spaCy/website/docs/api/doc.jade

//- 💫 DOCS > API > DOC

include ../../_includes/_mixins

p A container for accessing linguistic annotations.

p
    |  A #[code Doc] is a sequence of #[+api("token") #[code Token]] objects.
    |  Access sentences and named entities, export annotations to numpy arrays,
    |  losslessly serialize to compressed binary strings. The #[code Doc] object
    |  holds an array of #[code TokenC] structs. The Python-level #[code Token]
    |  and #[+api("span") #[code Span]] objects are views of this array, i.e.
    |  they don't own the data themselves.

+aside-code("Example").
    # Construction 1
    doc = nlp(u'Some text')

    # Construction 2
    from spacy.tokens import Doc
    doc = doc = Doc(nlp.vocab, words=[u'hello', u'world', u'!'],
                               spaces=[True, False, False])

+h(2, "init") Doc.__init__
    +tag method

p
    |  Construct a #[code Doc] object. The most common way to get a #[code Doc]
    |  object is via the #[code nlp] object.

+table(["Name", "Type", "Description"])
    +row
        +cell #[code vocab]
        +cell #[code Vocab]
        +cell A storage container for lexical types.

    +row
        +cell #[code words]
        +cell -
        +cell A list of strings to add to the container.

    +row
        +cell #[code spaces]
        +cell -
        +cell
            |  A list of boolean values indicating whether each word has a
            |  subsequent space. Must have the same length as #[code words], if
            |  specified. Defaults to a sequence of #[code True].

    +footrow
        +cell returns
        +cell #[code Doc]
        +cell The newly constructed object.

+h(2, "getitem") Doc.__getitem__
    +tag method

p
    |  Get a #[+api("token") #[code Token]] object at position #[code i], where
    |  #[code i] is an integer. Negative indexing is supported, and follows the
    |  usual Python semantics, i.e. #[code doc[-2]] is #[code doc[len(doc) - 2]].

+aside-code("Example").
    doc = nlp(u'Give it back! He pleaded.')
    assert doc[0].text == 'Give'
    assert doc[-1].text == '.'
    span = doc[1:1]
    assert span.text == 'it back'

+table(["Name", "Type", "Description"])
    +row
        +cell #[code i]
        +cell int
        +cell The index of the token.

    +footrow
        +cell returns
        +cell #[code Token]
        +cell The token at #[code doc[i]].

p
    |  Get a #[+api("span") #[code Span]] object, starting at position
    |  #[code start] (token index) and ending at position #[code end] (token
    |  index).

p
    |  For instance, #[code doc[2:5]] produces a span consisting of tokens 2, 3
    |  and 4. Stepped slices (e.g. #[code doc[start : end : step]]) are not
    |  supported, as #[code Span] objects must be contiguous (cannot have gaps).
    |  You can use negative indices and open-ended ranges, which have their
    |  normal Python semantics.

+table(["Name", "Type", "Description"])
    +row
        +cell #[code start_end]
        +cell tuple
        +cell The slice of the document to get.

    +footrow
        +cell returns
        +cell #[code Span]
        +cell The span at #[code doc[start : end]].

+h(2, "iter") Doc.__iter__
    +tag method

p
    |  Iterate over #[code Token] objects, from which the annotations can be
    |  easily accessed.

+aside-code("Example").
    doc = nlp(u'Give it back')
    assert [t.text for t in doc] == [u'Give', u'it', u'back']

p
    |  This is the main way of accessing #[+api("token") #[code Token]] objects,
    |  which are the main way annotations are accessed from Python. If
    |  faster-than-Python speeds are required, you can instead access the
    |  annotations as a numpy array, or access the underlying C data directly
    |  from Cython.

+table(["Name", "Type", "Description"])
    +footrow
        +cell yields
        +cell #[code Token]
        +cell A #[code Token] object.

+h(2, "len") Doc.__len__
    +tag method

p Get the number of tokens in the document.

+aside-code("Example").
    doc = nlp(u'Give it back! He pleaded.')
    assert len(doc) == 7

+table(["Name", "Type", "Description"])
    +footrow
        +cell returns
        +cell int
        +cell The number of tokens in the document.

+h(2, "similarity") Doc.similarity
    +tag method
    +tag-model("vectors")

p
    |  Make a semantic similarity estimate. The default estimate is cosine
    |  similarity using an average of word vectors.

+aside-code("Example").
    apples = nlp(u'I like apples')
    oranges = nlp(u'I like oranges')
    apples_oranges = apples.similarity(oranges)
    oranges_apples = oranges.similarity(apples)
    assert apples_oranges == oranges_apples

+table(["Name", "Type", "Description"])
    +row
        +cell #[code other]
        +cell -
        +cell
            |  The object to compare with. By default, accepts #[code Doc],
            |  #[code Span], #[code Token] and #[code Lexeme] objects.

    +footrow
        +cell returns
        +cell float
        +cell A scalar similarity score. Higher is more similar.

+h(2, "count_by") Doc.count_by
    +tag method

p
    |  Count the frequencies of a given attribute. Produces a dict of
    |  #[code {attr (int): count (ints)}] frequencies, keyed by the values
    |  of the given attribute ID.

+aside-code("Example").
    from spacy.attrs import ORTH
    doc = nlp(u'apple apple orange banana')
    assert doc.count_by(ORTH) == {7024L: 1, 119552L: 1, 2087L: 2}
    doc.to_array([attrs.ORTH])
    # array([[11880], [11880], [7561], [12800]])

+table(["Name", "Type", "Description"])
    +row
        +cell #[code attr_id]
        +cell int
        +cell The attribute ID

    +footrow
        +cell returns
        +cell dict
        +cell A dictionary mapping attributes to integer counts.

+h(2, "to_array") Doc.to_array
    +tag method

p
    |  Export the document annotations to a numpy array of shape #[code N*M]
    |  where #[code N] is the length of the document and #[code M] is the number
    |  of attribute IDs to export. The values will be 32-bit integers.

+aside-code("Example").
    from spacy.attrs import LOWER, POS, ENT_TYPE, IS_ALPHA
    doc = nlp(text)
    # All strings mapped to integers, for easy export to numpy
    np_array = doc.to_array([LOWER, POS, ENT_TYPE, IS_ALPHA])

+table(["Name", "Type", "Description"])
    +row
        +cell #[code attr_ids]
        +cell ints
        +cell A list of attribute ID ints.

    +footrow
        +cell returns
        +cell #[code numpy.ndarray[ndim=2, dtype='int32']]
        +cell
            |  The exported attributes as a 2D numpy array, with one row per
            |  token and one column per attribute.

+h(2, "from_array") Doc.from_array
    +tag method

p
    |  Load attributes from a numpy array. Write to a #[code Doc] object, from
    |  an #[code (M, N)] array of attributes.

+aside-code("Example").
    from spacy.attrs import LOWER, POS, ENT_TYPE, IS_ALPHA
    from spacy.tokens import Doc
    doc = nlp(text)
    np_array = doc.to_array([LOWER, POS, ENT_TYPE, IS_ALPHA])
    doc2 = Doc(doc.vocab)
    doc2.from_array([LOWER, POS, ENT_TYPE, IS_ALPHA], np_array)
    assert doc.text == doc2.text

+table(["Name", "Type", "Description"])
    +row
        +cell #[code attrs]
        +cell ints
        +cell A list of attribute ID ints.

    +row
        +cell #[code array]
        +cell #[code numpy.ndarray[ndim=2, dtype='int32']]
        +cell The attribute values to load.

    +footrow
        +cell returns
        +cell #[code Doc]
        +cell Itself.

+h(2, "to_bytes") Doc.to_bytes
    +tag method

p Serialize, i.e. export the document contents to a binary string.

+aside-code("Example").
    doc = nlp(u'Give it back! He pleaded.')
    doc_bytes = doc.to_bytes()

+table(["Name", "Type", "Description"])
    +footrow
        +cell returns
        +cell bytes
        +cell
            |  A losslessly serialized copy of the #[code Doc], including all
            |  annotations.

+h(2, "from_bytes") Doc.from_bytes
    +tag method

p Deserialize, i.e. import the document contents from a binary string.

+aside-code("Example").
    from spacy.tokens import Doc
    text = u'Give it back! He pleaded.'
    doc = nlp(text)
    bytes = doc.to_bytes()
    doc2 = Doc(doc.vocab).from_bytes(bytes)
    assert doc.text == doc2.text

+table(["Name", "Type", "Description"])
    +row
        +cell #[code data]
        +cell bytes
        +cell The string to load from.

    +footrow
        +cell returns
        +cell #[code Doc]
        +cell The #[code Doc] object.

+h(2, "merge") Doc.merge
    +tag method

p
    |  Retokenize the document, such that the span at
    |  #[code doc.text[start_idx : end_idx]] is merged into a single token. If
    |  #[code start_idx] and #[end_idx] do not mark start and end token
    |  boundaries, the document remains unchanged.

+aside-code("Example").
    doc = nlp(u'Los Angeles start.')
    doc.merge(0, len('Los Angeles'), 'NNP', 'Los Angeles', 'GPE')
    assert [t.text for t in doc] == [u'Los Angeles', u'start', u'.']

+table(["Name", "Type", "Description"])
    +row
        +cell #[code start_idx]
        +cell int
        +cell The character index of the start of the slice to merge.

    +row
        +cell #[code end_idx]
        +cell int
        +cell The character index after the end of the slice to merge.

    +row
        +cell #[code **attributes]
        +cell -
        +cell
            |  Attributes to assign to the merged token. By default,
            |  attributes are inherited from the syntactic root token of
            |  the span.

    +footrow
        +cell returns
        +cell #[code Token]
        +cell
            |  The newly merged token, or #[code None] if the start and end
            |  indices did not fall at token boundaries

+h(2, "print_tree") Doc.print_tree
    +tag method
    +tag-model("parse")

p
    |  Returns the parse trees in JSON (dict) format. Especially useful for
    |  web applications.

+aside-code("Example").
    doc = nlp('Alice ate the pizza.')
    trees = doc.print_tree()
    # {'modifiers': [
    #   {'modifiers': [], 'NE': 'PERSON', 'word': 'Alice', 'arc': 'nsubj', 'POS_coarse': 'PROPN', 'POS_fine': 'NNP', 'lemma': 'Alice'},
    #   {'modifiers': [{'modifiers': [], 'NE': '', 'word': 'the', 'arc': 'det', 'POS_coarse': 'DET', 'POS_fine': 'DT', 'lemma': 'the'}], 'NE': '', 'word': 'pizza', 'arc': 'dobj', 'POS_coarse': 'NOUN', 'POS_fine': 'NN', 'lemma': 'pizza'},
    #   {'modifiers': [], 'NE': '', 'word': '.', 'arc': 'punct', 'POS_coarse': 'PUNCT', 'POS_fine': '.', 'lemma': '.'}
    # ], 'NE': '', 'word': 'ate', 'arc': 'ROOT', 'POS_coarse': 'VERB', 'POS_fine': 'VBD', 'lemma': 'eat'}

+table(["Name", "Type", "Description"])
    +row
        +cell #[code light]
        +cell bool
        +cell Don't include lemmas or entities.

    +row
        +cell #[code flat]
        +cell bool
        +cell Don't include arcs or modifiers.

    +footrow
        +cell returns
        +cell dict
        +cell Parse tree as dict.

+h(2, "ents") Doc.ents
    +tag property
    +tag-model("NER")

p
    |  Iterate over the entities in the document. Yields named-entity
    |  #[code Span] objects, if the entity recognizer has been applied to the
    |  document.

+aside-code("Example").
    tokens = nlp(u'Mr. Best flew to New York on Saturday morning.')
    ents = list(tokens.ents)
    assert ents[0].label == 346
    assert ents[0].label_ == 'PERSON'
    assert ents[0].text == 'Mr. Best'

+table(["Name", "Type", "Description"])
    +footrow
        +cell yields
        +cell #[code Span]
        +cell Entities in the document.

+h(2, "noun_chunks") Doc.noun_chunks
    +tag property
    +tag-model("parse")

p
    |  Iterate over the base noun phrases in the document. Yields base
    |  noun-phrase #[code Span] objects, if the document has been syntactically
    |  parsed. A base noun phrase, or "NP chunk", is a noun phrase that does not
    |  permit other NPs to be nested within it – so no NP-level coordination, no
    |  prepositional phrases, and no relative clauses.

+aside-code("Example").
    doc = nlp(u'A phrase with another phrase occurs.')
    chunks = list(doc.noun_chunks)
    assert chunks[0].text == "A phrase"
    assert chunks[1].text == "another phrase"

+table(["Name", "Type", "Description"])
    +footrow
        +cell yields
        +cell #[code Span]
        +cell Noun chunks in the document.

+h(2, "sents") Doc.sents
    +tag property
    +tag-model("parse")

p
    |  Iterate over the sentences in the document. Sentence spans have no label.
    |  To improve accuracy on informal texts, spaCy calculates sentence boundaries
    |  from the syntactic dependency parse. If the parser is disabled,
    |  the #[code sents] iterator will be unavailable.

+aside-code("Example").
    doc = nlp(u"This is a sentence. Here's another...")
    sents = list(doc.sents)
    assert len(sents) == 2
    assert [s.root.text for s in sents] == ["is", "'s"]

+table(["Name", "Type", "Description"])
    +footrow
        +cell yields
        +cell #[code Span]
        +cell Sentences in the document.

+h(2, "has_vector") Doc.has_vector
    +tag property
    +tag-model("vectors")

p
    |  A boolean value indicating whether a word vector is associated with the
    |  object.

+aside-code("Example").
    doc = nlp(u'I like apples')
    assert doc.has_vector

+table(["Name", "Type", "Description"])
    +footrow
        +cell returns
        +cell bool
        +cell Whether the document has a vector data attached.

+h(2, "vector") Doc.vector
    +tag property
    +tag-model("vectors")

p
    |  A real-valued meaning representation. Defaults to an average of the
    |  token vectors.

+aside-code("Example").
    apples = nlp(u'I like apples')
    assert doc.vector.dtype == 'float32'
    assert doc.vector.shape == (300,)

+table(["Name", "Type", "Description"])
    +footrow
        +cell returns
        +cell #[code numpy.ndarray[ndim=1, dtype='float32']]
        +cell A 1D numpy array representing the document's semantics.

+h(2, "vector_norm") Doc.vector_norm
    +tag property
    +tag-model("vectors")

p
    |  The L2 norm of the document's vector representation.

+aside-code("Example").
    doc1 = nlp(u'I like apples')
    doc2 = nlp(u'I like oranges')
    doc1.vector_norm # 4.54232424414368
    doc2.vector_norm # 3.304373298575751
    assert doc1.vector_norm != doc2.vector_norm

+table(["Name", "Type", "Description"])
    +footrow
        +cell returns
        +cell float
        +cell The L2 norm of the vector representation.

+h(2, "attributes") Attributes

+table(["Name", "Type", "Description"])
    +row
        +cell #[code text]
        +cell unicode
        +cell A unicode representation of the document text.

    +row
        +cell #[code text_with_ws]
        +cell unicode
        +cell
            |  An alias of #[code Doc.text], provided for duck-type compatibility
            |  with #[code Span] and #[code Token].

    +row
        +cell #[code mem]
        +cell #[code Pool]
        +cell The document's local memory heap, for all C data it owns.

    +row
        +cell #[code vocab]
        +cell #[code Vocab]
        +cell The store of lexical types.

    +row
        +cell #[code tensor]
        +cell object
        +cell Container for dense vector representations.

    +row
        +cell #[code user_data]
        +cell -
        +cell A generic storage area, for user custom data.

    +row
        +cell #[code is_tagged]
        +cell bool
        +cell
            |  A flag indicating that the document has been part-of-speech
            |  tagged.

    +row
        +cell #[code is_parsed]
        +cell bool
        +cell A flag indicating that the document has been syntactically parsed.

    +row
        +cell #[code sentiment]
        +cell float
        +cell The document's positivity/negativity score, if available.

    +row
        +cell #[code user_hooks]
        +cell dict
        +cell
            |  A dictionary that allows customisation of the #[code Doc]'s
            |  properties.

    +row
        +cell #[code user_token_hooks]
        +cell dict
        +cell
            |  A dictionary that allows customisation of properties of
            |  #[code Token] children.

    +row
        +cell #[code user_span_hooks]
        +cell dict
        +cell
            |  A dictionary that allows customisation of properties of
            |  #[code Span] children.