layers.py

from inits import *
import tensorflow as tf

flags = tf.app.flags
FLAGS = flags.FLAGS

# global unique layer ID dictionary for layer name assignment
_LAYER_UIDS = {}


def get_layer_uid(layer_name=''):
    """Helper function, assigns unique layer IDs."""
    if layer_name not in _LAYER_UIDS:
        _LAYER_UIDS[layer_name] = 1
        return 1
    else:
        _LAYER_UIDS[layer_name] += 1
        return _LAYER_UIDS[layer_name]


def sparse_dropout(x, keep_prob, noise_shape, is_training):
    """Dropout for sparse tensors."""
    if is_training == True:
        random_tensor = keep_prob
        random_tensor += tf.random_uniform(noise_shape)
        dropout_mask = tf.cast(tf.floor(random_tensor), dtype=tf.bool)
        pre_out = tf.sparse_retain(x, dropout_mask)
        x = pre_out * (1./keep_prob)
    return x


def dot(x, y, sparse=False):
    """Wrapper for tf.matmul (sparse vs dense)."""
    if sparse:
        res = tf.sparse_tensor_dense_matmul(x, y)
    else:
        res = tf.matmul(x, y)
    return res


class Layer(object):
    """Base layer class. Defines basic API for all layer objects.
    Implementation inspired by keras (http://keras.io).

    # Properties
        name: String, defines the variable scope of the layer.
        logging: Boolean, switches Tensorflow histogram logging on/off

    # Methods
        _call(inputs): Defines computation graph of layer
            (i.e. takes input, returns output)
        __call__(inputs): Wrapper for _call()
        _log_vars(): Log all variables
    """

    def __init__(self, **kwargs):
        allowed_kwargs = {'name', 'logging'}
        for kwarg in kwargs.keys():
            assert kwarg in allowed_kwargs, 'Invalid keyword argument: ' + kwarg
        name = kwargs.get('name')
        if not name:
            layer = self.__class__.__name__.lower()
            name = layer + '_' + str(get_layer_uid(layer))
        self.name = name
        self.vars = {}
        logging = kwargs.get('logging', False)
        self.logging = logging
        self.sparse_inputs = False

    def _call(self, inputs):
        return inputs

    def __call__(self, inputs):
        with tf.name_scope(self.name):
            if self.logging and not self.sparse_inputs:
                tf.summary.histogram(self.name + '/inputs', inputs)
            outputs = self._call(inputs)
            if self.logging:
                tf.summary.histogram(self.name + '/outputs', outputs)
            return outputs

    def _log_vars(self):
        for var in self.vars:
            tf.summary.histogram(self.name + '/vars/' + var, self.vars[var])


class Dense(Layer):
    """Dense layer."""
    def __init__(self, input_dim, output_dim, placeholders, dropout=0., sparse_inputs=False,
                 act=tf.nn.relu, bias=False, featureless=False, **kwargs):
        super(Dense, self).__init__(**kwargs)

        if dropout:
            self.dropout = placeholders['dropout']
        else:
            self.dropout = 0.

        self.act = act
        self.sparse_inputs = sparse_inputs
        self.featureless = featureless
        self.bias = bias

        # helper variable for sparse dropout
        self.num_features_nonzero = placeholders['num_features_nonzero']

        with tf.variable_scope(self.name + '_vars'):
            self.vars['weights'] = glorot([input_dim, output_dim],
                                          name='weights')
            if self.bias:
                self.vars['bias'] = zeros([output_dim], name='bias')

        if self.logging:
            self._log_vars()

    def _call(self, inputs):
        x = inputs

        # dropout
        if self.sparse_inputs:
            x = sparse_dropout(x, 1-self.dropout, self.num_features_nonzero, self.is_training)
        else:
            x = tf.contrib.layers.dropout(x, 1-self.dropout, is_training=self.is_training)

        # transform
        output = dot(x, self.vars['weights'], sparse=self.sparse_inputs)

        # bias
        if self.bias:
            output += self.vars['bias']

        return self.act(output)


class GraphConvolution(Layer):
    """Graph convolution layer."""
    def __init__(self, input_dim, output_dim, placeholders, dropout=0.,
                 sparse_inputs=False, act=tf.nn.relu, bias=False,
                 featureless=False, **kwargs):
        super(GraphConvolution, self).__init__(**kwargs)

        if dropout:
            self.dropout = placeholders['dropout']
        else:
            self.dropout = 0.

        self.act = act
        self.support = placeholders['support']
        self.sparse_inputs = sparse_inputs
        self.featureless = featureless
        self.bias = bias
        self.is_training = placeholders['is_training']

        # helper variable for sparse dropout
        self.num_features_nonzero = placeholders['num_features_nonzero']

        with tf.variable_scope(self.name + '_vars'):
            self.vars['weights'] = glorot([input_dim, output_dim], name='weights')
            if self.bias:
                self.vars['bias'] = zeros([output_dim], name='bias')

        if self.logging:
            self._log_vars()

    def _call(self, inputs):
        x = inputs

        # dropout
        if self.sparse_inputs:
            x = sparse_dropout(x, 1-self.dropout, self.num_features_nonzero, self.is_training)
        else:
            x = tf.contrib.layers.dropout(x, 1-self.dropout, is_training=self.is_training)

        # convolve
        if not self.featureless:
            pre_sup = dot(x, self.vars['weights'], sparse=self.sparse_inputs)
        else:
            pre_sup = self.vars['weights']
        output = dot(self.support, pre_sup, sparse=False)

        # bias
        if self.bias:
            output += self.vars['bias']

        return self.act(output)


class FullGraphConvolution(Layer):
    """Graph convolution layer."""
    def __init__(self, input_dim, output_dim, placeholders, dropout=0.,
                 act=tf.nn.relu, bias=False, **kwargs):
        super(FullGraphConvolution, self).__init__(**kwargs)

        self.dropout = placeholders['dropout'] if dropout else 0.
        self.act = act
        self.support = placeholders['support']
        self.output_dim = output_dim
        self.bias = bias
        self.is_training = placeholders['is_training']

        if self.logging:
            self._log_vars()

    def _call(self, inputs):
        x = inputs
        x = tf.contrib.layers.dropout(x, 1-self.dropout, is_training=self.is_training)
        x = graph_conv(self.support, x, self.output_dim, 9, self.name)
        return x


def top_k_features(adj_m, fea_m, k, name='top_k_features'):
    adj_m = tf.expand_dims(adj_m, axis=1, name=name+'/expand1')
    fea_m = tf.expand_dims(fea_m, axis=-1, name=name+'/expand2')
    feas = tf.multiply(adj_m, fea_m, name=name+'/mul')
    feas = tf.transpose(feas, perm=[2, 1, 0], name=name+'/trans1')
    top_k = tf.nn.top_k(feas, k=k, name=name+'/top_k').values
    top_k = tf.transpose(top_k, perm=[0, 2, 1], name=name+'/trans2')
    return top_k


def graph_conv(adj_m, fea_m, num_out, k, name='graph_conv'):
    outs = top_k_features(adj_m, fea_m, k, name+'/top_k_fea')
    l2_func = tf.contrib.layers.l2_regularizer(1e-4, name)
    outs = tf.layers.conv1d(
        outs, num_out, k, activation=tf.nn.relu, name=name+'/conv',
        kernel_initializer=tf.random_normal_initializer(),
        kernel_regularizer=l2_func, bias_regularizer=l2_func)
    outs = tf.squeeze(outs, axis=[1], name=name+'/squeeze')
    return outs