Implement Vocab.lex_attr_data as a dataclass

spacy/lang/lex_attrs.py

@@ -1,4 +1,5 @@
 from typing import Set
+from dataclasses import dataclass, field
 import unicodedata
 import re
 
@@ -6,6 +7,32 @@ from .. import attrs
 from .tokenizer_exceptions import URL_MATCH
 
 
+@dataclass
+class LexAttrData:
+    lower: dict = field(default_factory=dict)
+    norm: dict = field(default_factory=dict)
+    prefix: dict = field(default_factory=dict)
+    suffix: dict = field(default_factory=dict)
+    is_alpha: dict = field(default_factory=dict)
+    is_digit: dict = field(default_factory=dict)
+    is_lower: dict = field(default_factory=dict)
+    is_space: dict = field(default_factory=dict)
+    is_title: dict = field(default_factory=dict)
+    is_upper: dict = field(default_factory=dict)
+    is_stop: dict = field(default_factory=dict)
+    like_email: dict = field(default_factory=dict)
+    like_num: dict = field(default_factory=dict)
+    is_punct: dict = field(default_factory=dict)
+    is_ascii: dict = field(default_factory=dict)
+    shape: dict = field(default_factory=dict)
+    is_bracket: dict = field(default_factory=dict)
+    is_quote: dict = field(default_factory=dict)
+    is_left_punct: dict = field(default_factory=dict)
+    is_right_punct: dict = field(default_factory=dict)
+    is_currency: dict = field(default_factory=dict)
+    like_url: dict = field(default_factory=dict)
+
+
 _like_email = re.compile(r"([a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+)").match
 _tlds = set(
     "com|org|edu|gov|net|mil|aero|asia|biz|cat|coop|info|int|jobs|mobi|museum|"
@@ -179,8 +206,8 @@ def is_upper(vocab, string: str) -> bool:
 
 
 def is_stop(vocab, string: str) -> bool:
-    stops = vocab.lex_attr_data.get("stops", [])
-    return string.lower() in stops
+    stop_words = vocab.lex_attr_data.is_stop.get("stop_words", set())
+    return string.lower() in stop_words
 
 
 def norm(vocab, string: str) -> str:

spacy/language.py

@@ -28,6 +28,7 @@ from .scorer import Scorer
 from .util import registry, SimpleFrozenList, _pipe, raise_error, _DEFAULT_EMPTY_PIPES
 from .util import SimpleFrozenDict, combine_score_weights, CONFIG_SECTION_ORDER
 from .util import warn_if_jupyter_cupy
+from .lang.lex_attrs import LexAttrData
 from .lang.tokenizer_exceptions import URL_MATCH, BASE_EXCEPTIONS
 from .lang.punctuation import TOKENIZER_PREFIXES, TOKENIZER_SUFFIXES
 from .lang.punctuation import TOKENIZER_INFIXES
@@ -74,7 +75,7 @@ class BaseDefaults:
     url_match: Optional[Callable] = URL_MATCH
     syntax_iterators: Dict[str, Callable] = {}
     lex_attr_getters: Dict[int, Callable[[str], Any]] = {}
-    lex_attr_data: Dict[str, Any] = {}
+    lex_attr_data: LexAttrData = LexAttrData()
     stop_words: Set[str] = set()
     writing_system = {"direction": "ltr", "has_case": True, "has_letters": True}
 

spacy/tests/serialize/test_serialize_vocab_strings.py

@@ -11,7 +11,7 @@ from spacy.tokens import Doc
 from spacy.util import ensure_path, load_model
 from spacy.vectors import Vectors
 from spacy.vocab import Vocab
-from spacy.lang.lex_attrs import norm
+from spacy.lang.lex_attrs import norm, LexAttrData
 
 from ..util import make_tempdir
 
@@ -142,7 +142,12 @@ def test_deserialize_vocab_seen_entries(strings, lex_attr):
 
 @pytest.mark.parametrize("strings,lex_attr", test_strings_attrs)
 def test_serialize_vocab_lex_attrs_disk(strings, lex_attr):
-    vocab1 = Vocab(strings=strings, lex_attr_getters={NORM: norm})
+    stop_words_data = {"stop_words": ["a", "b", "c"]}
+    vocab1 = Vocab(
+        strings=strings,
+        lex_attr_getters={NORM: norm},
+        lex_attr_data=LexAttrData(is_stop=stop_words_data),
+    )
     vocab2 = Vocab(lex_attr_getters={NORM: norm})
     s = next(iter(vocab1.strings))
     vocab1[s].norm_ = lex_attr
@@ -153,6 +158,8 @@ def test_serialize_vocab_lex_attrs_disk(strings, lex_attr):
         vocab1.to_disk(file_path)
         vocab2 = vocab2.from_disk(file_path)
     assert vocab2[s].norm_ == lex_attr
+    assert vocab1.lex_attr_data == vocab2.lex_attr_data
+    assert vocab2.lex_attr_data.is_stop == stop_words_data
 
 
 @pytest.mark.parametrize("strings1,strings2", test_strings)
@@ -197,7 +204,7 @@ def test_pickle_vocab(strings, lex_attr):
     ops = get_current_ops()
     vectors = Vectors(data=ops.xp.zeros((10, 10)), mode="floret", hash_count=1)
     vocab.vectors = vectors
-    vocab.lex_attr_data = {"a": 1}
+    vocab.lex_attr_data.is_stop = {"stop_words": [1, 2, 3]}
     vocab[strings[0]].norm_ = lex_attr
     vocab_pickled = pickle.dumps(vocab)
     vocab_unpickled = pickle.loads(vocab_pickled)

spacy/vocab.pyi

@@ -3,6 +3,7 @@ from typing import Any, Iterable
 from thinc.types import Floats1d, FloatsXd
 from . import Language
 from .strings import StringStore
+from .lang.lex_attrs import LexAttrData
 from .lexeme import Lexeme
 from .lookups import Lookups
 from .morphology import Morphology
@@ -31,6 +32,7 @@ class Vocab:
         oov_prob: float = ...,
         writing_system: Dict[str, Any] = ...,
         get_noun_chunks: Optional[Callable[[Union[Doc, Span]], Iterator[Span]]] = ...,
+        lex_attr_data=LexAttrData,
     ) -> None: ...
     @property
     def lang(self) -> str: ...

spacy/vocab.pyx

@@ -1,4 +1,6 @@
 # cython: profile=True
+from dataclasses import asdict
+
 from libc.string cimport memcpy
 
 import numpy
@@ -21,14 +23,15 @@ from .util import registry
 from .lookups import Lookups
 from . import util
 from .lang.norm_exceptions import BASE_NORMS
-from .lang.lex_attrs import LEX_ATTRS
+from .lang.lex_attrs import LEX_ATTRS, LexAttrData
 
 
 def create_vocab(lang, defaults):
     # If the spacy-lookups-data package is installed, we pre-populate the lookups
     # with lexeme data, if available
     lex_attrs = {**LEX_ATTRS, **defaults.lex_attr_getters}
-    defaults.lex_attr_data["stops"] = list(defaults.stop_words)
+    if len(defaults.stop_words) > 0:
+        defaults.lex_attr_data.is_stop["stop_words"] = list(defaults.stop_words)
     lookups = Lookups()
     lookups.add_table("lexeme_norm", BASE_NORMS)
     return Vocab(
@@ -63,7 +66,7 @@ cdef class Vocab:
             A function that yields base noun phrases used for Doc.noun_chunks.
         """
         lex_attr_getters = lex_attr_getters if lex_attr_getters is not None else {}
-        lex_attr_data = lex_attr_data if lex_attr_data is not None else {}
+        lex_attr_data = lex_attr_data if lex_attr_data is not None else LexAttrData()
         if lookups in (None, True, False):
             lookups = Lookups()
         self.cfg = {'oov_prob': oov_prob}
@@ -467,7 +470,7 @@ cdef class Vocab:
         if "lookups" not in exclude:
             self.lookups.to_disk(path)
         if "lex_attr_data" not in exclude:
-            srsly.write_msgpack(path / "lex_attr_data", self.lex_attr_data)
+            srsly.write_msgpack(path / "lex_attr_data", asdict(self.lex_attr_data))
 
     def from_disk(self, path, *, exclude=tuple()):
         """Loads state from a directory. Modifies the object in place and
@@ -488,7 +491,7 @@ cdef class Vocab:
         if "lookups" not in exclude:
             self.lookups.from_disk(path)
         if "lex_attr_data" not in exclude:
-            self.lex_attr_data = srsly.read_msgpack(path / "lex_attr_data")
+            self.lex_attr_data = LexAttrData(**srsly.read_msgpack(path / "lex_attr_data"))
         self._reset_lexeme_cache()
         return self
 
@@ -510,7 +513,7 @@ cdef class Vocab:
             "strings": lambda: self.strings.to_bytes(),
             "vectors": deserialize_vectors,
             "lookups": lambda: self.lookups.to_bytes(),
-            "lex_attr_data": lambda: srsly.msgpack_dumps(self.lex_attr_data)
+            "lex_attr_data": lambda: srsly.msgpack_dumps(asdict(self.lex_attr_data))
         }
         return util.to_bytes(getters, exclude)
 
@@ -530,7 +533,7 @@ cdef class Vocab:
                 return self.vectors.from_bytes(b, exclude=["strings"])
 
         def serialize_lex_attr_data(b):
-            self.lex_attr_data = srsly.msgpack_loads(b)
+            self.lex_attr_data = LexAttrData(**srsly.msgpack_loads(b))
 
         setters = {
             "strings": lambda b: self.strings.from_bytes(b),