Sequence.__len__() - Code Metrics - Inspection of "Make the Sequence brick implement the Sequence ADT..." - mila-udem/blocks - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — master (#1163)

by David

created 2016-11-11 22:40 UTC

Sequence.len() A

↳ Parent: Project

Complexity

Conditions

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	1
dl	0
loc	2
rs	10
c	0
b	0
f	0

"""Bricks that compose together other bricks in linear sequences."""
from collections import MutableSequence
import copy
from toolz import interleave, unique
from picklable_itertools.extras import equizip

from ..utils import pack
from .base import Brick, application, lazy
from .interfaces import Feedforward, Initializable
from .simple import Linear


class Sequence(Brick, MutableSequence):
    """A sequence of bricks.

    This brick applies a sequence of bricks, assuming that their in- and
    outputs are compatible.

    Parameters
    ----------
    application_methods : list
        List of :class:`.BoundApplication` or :class:`.Brick` to apply.
        For :class:`.Brick`s, the ``.apply`` method is used.

    """
    def __init__(self, application_methods, **kwargs):
        pairs = ((a.apply, a) if isinstance(a, Brick) else (a, a.brick)
                 for a in application_methods)
        self.application_methods, bricks = [list(s) for s in zip(*pairs)]
        kwargs.setdefault('children', []).extend(unique(bricks))
        super(Sequence, self).__init__(**kwargs)

    @application
    def apply(self, *args):
        child_input = args
        for application_method in self.application_methods:
            output = application_method(*pack(child_input))
            child_input = output
        return output

    @apply.property('inputs')
    def apply_inputs(self):
        return self.application_methods[0].inputs

    @apply.property('outputs')
    def apply_outputs(self):
        return self.application_methods[-1].outputs

    def __delitem__(self, index):
        application_method = self.application_methods[index]
        brick = application_method.brick
        try:
            del self.application_methods[index]
        finally:
            apps_with_brick = [a for a in self.application_methods
                               if a.brick is brick]
            if len(apps_with_brick) == 0:
                self.children.remove(brick)

    def __getitem__(self, index):
        return self.application_methods[index]

    def __setitem__(self, index, new_application):
        del self[index]
        self.insert(index, new_application)

    def insert(self, index, new_application):
        if new_application.brick not in self.children:
            self.children.append(new_application.brick)
        self.application_methods.insert(index, new_application)

    def __len__(self):
        return len(self.application_methods)


class FeedforwardSequence(Sequence, Feedforward):
    """A sequence where the first and last bricks are feedforward.

    Parameters
    ----------
    application_methods : list
        List of :class:`.BoundApplication` to apply. The first and last
        application method should belong to a :class:`Feedforward` brick.

    """
    @property
    def input_dim(self):
        return self.children[0].input_dim

    @input_dim.setter
    def input_dim(self, value):
        self.children[0].input_dim = value

    @property
    def output_dim(self):
        return self.children[-1].output_dim

    @output_dim.setter
    def output_dim(self, value):
        self.children[-1].output_dim = value


class MLP(FeedforwardSequence, Initializable):
    """A simple multi-layer perceptron.

    Parameters
    ----------
    activations : list of :class:`.Brick`, :class:`.BoundApplication`,
                  or ``None``
        A list of activations to apply after each linear transformation.
        Give ``None`` to not apply any activation. It is assumed that the
        application method to use is ``apply``. Required for
        :meth:`__init__`.
    dims : list of ints
        A list of input dimensions, as well as the output dimension of the
        last layer. Required for :meth:`~.Brick.allocate`.
    prototype : :class:`.Brick`, optional
        The transformation prototype. A copy will be created for every
        activation. If not provided, an instance of :class:`~simple.Linear`
        will be used.

    Notes
    -----
    See :class:`Initializable` for initialization parameters.

    Note that the ``weights_init``, ``biases_init`` (as well as
    ``use_bias`` if set to a value other than the default of ``None``)
    configurations will overwrite those of the layers each time the
    :class:`MLP` is re-initialized. For more fine-grained control, push the
    configuration to the child layers manually before initialization.

    >>> from blocks.bricks import Tanh
    >>> from blocks.initialization import IsotropicGaussian, Constant
    >>> mlp = MLP(activations=[Tanh(), None], dims=[30, 20, 10],
    ...           weights_init=IsotropicGaussian(),
    ...           biases_init=Constant(1))
    >>> mlp.push_initialization_config()  # Configure children
    >>> mlp.children[0].weights_init = IsotropicGaussian(0.1)
    >>> mlp.initialize()

    """
    @lazy(allocation=['dims'])
    def __init__(self, activations, dims, prototype=None, **kwargs):
        self.activations = activations
        self.prototype = Linear() if prototype is None else prototype
        self.linear_transformations = []
        for i in range(len(activations)):
            linear = copy.deepcopy(self.prototype)
            name = self.prototype.__class__.__name__.lower()
            linear.name = '{}_{}'.format(name, i)
            self.linear_transformations.append(linear)
        if not dims:
            dims = [None] * (len(activations) + 1)
        self.dims = dims
        # Interleave the transformations and activations
        applications = [a for a in interleave([self.linear_transformations,
                                               activations]) if a is not None]
        super(MLP, self).__init__(applications, **kwargs)

    @property
    def input_dim(self):
        return self.dims[0]

    @input_dim.setter
    def input_dim(self, value):

        self.dims[0] = value

    @property
    def output_dim(self):
        return self.dims[-1]

    @output_dim.setter
    def output_dim(self, value):

        self.dims[-1] = value

    def _push_allocation_config(self):
        if not len(self.dims) - 1 == len(self.linear_transformations):
            raise ValueError
        for input_dim, output_dim, layer in \
                equizip(self.dims[:-1], self.dims[1:],
                        self.linear_transformations):
            layer.input_dim = input_dim
            layer.output_dim = output_dim
            if getattr(self, 'use_bias', None) is not None:
                layer.use_bias = self.use_bias


1			"""Bricks that compose together other bricks in linear sequences."""
2			from collections import MutableSequence
3			import copy
4			from toolz import interleave, unique
5			from picklable_itertools.extras import equizip
6
7			from ..utils import pack
8			from .base import Brick, application, lazy
9			from .interfaces import Feedforward, Initializable
10			from .simple import Linear
11
12
13			class Sequence(Brick, MutableSequence):
14			"""A sequence of bricks.
15
16			This brick applies a sequence of bricks, assuming that their in- and
17			outputs are compatible.
18
19			Parameters
20			----------
21			application_methods : list
22			List of :class:`.BoundApplication` or :class:`.Brick` to apply.
23			For :class:`.Brick`s, the ``.apply`` method is used.
24
25			"""
26			def __init__(self, application_methods, **kwargs):
27			pairs = ((a.apply, a) if isinstance(a, Brick) else (a, a.brick)
28			for a in application_methods)
29			self.application_methods, bricks = [list(s) for s in zip(*pairs)]
30			kwargs.setdefault('children', []).extend(unique(bricks))
31			super(Sequence, self).__init__(**kwargs)
32
33			@application
34			def apply(self, *args):
35			child_input = args
36			for application_method in self.application_methods:
37			output = application_method(*pack(child_input))
38			child_input = output
39			return output
40
41			@apply.property('inputs')
42			def apply_inputs(self):
43			return self.application_methods[0].inputs
44
45			@apply.property('outputs')
46			def apply_outputs(self):
47			return self.application_methods[-1].outputs
48
49			def __delitem__(self, index):
50			application_method = self.application_methods[index]
51			brick = application_method.brick
52			try:
53			del self.application_methods[index]
54			finally:
55			apps_with_brick = [a for a in self.application_methods
56			if a.brick is brick]
57			if len(apps_with_brick) == 0:
58			self.children.remove(brick)
59
60			def __getitem__(self, index):
61			return self.application_methods[index]
62
63			def __setitem__(self, index, new_application):
64			del self[index]
65			self.insert(index, new_application)
66
67			def insert(self, index, new_application):
68			if new_application.brick not in self.children:
69			self.children.append(new_application.brick)
70			self.application_methods.insert(index, new_application)
71
72			def __len__(self):
73			return len(self.application_methods)
74
75
76			class FeedforwardSequence(Sequence, Feedforward):
77			"""A sequence where the first and last bricks are feedforward.
78
79			Parameters
80			----------
81			application_methods : list
82			List of :class:`.BoundApplication` to apply. The first and last
83			application method should belong to a :class:`Feedforward` brick.
84
85			"""
86			@property
87			def input_dim(self):
88			return self.children[0].input_dim
89
90			@input_dim.setter
91			def input_dim(self, value):
92			self.children[0].input_dim = value
93
94			@property
95			def output_dim(self):
96			return self.children[-1].output_dim
97
98			@output_dim.setter
99			def output_dim(self, value):
100			self.children[-1].output_dim = value
101
102
103			class MLP(FeedforwardSequence, Initializable):
104			"""A simple multi-layer perceptron.
105
106			Parameters
107			----------
108			activations : list of :class:`.Brick`, :class:`.BoundApplication`,
109			or ``None``
110			A list of activations to apply after each linear transformation.
111			Give ``None`` to not apply any activation. It is assumed that the
112			application method to use is ``apply``. Required for
113			:meth:`__init__`.
114			dims : list of ints
115			A list of input dimensions, as well as the output dimension of the
116			last layer. Required for :meth:`~.Brick.allocate`.
117			prototype : :class:`.Brick`, optional
118			The transformation prototype. A copy will be created for every
119			activation. If not provided, an instance of :class:`~simple.Linear`
120			will be used.
121
122			Notes
123			-----
124			See :class:`Initializable` for initialization parameters.
125
126			Note that the ``weights_init``, ``biases_init`` (as well as
127			``use_bias`` if set to a value other than the default of ``None``)
128			configurations will overwrite those of the layers each time the
129			:class:`MLP` is re-initialized. For more fine-grained control, push the
130			configuration to the child layers manually before initialization.
131
132			>>> from blocks.bricks import Tanh
133			>>> from blocks.initialization import IsotropicGaussian, Constant
134			>>> mlp = MLP(activations=[Tanh(), None], dims=[30, 20, 10],
135			... weights_init=IsotropicGaussian(),
136			... biases_init=Constant(1))
137			>>> mlp.push_initialization_config() # Configure children
138			>>> mlp.children[0].weights_init = IsotropicGaussian(0.1)
139			>>> mlp.initialize()
140
141			"""
142			@lazy(allocation=['dims'])
143			def __init__(self, activations, dims, prototype=None, **kwargs):
144			self.activations = activations
145			self.prototype = Linear() if prototype is None else prototype
146			self.linear_transformations = []
147			for i in range(len(activations)):
148			linear = copy.deepcopy(self.prototype)
149			name = self.prototype.__class__.__name__.lower()
150			linear.name = '{}_{}'.format(name, i)
151			self.linear_transformations.append(linear)
152			if not dims:
153			dims = [None] * (len(activations) + 1)
154			self.dims = dims
155			# Interleave the transformations and activations
156			applications = [a for a in interleave([self.linear_transformations,
157			activations]) if a is not None]
158			super(MLP, self).__init__(applications, **kwargs)
159
160			@property
161			def input_dim(self):
162			return self.dims[0]
163
164			@input_dim.setter
165			def input_dim(self, value):
			0 ignored issues – show Bug introduced 2016-03-10 17:46 UTC by Report Bug Copy Issue Report Arguments number differs from overridden 'input_dim' method Loading history...
166			self.dims[0] = value
167
168			@property
169			def output_dim(self):
170			return self.dims[-1]
171
172			@output_dim.setter
173			def output_dim(self, value):
			0 ignored issues – show Bug introduced 2016-03-10 17:46 UTC by Report Bug Copy Issue Report Arguments number differs from overridden 'output_dim' method Loading history...
174			self.dims[-1] = value
175
176			def _push_allocation_config(self):
177			if not len(self.dims) - 1 == len(self.linear_transformations):
178			raise ValueError
179			for input_dim, output_dim, layer in \
180			equizip(self.dims[:-1], self.dims[1:],
181			self.linear_transformations):
182			layer.input_dim = input_dim
183			layer.output_dim = output_dim
184			if getattr(self, 'use_bias', None) is not None:
185			layer.use_bias = self.use_bias
186

mila-udem / blocks

Pull Request — master (#1163)

Sequence.__len__() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like

Sequence.len() A