utils.py

from __future__ import print_function

import scipy.sparse as sp
import numpy as np
from scipy.sparse.linalg.eigen.arpack import eigsh, ArpackNoConvergence


def encode_onehot(labels):
    classes = set(labels)
    classes_dict = {c: np.identity(len(classes))[i, :] for i, c in enumerate(classes)}
    labels_onehot = np.array(list(map(classes_dict.get, labels)), dtype=np.int32)
    return labels_onehot


def load_data(path="data/cora/", dataset="cora"):
    """Load citation network dataset (cora only for now)"""
    print('Loading {} dataset...'.format(dataset))

    idx_features_labels = np.genfromtxt("{}{}.content".format(path, dataset), dtype=np.dtype(str))
    features = sp.csr_matrix(idx_features_labels[:, 1:-1], dtype=np.float32)
    labels = encode_onehot(idx_features_labels[:, -1])

    # build graph
    idx = np.array(idx_features_labels[:, 0], dtype=np.int32)
    idx_map = {j: i for i, j in enumerate(idx)}
    edges_unordered = np.genfromtxt("{}{}.cites".format(path, dataset), dtype=np.int32)
    edges = np.array(list(map(idx_map.get, edges_unordered.flatten())),
                     dtype=np.int32).reshape(edges_unordered.shape)
    adj = sp.coo_matrix((np.ones(edges.shape[0]), (edges[:, 0], edges[:, 1])),
                        shape=(labels.shape[0], labels.shape[0]), dtype=np.float32)

    # build symmetric adjacency matrix
    adj = adj + adj.T.multiply(adj.T > adj) - adj.multiply(adj.T > adj)

    # features = normalize_features(features)
    # adj = normalize_adj(adj + sp.eye(adj.shape[0]))

    print('Dataset has {} nodes, {} edges, {} features.'.format(adj.shape[0], edges.shape[0], features.shape[1]))

    return features.todense(), adj, labels


def load_data_attention(path="data/cora/", dataset="cora"):
    """Load citation network dataset (cora only for now)"""
    print('Loading {} dataset...'.format(dataset))

    idx_features_labels = np.genfromtxt("{}{}.content".format(path, dataset), dtype=np.dtype(str))
    features = sp.csr_matrix(idx_features_labels[:, 1:-1], dtype=np.float32)
    labels = encode_onehot(idx_features_labels[:, -1])

    # build graph
    idx = np.array(idx_features_labels[:, 0], dtype=np.int32)
    idx_map = {j: i for i, j in enumerate(idx)}
    edges_unordered = np.genfromtxt("{}{}.cites".format(path, dataset), dtype=np.int32)
    edges = np.array(list(map(idx_map.get, edges_unordered.flatten())),
                     dtype=np.int32).reshape(edges_unordered.shape)
    adj = sp.coo_matrix((np.ones(edges.shape[0]), (edges[:, 0], edges[:, 1])),
                        shape=(labels.shape[0], labels.shape[0]), dtype=np.float32)

    # build symmetric adjacency matrix
    adj = adj + adj.T.multiply(adj.T > adj) - adj.multiply(adj.T > adj)

    features = normalize_features(features)
    adj = normalize_adj(adj + sp.eye(adj.shape[0]))

    print('Dataset has {} nodes, {} edges, {} features.'.format(adj.shape[0], edges.shape[0], features.shape[1]))

    return features.todense(), adj, labels


def normalize_features(mx):
    """Row-normalize sparse matrix"""
    rowsum = np.array(mx.sum(1))
    r_inv = np.power(rowsum, -1).flatten()
    r_inv[np.isinf(r_inv)] = 0.
    r_mat_inv = sp.diags(r_inv)
    mx = r_mat_inv.dot(mx)
    return mx


def normalize_adj(adj, symmetric=True):
    if symmetric:
        d = sp.diags(np.power(np.array(adj.sum(1)), -0.5).flatten(), 0)
        a_norm = adj.dot(d).transpose().dot(d).tocsr()
    else:
        d = sp.diags(np.power(np.array(adj.sum(1)), -1).flatten(), 0)
        a_norm = d.dot(adj).tocsr()
    return a_norm


def normalize_adj_numpy(adj, symmetric=True):
    if symmetric:
        d = np.diag(np.power(np.array(adj.sum(1)), -0.5).flatten(), 0)
        a_norm = adj.dot(d).transpose().dot(d)
    else:
        d = np.diag(np.power(np.array(adj.sum(1)), -1).flatten(), 0)
        a_norm = d.dot(adj)
    return a_norm


def preprocess_adj(adj, symmetric=True):
    adj = adj + sp.eye(adj.shape[0])
    adj = normalize_adj(adj, symmetric)
    return adj


def preprocess_adj_numpy(adj, symmetric=True):
    adj = adj + np.eye(adj.shape[0])
    adj = normalize_adj_numpy(adj, symmetric)
    return adj


def preprocess_adj_tensor(adj_tensor, symmetric=True):
    adj_out_tensor = []
    for i in range(adj_tensor.shape[0]):
        adj = adj_tensor[i]
        adj = adj + np.eye(adj.shape[0])
        adj = normalize_adj_numpy(adj, symmetric)
        adj_out_tensor.append(adj)
    adj_out_tensor = np.array(adj_out_tensor)
    return adj_out_tensor


def preprocess_adj_tensor_with_identity(adj_tensor, symmetric=True):
    adj_out_tensor = []
    for i in range(adj_tensor.shape[0]):
        adj = adj_tensor[i]
        adj = adj + np.eye(adj.shape[0])
        adj = normalize_adj_numpy(adj, symmetric)
        adj = np.concatenate([np.eye(adj.shape[0]), adj], axis=0)
        adj_out_tensor.append(adj)
    adj_out_tensor = np.array(adj_out_tensor)
    return adj_out_tensor


def preprocess_adj_tensor_with_identity_concat(adj_tensor, symmetric=True):
    adj_out_tensor = []
    for i in range(adj_tensor.shape[0]):
        adj = adj_tensor[i]
        adj = adj + np.eye(adj.shape[0])
        adj = normalize_adj_numpy(adj, symmetric)
        adj = np.concatenate([np.eye(adj.shape[0]), adj], axis=0)
        adj_out_tensor.append(adj)
    adj_out_tensor = np.concatenate(adj_out_tensor, axis=0)
    return adj_out_tensor

def preprocess_adj_tensor_concat(adj_tensor, symmetric=True):
    adj_out_tensor = []
    for i in range(adj_tensor.shape[0]):
        adj = adj_tensor[i]
        adj = adj + np.eye(adj.shape[0])
        adj = normalize_adj_numpy(adj, symmetric)
        adj_out_tensor.append(adj)
    adj_out_tensor = np.concatenate(adj_out_tensor, axis=0)
    return adj_out_tensor

def preprocess_edge_adj_tensor(edge_adj_tensor, symmetric=True):
    edge_adj_out_tensor = []
    num_edge_features = int(edge_adj_tensor.shape[1]/edge_adj_tensor.shape[2])

    for i in range(edge_adj_tensor.shape[0]):
        edge_adj = edge_adj_tensor[i]
        edge_adj = np.split(edge_adj, num_edge_features, axis=0)
        edge_adj = np.array(edge_adj)
        edge_adj = preprocess_adj_tensor_concat(edge_adj, symmetric)
        edge_adj_out_tensor.append(edge_adj)

    edge_adj_out_tensor = np.array(edge_adj_out_tensor)
    return edge_adj_out_tensor


def sample_mask(idx, l):
    mask = np.zeros(l)
    mask[idx] = 1
    return np.array(mask, dtype=np.bool)


def get_splits(y):
    idx_train = range(140)
    idx_val = range(200, 500)
    idx_test = range(500, 1500)
    y_train = np.zeros(y.shape, dtype=np.int32)
    y_val = np.zeros(y.shape, dtype=np.int32)
    y_test = np.zeros(y.shape, dtype=np.int32)
    y_train[idx_train] = y[idx_train]
    y_val[idx_val] = y[idx_val]
    y_test[idx_test] = y[idx_test]
    train_mask = sample_mask(idx_train, y.shape[0])
    return y_train, y_val, y_test, idx_train, idx_val, idx_test, train_mask


def get_splits_v2(y):
    idx_train = range(1708)
    idx_val = range(1708, 1708 + 500)
    idx_test = range(1708 + 500, 2708)
    y_train = np.zeros(y.shape, dtype=np.int32)
    y_val = np.zeros(y.shape, dtype=np.int32)
    y_test = np.zeros(y.shape, dtype=np.int32)
    y_train[idx_train] = y[idx_train]
    y_val[idx_val] = y[idx_val]
    y_test[idx_test] = y[idx_test]
    train_mask = sample_mask(idx_train, y.shape[0])
    return y_train, y_val, y_test, idx_train, idx_val, idx_test, train_mask


def categorical_crossentropy(preds, labels):
    return np.mean(-np.log(np.extract(labels, preds)))


def accuracy(preds, labels):
    return np.mean(np.equal(np.argmax(labels, 1), np.argmax(preds, 1)))


def evaluate_preds(preds, labels, indices):
    split_loss = list()
    split_acc = list()

    for y_split, idx_split in zip(labels, indices):
        split_loss.append(categorical_crossentropy(preds[idx_split], y_split[idx_split]))
        split_acc.append(accuracy(preds[idx_split], y_split[idx_split]))

    return split_loss, split_acc


def normalized_laplacian(adj, symmetric=True):
    adj_normalized = normalize_adj(adj, symmetric)
    laplacian = sp.eye(adj.shape[0]) - adj_normalized
    return laplacian


def rescale_laplacian(laplacian):
    try:
        print('Calculating largest eigenvalue of normalized graph Laplacian...')
        largest_eigval = eigsh(laplacian, 1, which='LM', return_eigenvectors=False)[0]
    except ArpackNoConvergence:
        print('Eigenvalue calculation did not converge! Using largest_eigval=2 instead.')
        largest_eigval = 2

    scaled_laplacian = (2. / largest_eigval) * laplacian - sp.eye(laplacian.shape[0])
    return scaled_laplacian


def chebyshev_polynomial(X, k):
    """Calculate Chebyshev polynomials up to order k. Return a list of sparse matrices."""
    print("Calculating Chebyshev polynomials up to order {}...".format(k))

    T_k = list()
    T_k.append(sp.eye(X.shape[0]).tocsr())
    T_k.append(X)

    def chebyshev_recurrence(T_k_minus_one, T_k_minus_two, X):
        X_ = sp.csr_matrix(X, copy=True)
        return 2 * X_.dot(T_k_minus_one) - T_k_minus_two

    for i in range(2, k + 1):
        T_k.append(chebyshev_recurrence(T_k[-1], T_k[-2], X))

    return T_k


def sparse_to_tuple(sparse_mx):
    if not sp.isspmatrix_coo(sparse_mx):
        sparse_mx = sparse_mx.tocoo()
    coords = np.vstack((sparse_mx.row, sparse_mx.col)).transpose()
    values = sparse_mx.data
    shape = sparse_mx.shape
    return coords, values, shape





## GATED GRAPH NEURAL NETWORK

#!/usr/bin/env/python

import numpy as np
import tensorflow as tf
import queue
import threading

SMALL_NUMBER = 1e-7


def glorot_init(shape):
    initialization_range = np.sqrt(6.0 / (shape[-2] + shape[-1]))
    return np.random.uniform(low=-initialization_range, high=initialization_range, size=shape).astype(np.float32)


class ThreadedIterator:
    """An iterator object that computes its elements in a parallel thread to be ready to be consumed.
    The iterator should *not* return None"""

    def __init__(self, original_iterator, max_queue_size: int=2):
        self.__queue = queue.Queue(maxsize=max_queue_size)
        self.__thread = threading.Thread(target=lambda: self.worker(original_iterator))
        self.__thread.start()

    def worker(self, original_iterator):
        for element in original_iterator:
            assert element is not None, 'By convention, iterator elements much not be None'
            self.__queue.put(element, block=True)
        self.__queue.put(None, block=True)

    def __iter__(self):
        next_element = self.__queue.get(block=True)
        while next_element is not None:
            yield next_element
            next_element = self.__queue.get(block=True)
        self.__thread.join()


class MLP(object):
    def __init__(self, in_size, out_size, hid_sizes, dropout_keep_prob):
        self.in_size = in_size
        self.out_size = out_size
        self.hid_sizes = hid_sizes
        self.dropout_keep_prob = dropout_keep_prob
        self.params = self.make_network_params()

    def make_network_params(self):
        dims = [self.in_size] + self.hid_sizes + [self.out_size]
        weight_sizes = list(zip(dims[:-1], dims[1:]))
        weights = [tf.Variable(self.init_weights(s), name='MLP_W_layer%i' % i)
                   for (i, s) in enumerate(weight_sizes)]
        biases = [tf.Variable(np.zeros(s[-1]).astype(np.float32), name='MLP_b_layer%i' % i)
                  for (i, s) in enumerate(weight_sizes)]

        network_params = {
            "weights": weights,
            "biases": biases,
        }

        return network_params

    def init_weights(self, shape):
        return np.sqrt(6.0 / (shape[-2] + shape[-1])) * (2 * np.random.rand(*shape).astype(np.float32) - 1)

    def __call__(self, inputs):
        acts = inputs
        for W, b in zip(self.params["weights"], self.params["biases"]):
            hid = tf.matmul(acts, tf.nn.dropout(W, self.dropout_keep_prob)) + b
            acts = tf.nn.relu(hid)
        last_hidden = hid
        return last_hidden