# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
""" split_data """
import numpy as np
import scipy.sparse as sp
[文档]def split_data(x, val_ratio=0.05, test_ratio=0.1, graph_type='undirected'):
r"""
Cut the training set into training set, validation set and test set according to the proportion of user input,
and perform graph reconstruction on the training set, and then return.
Args:
x (mindspore_gl.dataloader.Dataset): Graph Structured Dataset
val_ratio(float, optional): Validation set proportion. Default: 0.05.
test_ratio(float, optional): Test set proportion. Default: 0.1.
graph_type(str, optional): The type of graph.'undirected': undirected graph, 'directed': directed graph.
Default: 'undirected'.
Returns:
- **train** (numpy.ndarray) - Train set positive examples, shape :math:`(train\_len, 2)` .
- **val** (numpy.ndarray) - Validation set positive example, shape :math:`(val\_len, 2)` .
- **test** (numpy.ndarray) - Test set positive examples, shape :math:`(test\_len, 2)` .
Supported Platforms:
``Ascend`` ``GPU``
Examples:
>>> from mindspore_gl.dataloader import split_data
>>> from mindspore_gl.dataset import CoraV2
>>> ds = CoraV2('data_path')
>>> adj_coo, (train, val, test) = split_data(ds)
>>> print(train.shape, val.shape, test.shape)
(11684, 2) (263, 2) (527, 2)
"""
col = x.adj_coo.col
row = x.adj_coo.row
# Construct an adjacency matrix
adj = []
for i in range(len(col)):
idx = []
idx.append(col[i])
idx.append(row[i])
adj.append(idx)
# Take the upper triangular matrix
adj_c = [i for i in adj if i[0] != i[1]]
if graph_type == 'undirected':
adj_cc = []
for i in adj_c:
if [i[1], i[0]] not in adj_cc:
adj_cc.append(i)
else:
adj_cc = adj_c
# Shuffle the subscript order, split the validation set and the test set
np.random.shuffle(adj_cc)
s = len(adj_cc)
val_l = int(s*val_ratio)
test_l = int(s*test_ratio)
idx = np.random.randint(val_l+test_l, s-val_l-test_l)
val = adj_cc[idx:idx+val_l]
test = adj_cc[idx+val_l:idx+val_l+test_l]
# Remove the validation and test sets from the training set
for i in val+test:
if i in adj:
adj.remove([i[1], i[0]])
adj.remove([i[0], i[1]])
train = adj
adj, val, test, train = np.array(adj), np.array(val), np.array(test), np.array(train)
# Refactored graph
data = np.ones(train.shape[0])
adj_train = sp.csr_matrix((data, (train[:, 0], train[:, 1])), shape=x.adj_coo.shape).tocoo(copy=False)
return adj_train, (train, val, test)