spaCy/spacy/tokens/span.pxd

cimport numpy as np

from .doc cimport Doc
from ..typedefs cimport attr_t


cdef class Span:
    cdef readonly Doc doc
    cdef readonly int start
    cdef readonly int end
    cdef readonly int start_char
    cdef readonly int end_char
    cdef readonly attr_t label

    cdef public _vector
    cdef public _vector_norm

    cpdef int _recalculate_indices(self) except -1
    cpdef np.ndarray to_array(self, object features)
Add tensor field to Lexeme, Token, Doc and Span, so that users have a place to hang neural network outputs 2016-10-14 04:24:13 +03:00			`cimport numpy as np`

* Break up tokens.pyx into tokens/doc.pyx, tokens/token.pyx, tokens/spans.pyx 2015-07-13 21:20:58 +03:00			`from .doc cimport Doc`
Fixes for new StringStore 2017-05-28 19:09:27 +03:00			`from ..typedefs cimport attr_t`
* Break up tokens.pyx into tokens/doc.pyx, tokens/token.pyx, tokens/spans.pyx 2015-07-13 21:20:58 +03:00

			`cdef class Span:`
* Rename _seq to doc attribute in Span 2015-09-29 16:03:55 +03:00			`cdef readonly Doc doc`
* Removed unused i variable from Span, and set attributes to read-only 2015-11-07 09:06:15 +03:00			`cdef readonly int start`
			`cdef readonly int end`
			`cdef readonly int start_char`
			`cdef readonly int end_char`
Fixes for new StringStore 2017-05-28 19:09:27 +03:00			`cdef readonly attr_t label`
* Try giving Doc and Span objects vector and vector_norm attributes, and .similarity functions. Turns out to be bad idea. 2015-09-17 04:50:11 +03:00
			`cdef public _vector`
			`cdef public _vector_norm`
* Rework the Span-merge patch, to avoid extending the interface of Doc, and avoid virtualizing the Span.start and Span.end indices, to keep Span usage efficient 2015-11-07 00:56:49 +03:00
			`cpdef int _recalculate_indices(self) except -1`
Add Span.to_array method 2017-08-19 13:20:45 +03:00			`cpdef np.ndarray to_array(self, object features)`