• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1# Copyright 2021 Huawei Technologies Co., Ltd
2#
3# Licensed under the Apache License, Version 2.0 (the "License");
4# you may not use this file except in compliance with the License.
5# You may obtain a copy of the License at
6#
7# http://www.apache.org/licenses/LICENSE-2.0
8#
9# Unless required by applicable law or agreed to in writing, software
10# distributed under the License is distributed on an "AS IS" BASIS,
11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12# See the License for the specific language governing permissions and
13# limitations under the License.
14# ==============================================================================
15import math
16
17import matplotlib.pyplot as plt
18import numpy as np
19import pytest
20
21import mindspore.dataset as ds
22from mindspore import log as logger
23import mindspore.dataset.vision.c_transforms as c_vision
24
25DATASET_DIR = "../data/dataset/testSBData/sbd"
26
27
28def visualize_dataset(images, labels, task):
29    """
30    Helper function to visualize the dataset samples
31    """
32    image_num = len(images)
33    subplot_rows = 1 if task == "Segmentation" else 4
34    for i in range(image_num):
35        plt.imshow(images[i])
36        plt.title('Original')
37        plt.savefig('./sbd_original_{}.jpg'.format(str(i)))
38        if task == "Segmentation":
39            plt.imshow(labels[i])
40            plt.title(task)
41            plt.savefig('./sbd_segmentation_{}.jpg'.format(str(i)))
42        else:
43            b_num = labels[i].shape[0]
44            for j in range(b_num):
45                plt.subplot(subplot_rows, math.ceil(b_num / subplot_rows), j + 1)
46                plt.imshow(labels[i][j])
47            plt.savefig('./sbd_boundaries_{}.jpg'.format(str(i)))
48        plt.close()
49
50
51def test_sbd_basic01(plot=False):
52    """
53    Validate SBDataset with different usage
54    """
55    task = 'Segmentation'  # Boundaries, Segmentation
56    data = ds.SBDataset(DATASET_DIR, task=task, usage='all', shuffle=False, decode=True)
57    count = 0
58    images_list = []
59    task_list = []
60    for item in data.create_dict_iterator(num_epochs=1, output_numpy=True):
61        images_list.append(item['image'])
62        task_list.append(item['task'])
63        count = count + 1
64    assert count == 6
65    if plot:
66        visualize_dataset(images_list, task_list, task)
67
68    data2 = ds.SBDataset(DATASET_DIR, task=task, usage='train', shuffle=False, decode=False)
69    count = 0
70    for item in data2.create_dict_iterator(num_epochs=1, output_numpy=True):
71        count = count + 1
72    assert count == 4
73
74    data3 = ds.SBDataset(DATASET_DIR, task=task, usage='val', shuffle=False, decode=False)
75    count = 0
76    for item in data3.create_dict_iterator(num_epochs=1, output_numpy=True):
77        count = count + 1
78    assert count == 2
79
80
81def test_sbd_basic02():
82    """
83    Validate SBDataset with repeat and batch operation
84    """
85    # Boundaries, Segmentation
86    # case 1: test num_samples
87    data1 = ds.SBDataset(DATASET_DIR, task='Boundaries', usage='train', num_samples=3, shuffle=False)
88    num_iter1 = 0
89    for _ in data1.create_dict_iterator(num_epochs=1):
90        num_iter1 += 1
91    assert num_iter1 == 3
92
93    # case 2: test repeat
94    data2 = ds.SBDataset(DATASET_DIR, task='Boundaries', usage='train', num_samples=4, shuffle=False)
95    data2 = data2.repeat(5)
96    num_iter2 = 0
97    for _ in data2.create_dict_iterator(num_epochs=1):
98        num_iter2 += 1
99    assert num_iter2 == 20
100
101    # case 3: test batch with drop_remainder=False
102    data3 = ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', shuffle=False, decode=True)
103    resize_op = c_vision.Resize((100, 100))
104    data3 = data3.map(operations=resize_op, input_columns=["image"], num_parallel_workers=1)
105    data3 = data3.map(operations=resize_op, input_columns=["task"], num_parallel_workers=1)
106    assert data3.get_dataset_size() == 4
107    assert data3.get_batch_size() == 1
108    data3 = data3.batch(batch_size=3)  # drop_remainder is default to be False
109    assert data3.get_dataset_size() == 2
110    assert data3.get_batch_size() == 3
111    num_iter3 = 0
112    for _ in data3.create_dict_iterator(num_epochs=1):
113        num_iter3 += 1
114    assert num_iter3 == 2
115
116    # case 4: test batch with drop_remainder=True
117    data4 = ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', shuffle=False, decode=True)
118    resize_op = c_vision.Resize((100, 100))
119    data4 = data4.map(operations=resize_op, input_columns=["image"], num_parallel_workers=1)
120    data4 = data4.map(operations=resize_op, input_columns=["task"], num_parallel_workers=1)
121    assert data4.get_dataset_size() == 4
122    assert data4.get_batch_size() == 1
123    data4 = data4.batch(batch_size=3, drop_remainder=True)  # the rest of incomplete batch will be dropped
124    assert data4.get_dataset_size() == 1
125    assert data4.get_batch_size() == 3
126    num_iter4 = 0
127    for _ in data4.create_dict_iterator(num_epochs=1):
128        num_iter4 += 1
129    assert num_iter4 == 1
130
131
132def test_sbd_sequential_sampler():
133    """
134    Test SBDataset with SequentialSampler
135    """
136    logger.info("Test SBDataset Op with SequentialSampler")
137    num_samples = 5
138    sampler = ds.SequentialSampler(num_samples=num_samples)
139    data1 = ds.SBDataset(DATASET_DIR, task='Segmentation', usage='all', sampler=sampler)
140    data2 = ds.SBDataset(DATASET_DIR, task='Segmentation', usage='all', shuffle=False, num_samples=num_samples)
141    num_iter = 0
142    for item1, item2 in zip(data1.create_dict_iterator(num_epochs=1, output_numpy=True),
143                            data2.create_dict_iterator(num_epochs=1, output_numpy=True)):
144        np.testing.assert_array_equal(item1["task"], item2["task"])
145        num_iter += 1
146    assert num_iter == num_samples
147
148
149def test_sbd_exception():
150    """
151    Validate SBDataset with error parameters
152    """
153    error_msg_1 = "sampler and shuffle cannot be specified at the same time"
154    with pytest.raises(RuntimeError, match=error_msg_1):
155        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', shuffle=False, sampler=ds.PKSampler(3))
156
157    error_msg_2 = "sampler and sharding cannot be specified at the same time"
158    with pytest.raises(RuntimeError, match=error_msg_2):
159        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', num_shards=2, shard_id=0,
160                     sampler=ds.PKSampler(3))
161
162    error_msg_3 = "num_shards is specified and currently requires shard_id as well"
163    with pytest.raises(RuntimeError, match=error_msg_3):
164        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', num_shards=10)
165
166    error_msg_4 = "shard_id is specified but num_shards is not"
167    with pytest.raises(RuntimeError, match=error_msg_4):
168        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', shard_id=0)
169
170    error_msg_5 = "Input shard_id is not within the required interval"
171    with pytest.raises(ValueError, match=error_msg_5):
172        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', num_shards=5, shard_id=-1)
173    with pytest.raises(ValueError, match=error_msg_5):
174        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', num_shards=5, shard_id=5)
175    with pytest.raises(ValueError, match=error_msg_5):
176        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', num_shards=2, shard_id=5)
177
178    error_msg_6 = "num_parallel_workers exceeds"
179    with pytest.raises(ValueError, match=error_msg_6):
180        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', shuffle=False, num_parallel_workers=0)
181    with pytest.raises(ValueError, match=error_msg_6):
182        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', shuffle=False, num_parallel_workers=256)
183    with pytest.raises(ValueError, match=error_msg_6):
184        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', shuffle=False, num_parallel_workers=-2)
185
186    error_msg_7 = "Argument shard_id"
187    with pytest.raises(TypeError, match=error_msg_7):
188        ds.SBDataset(DATASET_DIR, task='Segmentation', usage='train', num_shards=2, shard_id="0")
189
190
191def test_sbd_usage():
192    """
193    Validate SBDataset image readings
194    """
195
196    def test_config(usage):
197        try:
198            data = ds.SBDataset(DATASET_DIR, task='Segmentation', usage=usage)
199            num_rows = 0
200            for _ in data.create_dict_iterator(num_epochs=1, output_numpy=True):
201                num_rows += 1
202        except (ValueError, TypeError, RuntimeError) as e:
203            return str(e)
204        return num_rows
205
206    assert test_config("train") == 4
207    assert test_config("train_noval") == 4
208    assert test_config("val") == 2
209    assert test_config("all") == 6
210    assert "usage is not within the valid set of ['train', 'val', 'train_noval', 'all']" in test_config("invalid")
211    assert "Argument usage with value ['list'] is not of type [<class 'str'>]" in test_config(["list"])
212
213
214if __name__ == "__main__":
215    test_sbd_basic01()
216    test_sbd_basic02()
217    test_sbd_sequential_sampler()
218    test_sbd_exception()
219    test_sbd_usage()
220