spaCy/website/docs/api/vocab.md

title	teaser	tag	source
Vocab	A storage class for vocabulary and other data shared across a language	class	spacy/vocab.pyx

The Vocab object provides a lookup table that allows you to access Lexeme objects, as well as the StringStore. It also owns underlying C-data that is shared between Doc objects.

Vocab.__init__

Create the vocabulary.

Example

from spacy.vocab import Vocab
vocab = Vocab(strings=["hello", "world"])

Name	Description
lex_attr_getters	A dictionary mapping attribute IDs to functions to compute them. Defaults to None. Optional[Dict[str, Callablestr], Any]
strings	A StringStore that maps strings to hash values, and vice versa, or a list of strings. Union[List[str], StringStore]
lookups	A Lookups that stores the lexeme_norm and other large lookup tables. Defaults to None. Optional[Lookups]
oov_prob	The default OOV probability. Defaults to -20.0. float
vectors_name	A name to identify the vectors table. str
writing_system	A dictionary describing the language's writing system. Typically provided by Language.Defaults. Dict[str, Any]
get_noun_chunks	A function that yields base noun phrases used for Doc.noun_chunks. Optional[CallableUnion[Doc, Span], Iterator[Tuple[int, int, int]]]

Vocab.__len__

Get the current number of lexemes in the vocabulary.

Example

doc = nlp("This is a sentence.")
assert len(nlp.vocab) > 0

Name	Description
RETURNS	The number of lexemes in the vocabulary. int

Vocab.__getitem__

Retrieve a lexeme, given an int ID or a string. If a previously unseen string is given, a new lexeme is created and stored.

Example

apple = nlp.vocab.strings["apple"]
assert nlp.vocab[apple] == nlp.vocab["apple"]

Name	Description
id_or_string	The hash value of a word, or its string. Union[int, str]
RETURNS	The lexeme indicated by the given ID. Lexeme

Vocab.__iter__

Iterate over the lexemes in the vocabulary.

Example

stop_words = (lex for lex in nlp.vocab if lex.is_stop)

Name	Description
YIELDS	An entry in the vocabulary. Lexeme

Vocab.__contains__

Check whether the string has an entry in the vocabulary. To get the ID for a given string, you need to look it up in vocab.strings.

Example

apple = nlp.vocab.strings["apple"]
oov = nlp.vocab.strings["dskfodkfos"]
assert apple in nlp.vocab
assert oov not in nlp.vocab

Name	Description
string	The ID string. str
RETURNS	Whether the string has an entry in the vocabulary. bool

Vocab.add_flag

Set a new boolean flag to words in the vocabulary. The flag_getter function will be called over the words currently in the vocab, and then applied to new words as they occur. You'll then be able to access the flag value on each token, using token.check_flag(flag_id).

Example

def is_my_product(text):
    products = ["spaCy", "Thinc", "displaCy"]
    return text in products

MY_PRODUCT = nlp.vocab.add_flag(is_my_product)
doc = nlp("I like spaCy")
assert doc[2].check_flag(MY_PRODUCT) == True

Name	Description
flag_getter	A function that takes the lexeme text and returns the boolean flag value. Callable[[str], bool]
flag_id	An integer between 1 and 63 (inclusive), specifying the bit at which the flag will be stored. If -1, the lowest available bit will be chosen. int
RETURNS	The integer ID by which the flag value can be checked. int

Vocab.reset_vectors

Drop the current vector table. Because all vectors must be the same width, you have to call this to change the size of the vectors. Only one of the width and shape keyword arguments can be specified.

Example

nlp.vocab.reset_vectors(width=300)

Name	Description
keyword-only
width	The new width. int
shape	The new shape. int

Vocab.prune_vectors

Reduce the current vector table to nr_row unique entries. Words mapped to the discarded vectors will be remapped to the closest vector among those remaining. For example, suppose the original table had vectors for the words: ['sat', 'cat', 'feline', 'reclined']. If we prune the vector table to, two rows, we would discard the vectors for "feline" and "reclined". These words would then be remapped to the closest remaining vector – so "feline" would have the same vector as "cat", and "reclined" would have the same vector as "sat". The similarities are judged by cosine. The original vectors may be large, so the cosines are calculated in minibatches to reduce memory usage.

Example

nlp.vocab.prune_vectors(10000)
assert len(nlp.vocab.vectors) <= 10000

Name	Description
nr_row	The number of rows to keep in the vector table. int
batch_size	Batch of vectors for calculating the similarities. Larger batch sizes might be faster, while temporarily requiring more memory. int
RETURNS	A dictionary keyed by removed words mapped to (string, score) tuples, where string is the entry the removed word was mapped to, and score the similarity score between the two words. Dict[str, Tuple[str, float]]

Vocab.deduplicate_vectors

Example

nlp.vocab.deduplicate_vectors()

Remove any duplicate rows from the current vector table, maintaining the mappings for all words in the vectors.

Vocab.get_vector

Retrieve a vector for a word in the vocabulary. Words can be looked up by string or hash value. If the current vectors do not contain an entry for the word, a 0-vector with the same number of dimensions (Vocab.vectors_length) as the current vectors is returned.

Example

nlp.vocab.get_vector("apple")

Name	Description
orth	The hash value of a word, or its unicode string. Union[int, str]
RETURNS	A word vector. Size and shape are determined by the Vocab.vectors instance. numpy.ndarray[ndim=1, dtype=float32]

Vocab.set_vector

Set a vector for a word in the vocabulary. Words can be referenced by string or hash value.

Example

nlp.vocab.set_vector("apple", array([...]))

Name	Description
orth	The hash value of a word, or its unicode string. Union[int, str]
vector	The vector to set. numpy.ndarray[ndim=1, dtype=float32]

Vocab.has_vector

Check whether a word has a vector. Returns False if no vectors are loaded. Words can be looked up by string or hash value.

Example

if nlp.vocab.has_vector("apple"):
    vector = nlp.vocab.get_vector("apple")

Name	Description
orth	The hash value of a word, or its unicode string. Union[int, str]
RETURNS	Whether the word has a vector. bool

Vocab.to_disk

Save the current state to a directory.

Example

nlp.vocab.to_disk("/path/to/vocab")

Name	Description
path	A path to a directory, which will be created if it doesn't exist. Paths may be either strings or Path-like objects. Union[str, Path]
keyword-only
exclude	String names of serialization fields to exclude. Iterable[str]

Vocab.from_disk

Loads state from a directory. Modifies the object in place and returns it.

Example

from spacy.vocab import Vocab
vocab = Vocab().from_disk("/path/to/vocab")

Name	Description
path	A path to a directory. Paths may be either strings or Path-like objects. Union[str, Path]
keyword-only
exclude	String names of serialization fields to exclude. Iterable[str]
RETURNS	The modified Vocab object. Vocab

Vocab.to_bytes

Serialize the current state to a binary string.

Example

vocab_bytes = nlp.vocab.to_bytes()

Name	Description
keyword-only
exclude	String names of serialization fields to exclude. Iterable[str]
RETURNS	The serialized form of the Vocab object. Vocab

Vocab.from_bytes

Load state from a binary string.

Example

from spacy.vocab import Vocab
vocab_bytes = nlp.vocab.to_bytes()
vocab = Vocab()
vocab.from_bytes(vocab_bytes)

Name	Description
bytes_data	The data to load from. bytes
keyword-only
exclude	String names of serialization fields to exclude. Iterable[str]
RETURNS	The Vocab object. Vocab

Attributes

Example

apple_id = nlp.vocab.strings["apple"]
assert type(apple_id) == int
PERSON = nlp.vocab.strings["PERSON"]
assert type(PERSON) == int

Name	Description
strings	A table managing the string-to-int mapping. StringStore
vectors	A table associating word IDs to word vectors. Vectors
vectors_length	Number of dimensions for each word vector. int
lookups	The available lookup tables in this vocab. Lookups
writing_system	A dict with information about the language's writing system. Dict[str, Any]
get_noun_chunks 3.0	A function that yields base noun phrases used for Doc.noun_chunks. Optional[CallableUnion[Doc, Span], Iterator[Tuple[int, int, int]]]

Serialization fields

During serialization, spaCy will export several data fields used to restore different aspects of the object. If needed, you can exclude them from serialization by passing in the string names via the exclude argument.

Example

data = vocab.to_bytes(exclude=["strings", "vectors"])
vocab.from_disk("./vocab", exclude=["strings"])

Name	Description
strings	The strings in the StringStore.
vectors	The word vectors, if available.
lookups	The lookup tables, if available.