_disk(p),
"patterns": lambda p: srsly.write_msgpack(p, self.patterns),
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
}
util.to_disk(path, serialize, exclude)
def from_disk(
self, path: Union[Path, str], exclude: Iterable[str] = SimpleFrozenList()
2020-10-04 14:54:05 +03:00
) -> "AttributeRuler":
"""Load the AttributeRuler from disk.
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
path (Union[Path, str]): A path to a directory.
exclude (Iterable[str]): String names of serialization fields to exclude.
2020-10-04 14:54:05 +03:00
RETURNS (AttributeRuler): The loaded object.
2020-09-04 13:58:50 +03:00
DOCS: https://nightly.spacy.io/api/attributeruler#from_disk
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""
2020-08-29 14:01:10 +03:00
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
def load_patterns(p):
self.add_patterns(srsly.read_msgpack(p))
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
deserialize = {
2020-10-10 19:55:07 +03:00
"vocab": lambda p: self.vocab.from_disk(p),
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"patterns": load_patterns,
}
util.from_disk(path, deserialize, exclude)
return self
2020-10-04 14:54:05 +03:00
def _split_morph_attrs(attrs: dict) -> Tuple[dict, dict]:
Add AttributeRuler for token attribute exceptions (#5842) * Add AttributeRuler for token attribute exceptions Add the `AttributeRuler` to handle exceptions for token-level attributes. The `AttributeRuler` uses `Matcher` patterns to identify target spans and applies the specified attributes to the token at the provided index in the matched span. A negative index can be used to index from the end of the matched span. The retokenizer is used to "merge" the individual tokens and assign them the provided attributes. Helper functions can import existing tag maps and morph rules to the corresponding `Matcher` patterns. There is an additional minor bug fix for `MORPH` attributes in the retokenizer to correctly normalize the values and to handle `MORPH` alongside `_` in an attrs dict. * Fix default name * Update name in error message * Extend AttributeRuler functionality * Add option to initialize with a dict of AttributeRuler patterns * Instead of silently discarding overlapping matches (the default behavior for the retokenizer if only the attrs differ), split the matches into disjoint sets and retokenize each set separately. This allows, for instance, one pattern to set the POS and another pattern to set the lemma. (If two matches modify the same attribute, it looks like the attrs are applied in the order they were added, but it may not be deterministic?) * Improve types * Sort spans before processing * Fix index boundaries in Span * Refactor retokenizer to separate attrs methods Add top-level `normalize_token_attrs` and `set_token_attrs` methods. * Update AttributeRuler to use refactored methods Update `AttributeRuler` to replace use of full retokenizer with only the relevant methods for normalizing and setting attributes for a single token. * Update spacy/pipeline/attributeruler.py Co-authored-by: Ines Montani <ines@ines.io> * Make API more similar to EntityRuler * Add `AttributeRuler.add_patterns` to add patterns from a list of dicts * Return list of dicts as property `AttributeRuler.patterns` * Make attrs_unnormed private * Add test loading patterns from assets * Revert "Fix index boundaries in Span" This reverts commit 8f8a5c33861bff2d7c3f19914e289139ab3a2c28. * Add Span index boundary checks (#5861) * Add Span index boundary checks * Return Span-specific IndexError in all cases * Simplify and fix if/else Co-authored-by: Ines Montani <ines@ines.io>
2020-08-04 18:02:39 +03:00
"""Split entries from a tag map or morph rules dict into to two dicts, one
with the token-level features (POS, LEMMA) and one with the remaining
features, which are presumed to be individual MORPH features."""
other_attrs = {}
morph_attrs = {}
for k, v in attrs.items():
if k in "_" or k in IDS.keys() or k in IDS.values():
other_attrs[k] = v
else:
morph_attrs[k] = v
return other_attrs, morph_attrs