spaCy/spacy/tokens/doc.pyx

# coding: utf8
# cython: infer_types=True
# cython: bounds_check=False
# cython: profile=True
from __future__ import unicode_literals

cimport cython
cimport numpy as np
import numpy
import numpy.linalg
import struct
import dill
import msgpack
from thinc.neural.util import get_array_module, copy_array

from libc.string cimport memcpy, memset
from libc.math cimport sqrt

from .span cimport Span
from .token cimport Token
from .span cimport Span
from .token cimport Token
from .printers import parse_tree
from ..lexeme cimport Lexeme, EMPTY_LEXEME
from ..typedefs cimport attr_t, flags_t
from ..attrs import intify_attrs, IDS
from ..attrs cimport attr_id_t
from ..attrs cimport ID, ORTH, NORM, LOWER, SHAPE, PREFIX, SUFFIX, CLUSTER
from ..attrs cimport LENGTH, POS, LEMMA, TAG, DEP, HEAD, SPACY, ENT_IOB
from ..attrs cimport ENT_TYPE, SENT_START
from ..parts_of_speech cimport CCONJ, PUNCT, NOUN, univ_pos_t
from ..util import normalize_slice
from ..compat import is_config, copy_reg, pickle, basestring_
from .. import about
from .. import util
from .underscore import Underscore

DEF PADDING = 5


cdef int bounds_check(int i, int length, int padding) except -1:
    if (i + padding) < 0:
        raise IndexError
    if (i - padding) >= length:
        raise IndexError


cdef attr_t get_token_attr(const TokenC* token, attr_id_t feat_name) nogil:
    if feat_name == LEMMA:
        return token.lemma
    elif feat_name == POS:
        return token.pos
    elif feat_name == TAG:
        return token.tag
    elif feat_name == DEP:
        return token.dep
    elif feat_name == HEAD:
        return token.head
    elif feat_name == SENT_START:
        return token.sent_start
    elif feat_name == SPACY:
        return token.spacy
    elif feat_name == ENT_IOB:
        return token.ent_iob
    elif feat_name == ENT_TYPE:
        return token.ent_type
    else:
        return Lexeme.get_struct_attr(token.lex, feat_name)


def _get_chunker(lang):
    try:
        cls = util.get_lang_class(lang)
    except ImportError:
        return None
    except KeyError:
        return None
    return cls.Defaults.syntax_iterators.get(u'noun_chunks')


cdef class Doc:
    """A sequence of Token objects. Access sentences and named entities, export
    annotations to numpy arrays, losslessly serialize to compressed binary
    strings. The `Doc` object holds an array of `TokenC` structs. The
    Python-level `Token` and `Span` objects are views of this array, i.e.
    they don't own the data themselves.

    EXAMPLE: Construction 1
        >>> doc = nlp(u'Some text')

        Construction 2
        >>> from spacy.tokens import Doc
        >>> doc = Doc(nlp.vocab, words=[u'hello', u'world', u'!'],
                      spaces=[True, False, False])
    """
    @classmethod
    def set_extension(cls, name, default=None, method=None,
                      getter=None, setter=None):
        nr_defined = sum(t is not None for t in (default, getter, setter, method))
        assert nr_defined == 1
        Underscore.doc_extensions[name] = (default, method, getter, setter)

    @classmethod
    def get_extension(cls, name):
        return Underscore.doc_extensions.get(name)

    @classmethod
    def has_extension(cls, name):
        return name in Underscore.doc_extensions

    def __init__(self, Vocab vocab, words=None, spaces=None, user_data=None,
                 orths_and_spaces=None):
        """Create a Doc object.

        vocab (Vocab): A vocabulary object, which must match any models you
            want to use (e.g. tokenizer, parser, entity recognizer).
        words (list or None): A list of unicode strings to add to the document
            as words. If `None`, defaults to empty list.
        spaces (list or None): A list of boolean values, of the same length as
            words. True means that the word is followed by a space, False means
            it is not. If `None`, defaults to `[True]*len(words)`
        user_data (dict or None): Optional extra data to attach to the Doc.
        RETURNS (Doc): The newly constructed object.
        """
        self.vocab = vocab
        size = 20
        self.mem = Pool()
        # Guarantee self.lex[i-x], for any i >= 0 and x < padding is in bounds
        # However, we need to remember the true starting places, so that we can
        # realloc.
        data_start = <TokenC*>self.mem.alloc(size + (PADDING*2), sizeof(TokenC))
        cdef int i
        for i in range(size + (PADDING*2)):
            data_start[i].lex = &EMPTY_LEXEME
            data_start[i].l_edge = i
            data_start[i].r_edge = i
        self.c = data_start + PADDING
        self.max_length = size
        self.length = 0
        self.is_tagged = False
        self.is_parsed = False
        self.sentiment = 0.0
        self.cats = {}
        self.user_hooks = {}
        self.user_token_hooks = {}
        self.user_span_hooks = {}
        self.tensor = numpy.zeros((0,), dtype='float32')
        self.user_data = {} if user_data is None else user_data
        self._vector = None
        self.noun_chunks_iterator = _get_chunker(self.vocab.lang)
        cdef unicode orth
        cdef bint has_space
        if orths_and_spaces is None and words is not None:
            if spaces is None:
                spaces = [True] * len(words)
            elif len(spaces) != len(words):
                raise ValueError(
                    "Arguments 'words' and 'spaces' should be sequences of "
                    "the same length, or 'spaces' should be left default at "
                    "None. spaces should be a sequence of booleans, with True "
                    "meaning that the word owns a ' ' character following it.")
            orths_and_spaces = zip(words, spaces)
        if orths_and_spaces is not None:
            for orth_space in orths_and_spaces:
                if isinstance(orth_space, unicode):
                    orth = orth_space
                    has_space = True
                elif isinstance(orth_space, bytes):
                    raise ValueError(
                        "orths_and_spaces expects either List(unicode) or "
                        "List((unicode, bool)). "
                        "Got bytes instance: %s" % (str(orth_space)))
                else:
                    orth, has_space = orth_space
                # Note that we pass self.mem here --- we have ownership, if LexemeC
                # must be created.
                self.push_back(
                    <const LexemeC*>self.vocab.get(self.mem, orth), has_space)
        # Tough to decide on policy for this. Is an empty doc tagged and parsed?
        # There's no information we'd like to add to it, so I guess so?
        if self.length == 0:
            self.is_tagged = True
            self.is_parsed = True

    @property
    def _(self):
        return Underscore(Underscore.doc_extensions, self)

    def __getitem__(self, object i):
        """Get a `Token` or `Span` object.

        i (int or tuple) The index of the token, or the slice of the document
            to get.
        RETURNS (Token or Span): The token at `doc[i]]`, or the span at
            `doc[start : end]`.

        EXAMPLE:
            >>> doc[i]
            Get the `Token` object at position `i`, where `i` is an integer.
            Negative indexing is supported, and follows the usual Python
            semantics, i.e. `doc[-2]` is `doc[len(doc) - 2]`.

            >>> doc[start : end]]
            Get a `Span` object, starting at position `start` and ending at
            position `end`, where `start` and `end` are token indices. For
            instance, `doc[2:5]` produces a span consisting of tokens 2, 3 and
            4. Stepped slices (e.g. `doc[start : end : step]`) are not
            supported, as `Span` objects must be contiguous (cannot have gaps).
            You can use negative indices and open-ended ranges, which have
            their normal Python semantics.
        """
        if isinstance(i, slice):
            start, stop = normalize_slice(len(self), i.start, i.stop, i.step)
            return Span(self, start, stop, label=0)

        if i < 0:
            i = self.length + i
        bounds_check(i, self.length, PADDING)
        return Token.cinit(self.vocab, &self.c[i], i, self)

    def __iter__(self):
        """Iterate over `Token`  objects, from which the annotations can be
        easily accessed. This is the main way of accessing `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.

        EXAMPLE:
            >>> for token in doc
        """
        cdef int i
        for i in range(self.length):
            yield Token.cinit(self.vocab, &self.c[i], i, self)

    def __len__(self):
        """The number of tokens in the document.

        RETURNS (int): The number of tokens in the document.

        EXAMPLE:
            >>> len(doc)
        """
        return self.length

    def __unicode__(self):
        return u''.join([t.text_with_ws for t in self])

    def __bytes__(self):
        return u''.join([t.text_with_ws for t in self]).encode('utf-8')

    def __str__(self):
        if is_config(python3=True):
            return self.__unicode__()
        return self.__bytes__()

    def __repr__(self):
        return self.__str__()

    @property
    def doc(self):
        return self

    def char_span(self, int start_idx, int end_idx, label=0, vector=None):
        """Create a `Span` object from the slice `doc.text[start : end]`.

        doc (Doc): The parent document.
        start (int): The index of the first character of the span.
        end (int): The index of the first character after the span.
        label (uint64 or string): A label to attach to the Span, e.g. for
            named entities.
        vector (ndarray[ndim=1, dtype='float32']): A meaning representation of
            the span.
        RETURNS (Span): The newly constructed object.
        """
        if not isinstance(label, int):
            label = self.vocab.strings.add(label)
        cdef int start = token_by_start(self.c, self.length, start_idx)
        if start == -1:
            return None
        cdef int end = token_by_end(self.c, self.length, end_idx)
        if end == -1:
            return None
        # Currently we have the token index, we want the range-end index
        end += 1
        cdef Span span = Span(self, start, end, label=label, vector=vector)
        return span

    def similarity(self, other):
        """Make a semantic similarity estimate. The default estimate is cosine
        similarity using an average of word vectors.

        other (object): The object to compare with. By default, accepts `Doc`,
            `Span`, `Token` and `Lexeme` objects.
        RETURNS (float): A scalar similarity score. Higher is more similar.
        """
        if 'similarity' in self.user_hooks:
            return self.user_hooks['similarity'](self, other)
        if isinstance(other, (Lexeme, Token)) and self.length == 1:
            if self.c[0].lex.orth == other.orth:
                return 1.0
        elif isinstance(other, (Span, Doc)):
            if len(self) == len(other):
                for i in range(self.length):
                    if self[i].orth != other[i].orth:
                        break
                else:
                    return 1.0
 
        if self.vector_norm == 0 or other.vector_norm == 0:
            return 0.0
        return numpy.dot(self.vector, other.vector) / (self.vector_norm * other.vector_norm)

    property has_vector:
        """A boolean value indicating whether a word vector is associated with
        the object.

        RETURNS (bool): Whether a word vector is associated with the object.
        """
        def __get__(self):
            if 'has_vector' in self.user_hooks:
                return self.user_hooks['has_vector'](self)
            elif self.vocab.vectors.data.size:
                return True
            elif self.tensor.size:
                return True
            else:
                return False

    property vector:
        """A real-valued meaning representation. Defaults to an average of the
        token vectors.

        RETURNS (numpy.ndarray[ndim=1, dtype='float32']): A 1D numpy array
            representing the document's semantics.
        """
        def __get__(self):
            if 'vector' in self.user_hooks:
                return self.user_hooks['vector'](self)
            if self._vector is not None:
                return self._vector
            elif not len(self):
                self._vector = numpy.zeros((self.vocab.vectors_length,),
                                           dtype='f')
                return self._vector
            elif self.vocab.vectors.data.size > 0:
                vector = numpy.zeros((self.vocab.vectors_length,), dtype='f')
                for token in self.c[:self.length]:
                    vector += self.vocab.get_vector(token.lex.orth)
                self._vector = vector / len(self)
                return self._vector
            elif self.tensor.size > 0:
                self._vector = self.tensor.mean(axis=0)
                return self._vector
            else:
                return numpy.zeros((self.vocab.vectors_length,),
                                   dtype='float32')

        def __set__(self, value):
            self._vector = value

    property vector_norm:
        """The L2 norm of the document's vector representation.

        RETURNS (float): The L2 norm of the vector representation.
        """
        def __get__(self):
            if 'vector_norm' in self.user_hooks:
                return self.user_hooks['vector_norm'](self)
            cdef float value
            cdef double norm = 0
            if self._vector_norm is None:
                norm = 0.0
                for value in self.vector:
                    norm += value * value
                self._vector_norm = sqrt(norm) if norm != 0 else 0
            return self._vector_norm

        def __set__(self, value):
            self._vector_norm = value

    property text:
        """A unicode representation of the document text.

        RETURNS (unicode): The original verbatim text of the document.
        """
        def __get__(self):
            return u''.join(t.text_with_ws for t in self)

    property text_with_ws:
        """An alias of `Doc.text`, provided for duck-type compatibility with
        `Span` and `Token`.

        RETURNS (unicode): The original verbatim text of the document.
        """
        def __get__(self):
            return self.text

    property ents:
        """Iterate over the entities in the document. Yields named-entity
        `Span` objects, if the entity recognizer has been applied to the
        document.

        YIELDS (Span): Entities in the document.

        EXAMPLE: Iterate over the span to get individual Token objects,
            or access the label:

            >>> 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].orth_ == 'Best'
            >>> assert ents[0].text == 'Mr. Best'
        """
        def __get__(self):
            cdef int i
            cdef const TokenC* token
            cdef int start = -1
            cdef attr_t label = 0
            output = []
            for i in range(self.length):
                token = &self.c[i]
                if token.ent_iob == 1:
                    assert start != -1
                elif token.ent_iob == 2 or token.ent_iob == 0:
                    if start != -1:
                        output.append(Span(self, start, i, label=label))
                    start = -1
                    label = 0
                elif token.ent_iob == 3:
                    if start != -1:
                        output.append(Span(self, start, i, label=label))
                    start = i
                    label = token.ent_type
            if start != -1:
                output.append(Span(self, start, self.length, label=label))
            return tuple(output)

        def __set__(self, ents):
            # TODO:
            # 1. Allow negative matches
            # 2. Ensure pre-set NERs are not over-written during statistical
            #    prediction
            # 3. Test basic data-driven ORTH gazetteer
            # 4. Test more nuanced date and currency regex
            cdef int i
            for i in range(self.length):
                self.c[i].ent_type = 0
                # At this point we don't know whether the NER has run over the
                # Doc. If the ent_iob is missing, leave it missing.
                if self.c[i].ent_iob != 0:
                    self.c[i].ent_iob = 2  # Means O. Non-O are set from ents.
            cdef attr_t ent_type
            cdef int start, end
            for ent_info in ents:
                if isinstance(ent_info, Span):
                    ent_id = ent_info.ent_id
                    ent_type = ent_info.label
                    start = ent_info.start
                    end = ent_info.end
                elif len(ent_info) == 3:
                    ent_type, start, end = ent_info
                else:
                    ent_id, ent_type, start, end = ent_info
                if ent_type is None or ent_type < 0:
                    # Mark as O
                    for i in range(start, end):
                        self.c[i].ent_type = 0
                        self.c[i].ent_iob = 2
                else:
                    # Mark (inside) as I
                    for i in range(start, end):
                        self.c[i].ent_type = ent_type
                        self.c[i].ent_iob = 1
                    # Set start as B
                    self.c[start].ent_iob = 3

    property noun_chunks:
        """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.

        YIELDS (Span): Noun chunks in the document.
        """
        def __get__(self):
            if not self.is_parsed:
                raise ValueError(
                    "noun_chunks requires the dependency parse, which "
                    "requires a statistical model to be installed and loaded. "
                    "For more info, see the "
                    "documentation: \n%s\n" % about.__docs_models__)
            # Accumulate the result before beginning to iterate over it. This
            # prevents the tokenisation from being changed out from under us
            # during the iteration. The tricky thing here is that Span accepts
            # its tokenisation changing, so it's okay once we have the Span
            # objects. See Issue #375.
            spans = []
            for start, end, label in self.noun_chunks_iterator(self):
                spans.append(Span(self, start, end, label=label))
            for span in spans:
                yield span

    property sents:
        """Iterate over the sentences in the document. Yields sentence `Span`
        objects. 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 `sents` iterator will
        be unavailable.

        EXAMPLE:
            >>> doc = nlp("This is a sentence. Here's another...")
            >>> assert [s.root.text for s in doc.sents] == ["is", "'s"]
        """
        def __get__(self):
            if 'sents' in self.user_hooks:
                yield from self.user_hooks['sents'](self)
                return

            cdef int i
            if not self.is_parsed:
                for i in range(1, self.length):
                    if self.c[i].sent_start != 0:
                        break
                else:
                    raise ValueError(
                        "Sentence boundaries unset. You can add the 'sentencizer' "
                        "component to the pipeline with: "
                        "nlp.add_pipe(nlp.create_pipe('sentencizer')) "
                        "Alternatively, add the dependency parser, or set "
                        "sentence boundaries by setting doc[i].sent_start")
            start = 0
            for i in range(1, self.length):
                if self.c[i].sent_start == 1:
                    yield Span(self, start, i)
                    start = i
            if start != self.length:
                yield Span(self, start, self.length)

    cdef int push_back(self, LexemeOrToken lex_or_tok, bint has_space) except -1:
        if self.length == 0:
            # Flip these to false when we see the first token.
            self.is_tagged = False
            self.is_parsed = False
        if self.length == self.max_length:
            self._realloc(self.length * 2)
        cdef TokenC* t = &self.c[self.length]
        if LexemeOrToken is const_TokenC_ptr:
            t[0] = lex_or_tok[0]
        else:
            t.lex = lex_or_tok
        if self.length == 0:
            t.idx = 0
        else:
            t.idx = (t-1).idx + (t-1).lex.length + (t-1).spacy
        t.l_edge = self.length
        t.r_edge = self.length
        assert t.lex.orth != 0
        t.spacy = has_space
        self.length += 1
        return t.idx + t.lex.length + t.spacy

    @cython.boundscheck(False)
    cpdef np.ndarray to_array(self, object py_attr_ids):
        """Export given token attributes to a numpy `ndarray`.
        If `attr_ids` is a sequence of M attributes, the output array will be
        of shape `(N, M)`, where N is the length of the `Doc` (in tokens). If
        `attr_ids` is a single attribute, the output shape will be (N,). You
        can specify attributes by integer ID (e.g. spacy.attrs.LEMMA) or
        string name (e.g. 'LEMMA' or 'lemma').

        attr_ids (list[]): A list of attributes (int IDs or string names).
        RETURNS (numpy.ndarray[long, ndim=2]): A feature matrix, with one row
            per word, and one column per attribute indicated in the input
            `attr_ids`.

        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])
        """
        cdef int i, j
        cdef attr_id_t feature
        cdef np.ndarray[attr_t, ndim=2] output
        # Handle scalar/list inputs of strings/ints for py_attr_ids
        if not hasattr(py_attr_ids, '__iter__') \
        and not isinstance(py_attr_ids, basestring_):
            py_attr_ids = [py_attr_ids]

        # Allow strings, e.g. 'lemma' or 'LEMMA'
        py_attr_ids = [(IDS[id_.upper()] if hasattr(id_, 'upper') else id_)
                       for id_ in py_attr_ids]
        # Make an array from the attributes --- otherwise our inner loop is
        # Python dict iteration.
        cdef np.ndarray attr_ids = numpy.asarray(py_attr_ids, dtype='i')
        output = numpy.ndarray(shape=(self.length, len(attr_ids)),
                               dtype=numpy.uint64)
        c_output = <attr_t*>output.data
        c_attr_ids = <attr_id_t*>attr_ids.data
        cdef TokenC* token
        cdef int nr_attr = attr_ids.shape[0]
        for i in range(self.length):
            token = &self.c[i]
            for j in range(nr_attr):
                c_output[i*nr_attr + j] = get_token_attr(token, c_attr_ids[j])
        # Handle 1d case
        return output if len(attr_ids) >= 2 else output.reshape((self.length,))

    def count_by(self, attr_id_t attr_id, exclude=None,
                 PreshCounter counts=None):
        """Count the frequencies of a given attribute. Produces a dict of
        `{attribute (int): count (ints)}` frequencies, keyed by the values of
        the given attribute ID.

        attr_id (int): The attribute ID to key the counts.
        RETURNS (dict): A dictionary mapping attributes to integer counts.

        EXAMPLE:
            >>> from spacy import attrs
            >>> doc = nlp(u'apple apple orange banana')
            >>> tokens.count_by(attrs.ORTH)
            {12800L: 1, 11880L: 2, 7561L: 1}
            >>> tokens.to_array([attrs.ORTH])
            array([[11880], [11880], [7561], [12800]])
        """
        cdef int i
        cdef attr_t attr
        cdef size_t count

        if counts is None:
            counts = PreshCounter()
            output_dict = True
        else:
            output_dict = False
        # Take this check out of the loop, for a bit of extra speed
        if exclude is None:
            for i in range(self.length):
                counts.inc(get_token_attr(&self.c[i], attr_id), 1)
        else:
            for i in range(self.length):
                if not exclude(self[i]):
                    attr = get_token_attr(&self.c[i], attr_id)
                    counts.inc(attr, 1)
        if output_dict:
            return dict(counts)

    def _realloc(self, new_size):
        self.max_length = new_size
        n = new_size + (PADDING * 2)
        # What we're storing is a "padded" array. We've jumped forward PADDING
        # places, and are storing the pointer to that. This way, we can access
        # words out-of-bounds, and get out-of-bounds markers.
        # Now that we want to realloc, we need the address of the true start,
        # so we jump the pointer back PADDING places.
        cdef TokenC* data_start = self.c - PADDING
        data_start = <TokenC*>self.mem.realloc(data_start, n * sizeof(TokenC))
        self.c = data_start + PADDING
        cdef int i
        for i in range(self.length, self.max_length + PADDING):
            self.c[i].lex = &EMPTY_LEXEME

    cdef void set_parse(self, const TokenC* parsed) nogil:
        # TODO: This method is fairly misleading atm. It's used by Parser
        # to actually apply the parse calculated. Need to rethink this.

        # Probably we should use from_array?
        self.is_parsed = True
        for i in range(self.length):
            self.c[i] = parsed[i]

    def from_array(self, attrs, array):
        if SENT_START in attrs and HEAD in attrs:
            raise ValueError(
                "Conflicting attributes specified in doc.from_array(): "
                "(HEAD, SENT_START)\n"
                "The HEAD attribute currently sets sentence boundaries "
                "implicitly, based on the tree structure. This means the HEAD "
                "attribute would potentially override the sentence boundaries "
                "set by SENT_START.")
        cdef int i, col
        cdef attr_id_t attr_id
        cdef TokenC* tokens = self.c
        cdef int length = len(array)
        # Get set up for fast loading
        cdef Pool mem = Pool()
        cdef int n_attrs = len(attrs)
        attr_ids = <attr_id_t*>mem.alloc(n_attrs, sizeof(attr_id_t))
        for i, attr_id in enumerate(attrs):
            attr_ids[i] = attr_id
        # Now load the data
        for i in range(self.length):
            token = &self.c[i]
            for j in range(n_attrs):
                Token.set_struct_attr(token, attr_ids[j], array[i, j])
        # Auxiliary loading logic
        for col, attr_id in enumerate(attrs):
            if attr_id == TAG:
                for i in range(length):
                    if array[i, col] != 0:
                        self.vocab.morphology.assign_tag(&tokens[i], array[i, col])
        set_children_from_heads(self.c, self.length)
        self.is_parsed = bool(HEAD in attrs or DEP in attrs)
        self.is_tagged = bool(TAG in attrs or POS in attrs)
        return self

    def get_lca_matrix(self):
        """Calculates the lowest common ancestor matrix for a given `Doc`.
        Returns LCA matrix containing the integer index of the ancestor, or -1
        if no common ancestor is found (ex if span excludes a necessary
        ancestor). Apologies about the recursion, but the impact on
        performance is negligible given the natural limitations on the depth
        of a typical human sentence.
        """
        # Efficiency notes:
        # We can easily improve the performance here by iterating in Cython.
        # To loop over the tokens in Cython, the easiest way is:
        # for token in doc.c[:doc.c.length]:
        #     head = token + token.head
        # Both token and head will be TokenC* here. The token.head attribute
        # is an integer offset.
        def __pairwise_lca(token_j, token_k, lca_matrix):
            if lca_matrix[token_j.i][token_k.i] != -2:
                return lca_matrix[token_j.i][token_k.i]
            elif token_j == token_k:
                lca_index = token_j.i
            elif token_k.head == token_j:
                lca_index = token_j.i
            elif token_j.head == token_k:
                lca_index = token_k.i
            elif (token_j.head == token_j) and (token_k.head == token_k):
                lca_index = -1
            else:
                lca_index = __pairwise_lca(token_j.head, token_k.head,
                                           lca_matrix)
            lca_matrix[token_j.i][token_k.i] = lca_index
            lca_matrix[token_k.i][token_j.i] = lca_index

            return lca_index

        lca_matrix = numpy.empty((len(self), len(self)), dtype=numpy.int32)
        lca_matrix.fill(-2)
        for j in range(len(self)):
            token_j = self[j]
            for k in range(j, len(self)):
                token_k = self[k]
                lca_matrix[j][k] = __pairwise_lca(token_j, token_k, lca_matrix)
                lca_matrix[k][j] = lca_matrix[j][k]
        return lca_matrix

    def to_disk(self, path, **exclude):
        """Save the current state to a directory.

        path (unicode or Path): A path to a directory, which will be created if
            it doesn't exist. Paths may be either strings or Path-like objects.
        """
        path = util.ensure_path(path)
        with path.open('wb') as file_:
            file_.write(self.to_bytes(**exclude))

    def from_disk(self, path, **exclude):
        """Loads state from a directory. Modifies the object in place and
        returns it.

        path (unicode or Path): A path to a directory. Paths may be either
            strings or `Path`-like objects.
        RETURNS (Doc): The modified `Doc` object.
        """
        path = util.ensure_path(path)
        with path.open('rb') as file_:
            bytes_data = file_.read()
        return self.from_bytes(bytes_data, **exclude)

    def to_bytes(self, **exclude):
        """Serialize, i.e. export the document contents to a binary string.

        RETURNS (bytes): A losslessly serialized copy of the `Doc`, including
            all annotations.
        """
        array_head = [LENGTH, SPACY, TAG, LEMMA, HEAD, DEP, ENT_IOB, ENT_TYPE]
        # Msgpack doesn't distinguish between lists and tuples, which is
        # vexing for user data. As a best guess, we *know* that within
        # keys, we must have tuples. In values we just have to hope
        # users don't mind getting a list instead of a tuple.
        serializers = {
            'text': lambda: self.text,
            'array_head': lambda: array_head,
            'array_body': lambda: self.to_array(array_head),
            'sentiment': lambda: self.sentiment,
            'tensor': lambda: self.tensor,
        }
        if 'user_data' not in exclude and self.user_data:
            user_data_keys, user_data_values = list(zip(*self.user_data.items()))
            serializers['user_data_keys'] = lambda: msgpack.dumps(user_data_keys)
            serializers['user_data_values'] = lambda: msgpack.dumps(user_data_values)

        return util.to_bytes(serializers, exclude)

    def from_bytes(self, bytes_data, **exclude):
        """Deserialize, i.e. import the document contents from a binary string.

        data (bytes): The string to load from.
        RETURNS (Doc): Itself.
        """
        if self.length != 0:
            raise ValueError("Cannot load into non-empty Doc")
        deserializers = {
            'text': lambda b: None,
            'array_head': lambda b: None,
            'array_body': lambda b: None,
            'sentiment': lambda b: None,
            'tensor': lambda b: None,
            'user_data_keys': lambda b: None,
            'user_data_values': lambda b: None,
        }

        msg = util.from_bytes(bytes_data, deserializers, exclude)
        # Msgpack doesn't distinguish between lists and tuples, which is
        # vexing for user data. As a best guess, we *know* that within
        # keys, we must have tuples. In values we just have to hope
        # users don't mind getting a list instead of a tuple.
        if 'user_data' not in exclude and 'user_data_keys' in msg:
            user_data_keys = msgpack.loads(msg['user_data_keys'],
                                           use_list=False)
            user_data_values = msgpack.loads(msg['user_data_values'])
            for key, value in zip(user_data_keys, user_data_values):
                self.user_data[key] = value

        cdef attr_t[:, :] attrs
        cdef int i, start, end, has_space
        self.sentiment = msg['sentiment']
        self.tensor = msg['tensor']

        start = 0
        cdef const LexemeC* lex
        cdef unicode orth_
        text = msg['text']
        attrs = msg['array_body']
        for i in range(attrs.shape[0]):
            end = start + attrs[i, 0]
            has_space = attrs[i, 1]
            orth_ = text[start:end]
            lex = self.vocab.get(self.mem, orth_)
            self.push_back(lex, has_space)
            start = end + has_space
        self.from_array(msg['array_head'][2:],
                        attrs[:, 2:])
        return self

    def extend_tensor(self, tensor):
        '''Concatenate a new tensor onto the doc.tensor object.

        The doc.tensor attribute holds dense feature vectors
        computed by the models in the pipeline. Let's say a
        document with 30 words has a tensor with 128 dimensions
        per word. doc.tensor.shape will be (30, 128). After
        calling doc.extend_tensor with an array of hape (30, 64),
        doc.tensor == (30, 192).
        '''
        xp = get_array_module(self.tensor)
        if self.tensor.size == 0:
            self.tensor.resize(tensor.shape)
            copy_array(self.tensor, tensor)
        else:
            self.tensor = xp.hstack((self.tensor, tensor))

    def merge(self, int start_idx, int end_idx, *args, **attributes):
        """Retokenize the document, such that the span at
        `doc.text[start_idx : end_idx]` is merged into a single token. If
        `start_idx` and `end_idx `do not mark start and end token boundaries,
        the document remains unchanged.

        start_idx (int): Character index of the start of the slice to merge.
        end_idx (int): Character index after the end of the slice to merge.
        **attributes: Attributes to assign to the merged token. By default,
            attributes are inherited from the syntactic root of the span.
        RETURNS (Token): The newly merged token, or `None` if the start and end
            indices did not fall at token boundaries.
        """
        cdef unicode tag, lemma, ent_type
        if len(args) == 3:
            util.deprecated(
                "Positional arguments to Doc.merge are deprecated. Instead, "
                "use the keyword arguments, for example tag=, lemma= or "
                "ent_type=.")
            tag, lemma, ent_type = args
            attributes[TAG] = tag
            attributes[LEMMA] = lemma
            attributes[ENT_TYPE] = ent_type
        elif not args:
            if 'label' in attributes and 'ent_type' not in attributes:
                if isinstance(attributes['label'], int):
                    attributes[ENT_TYPE] = attributes['label']
                else:
                    attributes[ENT_TYPE] = self.vocab.strings[attributes['label']]
            if 'ent_type' in attributes:
                attributes[ENT_TYPE] = attributes['ent_type']
        elif args:
            raise ValueError(
                "Doc.merge received %d non-keyword arguments. Expected either "
                "3 arguments (deprecated), or 0 (use keyword arguments). "
                "Arguments supplied:\n%s\n"
                "Keyword arguments: %s\n" % (len(args), repr(args),
                                             repr(attributes)))

        # More deprecated attribute handling =/
        if 'label' in attributes:
            attributes['ent_type'] = attributes.pop('label')

        attributes = intify_attrs(attributes, strings_map=self.vocab.strings)

        cdef int start = token_by_start(self.c, self.length, start_idx)
        if start == -1:
            return None
        cdef int end = token_by_end(self.c, self.length, end_idx)
        if end == -1:
            return None
        # Currently we have the token index, we want the range-end index
        end += 1
        cdef Span span = self[start:end]
        # Get LexemeC for newly merged token
        new_orth = ''.join([t.text_with_ws for t in span])
        if span[-1].whitespace_:
            new_orth = new_orth[:-len(span[-1].whitespace_)]
        cdef const LexemeC* lex = self.vocab.get(self.mem, new_orth)
        # House the new merged token where it starts
        cdef TokenC* token = &self.c[start]
        token.spacy = self.c[end-1].spacy
        for attr_name, attr_value in attributes.items():
            if attr_name == TAG:
                self.vocab.morphology.assign_tag(token, attr_value)
            else:
                Token.set_struct_attr(token, attr_name, attr_value)
        # Begin by setting all the head indices to absolute token positions
        # This is easier to work with for now than the offsets
        # Before thinking of something simpler, beware the case where a
        # dependency bridges over the entity. Here the alignment of the
        # tokens changes.
        span_root = span.root.i
        token.dep = span.root.dep
        # We update token.lex after keeping span root and dep, since
        # setting token.lex will change span.start and span.end properties
        # as it modifies the character offsets in the doc
        token.lex = lex
        for i in range(self.length):
            self.c[i].head += i
        # Set the head of the merged token, and its dep relation, from the Span
        token.head = self.c[span_root].head
        # Adjust deps before shrinking tokens
        # Tokens which point into the merged token should now point to it
        # Subtract the offset from all tokens which point to >= end
        offset = (end - start) - 1
        for i in range(self.length):
            head_idx = self.c[i].head
            if start <= head_idx < end:
                self.c[i].head = start
            elif head_idx >= end:
                self.c[i].head -= offset
        # Now compress the token array
        for i in range(end, self.length):
            self.c[i - offset] = self.c[i]
        for i in range(self.length - offset, self.length):
            memset(&self.c[i], 0, sizeof(TokenC))
            self.c[i].lex = &EMPTY_LEXEME
        self.length -= offset
        for i in range(self.length):
            # ...And, set heads back to a relative position
            self.c[i].head -= i
        # Set the left/right children, left/right edges
        set_children_from_heads(self.c, self.length)
        # Clear the cached Python objects
        # Return the merged Python object
        return self[start]

    def print_tree(self, light=False, flat=False):
        """Returns the parse trees in JSON (dict) format.

        light (bool): Don't include lemmas or entities.
        flat (bool): Don't include arcs or modifiers.
        RETURNS (dict): Parse tree as dict.

        EXAMPLE:
            >>> doc = nlp('Bob brought Alice the pizza. Alice ate the pizza.')
            >>> trees = doc.print_tree()
            >>> trees[1]
            {'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'}
        """
        return parse_tree(self, light=light, flat=flat)


cdef int token_by_start(const TokenC* tokens, int length, int start_char) except -2:
    cdef int i
    for i in range(length):
        if tokens[i].idx == start_char:
            return i
    else:
        return -1


cdef int token_by_end(const TokenC* tokens, int length, int end_char) except -2:
    cdef int i
    for i in range(length):
        if tokens[i].idx + tokens[i].lex.length == end_char:
            return i
    else:
        return -1


cdef int set_children_from_heads(TokenC* tokens, int length) except -1:
    cdef TokenC* head
    cdef TokenC* child
    cdef int i
    # Set number of left/right children to 0. We'll increment it in the loops.
    for i in range(length):
        tokens[i].l_kids = 0
        tokens[i].r_kids = 0
        tokens[i].l_edge = i
        tokens[i].r_edge = i
    # Set left edges
    for i in range(length):
        child = &tokens[i]
        head = &tokens[i + child.head]
        if child < head:
            head.l_kids += 1
        if child.l_edge < head.l_edge:
            head.l_edge = child.l_edge

    # Set right edges --- same as above, but iterate in reverse
    for i in range(length-1, -1, -1):
        child = &tokens[i]
        head = &tokens[i + child.head]
        if child > head:
            head.r_kids += 1
        if child.r_edge > head.r_edge:
            head.r_edge = child.r_edge


    # Set sentence starts
    for i in range(length):
        if tokens[i].head == 0 and tokens[i].dep != 0:
            tokens[tokens[i].l_edge].sent_start = True


def pickle_doc(doc):
    bytes_data = doc.to_bytes(vocab=False, user_data=False)
    hooks_and_data = (doc.user_data, doc.user_hooks, doc.user_span_hooks,
                      doc.user_token_hooks)
    return (unpickle_doc, (doc.vocab, dill.dumps(hooks_and_data), bytes_data))


def unpickle_doc(vocab, hooks_and_data, bytes_data):
    user_data, doc_hooks, span_hooks, token_hooks = dill.loads(hooks_and_data)

    doc = Doc(vocab, user_data=user_data).from_bytes(bytes_data,
                                                     exclude='user_data')
    doc.user_hooks.update(doc_hooks)
    doc.user_span_hooks.update(span_hooks)
    doc.user_token_hooks.update(token_hooks)
    return doc


copy_reg.pickle(Doc, pickle_doc, unpickle_doc)