1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
|
// Copyright 2023 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "chrome/browser/navigation_predictor/preloading_model_executor.h"
#include "base/base_paths.h"
#include "base/path_service.h"
#include "base/task/sequenced_task_runner.h"
#include "base/test/scoped_feature_list.h"
#include "base/test/task_environment.h"
#include "testing/gtest/include/gtest/gtest.h"
#include "third_party/blink/public/common/features.h"
using ModelInput = PreloadingModelExecutor::ModelInput;
using ModelOutput = PreloadingModelExecutor::ModelOutput;
class PreloadingModelExecutorTest : public testing::Test {
public:
PreloadingModelExecutorTest() {
scoped_feature_list_.InitAndEnableFeature(
blink::features::kPreloadingHeuristicsMLModel);
}
~PreloadingModelExecutorTest() override = default;
void SetUp() override {
base::FilePath source_root_dir;
base::PathService::Get(base::DIR_SRC_TEST_DATA_ROOT, &source_root_dir);
model_file_path_ = source_root_dir.AppendASCII("chrome")
.AppendASCII("browser")
.AppendASCII("navigation_predictor")
.AppendASCII("test")
.AppendASCII("preloading_heuristics.tflite");
execution_task_runner_ = base::ThreadPool::CreateSequencedTaskRunner(
{base::MayBlock(), base::TaskPriority::BEST_EFFORT});
model_executor_ = std::make_unique<PreloadingModelExecutor>();
model_executor_->InitializeAndMoveToExecutionThread(
/*model_inference_timeout=*/std::nullopt,
optimization_guide::proto::OPTIMIZATION_TARGET_OMNIBOX_URL_SCORING,
execution_task_runner_, base::SequencedTaskRunner::GetCurrentDefault());
}
void TearDown() override {
// Destroy model executor.
execution_task_runner_->DeleteSoon(FROM_HERE, std::move(model_executor_));
RunUntilIdle();
}
void RunUntilIdle() { task_environment_.RunUntilIdle(); }
protected:
base::test::ScopedFeatureList scoped_feature_list_;
base::test::TaskEnvironment task_environment_;
base::FilePath model_file_path_;
scoped_refptr<base::SequencedTaskRunner> execution_task_runner_;
std::unique_ptr<PreloadingModelExecutor> model_executor_;
};
TEST_F(PreloadingModelExecutorTest, ExecuteModel) {
// Update model file.
execution_task_runner_->PostTask(
FROM_HERE,
base::BindOnce(
&optimization_guide::ModelExecutor<ModelOutput,
ModelInput>::UpdateModelFile,
model_executor_->GetWeakPtrForExecutionThread(), model_file_path_));
// Execute model.
std::unique_ptr<base::RunLoop> run_loop = std::make_unique<base::RunLoop>();
base::OnceCallback<void(const std::optional<ModelOutput>&)>
execution_callback = base::BindOnce(
[](base::RunLoop* run_loop,
const std::optional<ModelOutput>& output) {
ASSERT_TRUE(output.has_value());
// TODO(isaboori): After the trained model is approved, use
// realistic inputs and check the output value.
run_loop->Quit();
},
run_loop.get());
base::TimeTicks now = base::TimeTicks::Now();
ModelInput input = std::vector<float>(/*count=*/17, /*value=*/0.0);
execution_task_runner_->PostTask(
FROM_HERE, base::BindOnce(&optimization_guide::ModelExecutor<
ModelOutput, ModelInput>::SendForExecution,
model_executor_->GetWeakPtrForExecutionThread(),
std::move(execution_callback), now, input));
run_loop->Run();
}
|
