1 /* Copyright 2016 The TensorFlow Authors. All Rights Reserved.
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 ==============================================================================*/
15
16 #include <functional>
17 #include <memory>
18
19 #include "tensorflow/core/framework/allocator.h"
20 #include "tensorflow/core/framework/fake_input.h"
21 #include "tensorflow/core/framework/node_def_builder.h"
22 #include "tensorflow/core/framework/op_kernel.h"
23 #include "tensorflow/core/framework/summary.pb.h"
24 #include "tensorflow/core/framework/tensor.h"
25 #include "tensorflow/core/framework/types.h"
26 #include "tensorflow/core/kernels/ops_testutil.h"
27 #include "tensorflow/core/kernels/ops_util.h"
28 #include "tensorflow/core/lib/core/status_test_util.h"
29 #include "tensorflow/core/lib/histogram/histogram.h"
30 #include "tensorflow/core/lib/strings/strcat.h"
31 #include "tensorflow/core/platform/env.h"
32 #include "tensorflow/core/platform/logging.h"
33 #include "tensorflow/core/platform/protobuf.h"
34 #include "tensorflow/core/platform/test.h"
35
36 namespace tensorflow {
37 namespace {
38
EXPECT_SummaryMatches(const Summary & actual,const string & expected_str)39 static void EXPECT_SummaryMatches(const Summary& actual,
40 const string& expected_str) {
41 Summary expected;
42 CHECK(protobuf::TextFormat::ParseFromString(expected_str, &expected));
43 EXPECT_EQ(expected.DebugString(), actual.DebugString());
44 }
45
46 // --------------------------------------------------------------------------
47 // SummaryAudioOp
48 // --------------------------------------------------------------------------
49 class SummaryAudioOpTest : public OpsTestBase {
50 protected:
MakeOp(const int max_outputs)51 void MakeOp(const int max_outputs) {
52 TF_ASSERT_OK(NodeDefBuilder("myop", "AudioSummaryV2")
53 .Input(FakeInput())
54 .Input(FakeInput())
55 .Input(FakeInput())
56 .Attr("max_outputs", max_outputs)
57 .Finalize(node_def()));
58 TF_ASSERT_OK(InitOp());
59 }
60
CheckAndRemoveEncodedAudio(Summary * summary)61 void CheckAndRemoveEncodedAudio(Summary* summary) {
62 for (int i = 0; i < summary->value_size(); ++i) {
63 Summary::Value* value = summary->mutable_value(i);
64 ASSERT_TRUE(value->has_audio()) << "No audio for value: " << value->tag();
65 ASSERT_FALSE(value->audio().encoded_audio_string().empty())
66 << "No encoded_audio_string for value: " << value->tag();
67 if (VLOG_IS_ON(2)) {
68 // When LOGGING, output the audio to disk for manual inspection.
69 TF_CHECK_OK(WriteStringToFile(
70 Env::Default(), strings::StrCat("/tmp/", value->tag(), ".wav"),
71 value->audio().encoded_audio_string()));
72 }
73 value->mutable_audio()->clear_encoded_audio_string();
74 }
75 }
76 };
77
TEST_F(SummaryAudioOpTest,Basic3D)78 TEST_F(SummaryAudioOpTest, Basic3D) {
79 const float kSampleRate = 44100.0f;
80 const int kMaxOutputs = 3;
81 MakeOp(kMaxOutputs);
82
83 // Feed and run
84 AddInputFromArray<tstring>(TensorShape({}), {"tag"});
85 AddInputFromArray<float>(TensorShape({4, 2, 2}),
86 {0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,
87 0.0, 0.0, 0.0, 0.0, 0.0, 0.0});
88 AddInputFromArray<float>(TensorShape({}), {kSampleRate});
89
90 TF_ASSERT_OK(RunOpKernel());
91
92 // Check the output size.
93 Tensor* out_tensor = GetOutput(0);
94 ASSERT_EQ(0, out_tensor->dims());
95 Summary summary;
96 ParseProtoUnlimited(&summary, out_tensor->scalar<tstring>()());
97
98 CheckAndRemoveEncodedAudio(&summary);
99 EXPECT_SummaryMatches(summary, R"(
100 value { tag: 'tag/audio/0'
101 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 2
102 length_frames: 2 } }
103 value { tag: 'tag/audio/1'
104 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 2
105 length_frames: 2 } }
106 value { tag: 'tag/audio/2'
107 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 2
108 length_frames: 2 } }
109 )");
110 }
111
TEST_F(SummaryAudioOpTest,Basic2D)112 TEST_F(SummaryAudioOpTest, Basic2D) {
113 const float kSampleRate = 44100.0f;
114 const int kMaxOutputs = 3;
115 MakeOp(kMaxOutputs);
116
117 // Feed and run
118 AddInputFromArray<tstring>(TensorShape({}), {"tag"});
119 AddInputFromArray<float>(TensorShape({4, 4}),
120 {0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,
121 0.0, 0.0, 0.0, 0.0, 0.0, 0.0});
122 AddInputFromArray<float>(TensorShape({}), {kSampleRate});
123
124 TF_ASSERT_OK(RunOpKernel());
125
126 // Check the output size.
127 Tensor* out_tensor = GetOutput(0);
128 ASSERT_EQ(0, out_tensor->dims());
129 Summary summary;
130 ParseProtoUnlimited(&summary, out_tensor->scalar<tstring>()());
131
132 CheckAndRemoveEncodedAudio(&summary);
133 EXPECT_SummaryMatches(summary, R"(
134 value { tag: 'tag/audio/0'
135 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 1
136 length_frames: 4 } }
137 value { tag: 'tag/audio/1'
138 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 1
139 length_frames: 4 } }
140 value { tag: 'tag/audio/2'
141 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 1
142 length_frames: 4 } }
143 )");
144 }
145
TEST_F(SummaryAudioOpTest,ZeroLength)146 TEST_F(SummaryAudioOpTest, ZeroLength) {
147 const float kSampleRate = 44100.0f;
148 const int kMaxOutputs = 3;
149 MakeOp(kMaxOutputs);
150
151 // Feed and run
152 AddInputFromArray<tstring>(TensorShape({}), {"tag"});
153 AddInputFromArray<float>(TensorShape({4, 0}), {});
154 AddInputFromArray<float>(TensorShape({}), {kSampleRate});
155
156 TF_ASSERT_OK(RunOpKernel());
157
158 // Check the output size.
159 Tensor* out_tensor = GetOutput(0);
160 ASSERT_EQ(0, out_tensor->dims());
161 Summary summary;
162 ParseProtoUnlimited(&summary, out_tensor->scalar<tstring>()());
163
164 CheckAndRemoveEncodedAudio(&summary);
165 EXPECT_SummaryMatches(summary, R"(
166 value { tag: 'tag/audio/0'
167 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 1
168 length_frames: 0 } }
169 value { tag: 'tag/audio/1'
170 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 1
171 length_frames: 0 } }
172 value { tag: 'tag/audio/2'
173 audio { content_type: "audio/wav" sample_rate: 44100 num_channels: 1
174 length_frames: 0 } }
175 )");
176 }
177
178 } // namespace
179 } // namespace tensorflow
180