spaCy/website/docs/usage/entity-recognition.jade

//- 💫 DOCS > USAGE > NAMED ENTITY RECOGNITION

include ../../_includes/_mixins

p
    |  spaCy features an extremely fast statistical entity recognition system,
    |  that assigns labels to contiguous spans of tokens. The default model
    |  identifies a variety of named and numeric entities, including companies,
    |  locations, organizations and products. You can add arbitrary classes to
    |  the entity recognition system, and update the model with new examples.

+aside-code("Example").
    import spacy
    nlp = spacy.load('en')
    doc = nlp(u'London is a big city in the United Kingdom.')
    for ent in doc.ents:
        print(ent.label_, ent.text)
        # GPE London
        # GPE United Kingdom

p
    |  The standard way to access entity annotations is the
    |  #[+api("doc#ents") #[code doc.ents]] property, which produces a sequence
    |  of #[+api("span") #[code Span]] objects. The entity type is accessible
    |  either as an integer ID or as a string, using the attributes
    |  #[code ent.label] and #[code ent.label_]. The #[code Span] object acts
    |  as a sequence of tokens, so you can iterate over the entity or index into
    |  it. You can also get the text form of the whole entity, as though it were
    |  a single token. See the #[+api("span") API reference] for more details.

p
    |  You can access token entity annotations using the #[code token.ent_iob]
    |  and #[code token.ent_type] attributes. The #[code token.ent_iob]
    |  attribute indicates whether an entity starts, continues or ends on the
    |  tag (In, Begin, Out).

+code("Example").
    doc = nlp(u'London is a big city in the United Kingdom.')
    print(doc[0].text, doc[0].ent_iob, doc[0].ent_type_)
    # (u'London', 2, u'GPE')
    print(doc[1].text, doc[1].ent_iob, doc[1].ent_type_)
    # (u'is', 3, u'')

+h(2, "setting") Setting entity annotations

p
    |  To ensure that the sequence of token annotations remains consistent, you
    |  have to set entity annotations at the document level — you can't write
    |  directly to the #[code token.ent_iob] or #[code token.ent_type]
    |  attributes. The easiest way to set entities is to assign to the
    |  #[code doc.ents] attribute.

+code("Example").
    doc = nlp(u'London is a big city in the United Kingdom.')
    doc.ents = []
    assert doc[0].ent_type_ == ''
    doc.ents = [Span(doc, 0, 1, label=doc.vocab.strings['GPE'])]
    assert doc[0].ent_type_ == 'GPE'
    doc.ents = []
    doc.ents = [(u'LondonCity', doc.vocab.strings['GPE']), 0, 1)]

p
    |  The value you assign should be a sequence, the values of which
    |  can either be #[code Span] objects, or #[code (ent_id, ent_type, start, end)]
    |  tuples, where #[code start] and #[code end] are token offsets that
    |  describe the slice of the document that should be annotated.

p
    |  You can also assign entity annotations using the #[code doc.from_array()]
    |  method. To do this, you should include both the #[code ENT_TYPE] and the
    |  #[code ENT_IOB] attributes in the array you're importing from.

+code("Example").
    from spacy.attrs import ENT_IOB, ENT_TYPE
    import numpy

    doc = nlp.make_doc(u'London is a big city in the United Kingdom.')
    assert list(doc.ents) == []
    header = [ENT_IOB, ENT_TYPE]
    attr_array = numpy.zeros((len(doc), len(header)))
    attr_array[0, 0] = 2 # B
    attr_array[0, 1] = doc.vocab.strings[u'GPE']
    doc.from_array(header, attr_array)
    assert list(doc.ents)[0].text == u'London'

p
    |  Finally, you can always write to the underlying struct, if you compile
    |  a Cython function. This is easy to do, and allows you to write efficient
    |  native code.

+code("Example").
    # cython: infer_types=True
    from spacy.tokens.doc cimport Doc

    cpdef set_entity(Doc doc, int start, int end, int ent_type):
        for i in range(start, end):
            doc.c[i].ent_type = ent_type
        doc.c[start].ent_iob = 3
        for i in range(start+1, end):
            doc.c[i].ent_iob = 2

p
    |  Obviously, if you write directly to the array of #[code TokenC*] structs,
    |  you'll have responsibility for ensuring that the data is left in a
    |  consistent state.


+h(2, "displacy") The displaCy #[sup ENT] visualizer

p
    |  The #[+a(DEMOS_URL + "/displacy-ent/") displaCy #[sup ENT] visualizer]
    |  lets you explore an entity recognition model's behaviour interactively.
    |  If you're training a model, it's very useful to run the visualization
    |  server yourself. To help you do that, we've open-sourced both the
    |  #[+a(gh("spacy-services")) back-end service] and the
    |  #[+a(gh("displacy-ent")) front-end client].

+codepen("ALxpQO", 450)

+h(2, "entity-types") Built-in entity types

include ../api/_annotation/_named-entities

+aside("Install")
    |  The #[+api("load") spacy.load()] function configures a pipeline that
    |  includes all of the available annotators for the given ID. In the example
    |  above, the #[code 'en'] ID tells spaCy to load the default English
    |  pipeline. If you have installed the data with
    |  #[code python -m spacy.en.download] this will include the entity
    |  recognition model.

+h(2, "updating") Training and updating

p
    |  To provide training examples to the entity recogniser, you'll first need
    |  to create an instance of the #[code GoldParse] class. You can specify
    |  your annotations in a stand-off format or as token tags.

+code.
    import spacy
    from spacy.gold import GoldParse

    train_data = [
        ('Who is Chaka Khan?', [(7, 17, 'PERSON')]),
        ('I like London and Berlin.', [(7, 13, 'LOC'), (18, 24, 'LOC')])
    ]

    nlp = spacy.load('en', entity=False, parser=False)
    ner = EntityRecognizer(nlp.vocab, entity_types=['PERSON', 'LOC'])

    for itn in range(5):
        random.shuffle(train_data)
        for raw_text, entity_offsets in train_data:
            doc = nlp.make_doc(raw_text)
            gold = GoldParse(doc, entities=entity_offsets)

            nlp.tagger(doc)
            ner.update(doc, gold)
    ner.model.end_training()

p
    |  If a character offset in your entity annotations don't fall on a token
    |  boundary, the #[code GoldParse] class will treat that annotation as a
    |  missing value.  This allows for more realistic training, because the
    |  entity recogniser is allowed to learn from examples that may feature
    |  tokenizer errors.

+aside-code("Example").
    doc = Doc(nlp.vocab, [u'rats', u'make', u'good', u'pets'])
    gold = GoldParse(doc, [u'U-ANIMAL', u'O', u'O', u'O'])
    ner = EntityRecognizer(nlp.vocab, entity_types=['ANIMAL'])
    ner.update(doc, gold)

p
    |  You can also provide token-level entity annotation, using the
    |  following tagging scheme to describe the entity boundaries:

+table([ "Tag", "Description" ])
    +row
        +cell #[code #[span.u-color-theme B] EGIN]
        +cell The first token of a multi-token entity.

    +row
        +cell #[code #[span.u-color-theme I] N]
        +cell An inner token of a multi-token entity.

    +row
        +cell #[code #[span.u-color-theme L] AST]
        +cell The final token of a multi-token entity.

    +row
        +cell #[code #[span.u-color-theme U] NIT]
        +cell A single-token entity.

    +row
        +cell #[code #[span.u-color-theme O] UT]
        +cell A non-entity token.

+aside("Why BILUO, not IOB?")
    |  There are several coding schemes for encoding entity annotations as
    |  token tags.  These coding schemes are equally expressive, but not
    |  necessarily equally learnable.
    |  #[+a("http://www.aclweb.org/anthology/W09-1119") Ratinov and Roth]
    |  showed that the minimal #[strong Begin], #[strong In], #[strong Out]
    |  scheme was more difficult to learn than the #[strong BILUO] scheme that
    |  we use, which explicitly marks boundary tokens.

p
    |  spaCy translates the character offsets into this scheme, in order to
    |  decide the cost of each action given the current state of the entity
    |  recogniser. The costs are then used to calculate the gradient of the
    |  loss, to train the model. The exact algorithm is a pastiche of
    |  well-known methods, and is not currently described in any single
    |  publication. The model is a greedy transition-based parser guided by a
    |  linear model whose weights are learned using the averaged perceptron
    |  loss, via the #[+a("http://www.aclweb.org/anthology/C12-1059") dynamic oracle]
    |  imitation learning strategy. The transition system is equivalent to the
    |  BILOU tagging scheme.