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
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
|
// Copyright 2022 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "components/omnibox/browser/on_device_tail_tokenizer.h"
#include <string>
#include <vector>
#include "base/files/file_path.h"
#include "base/files/file_util.h"
#include "base/path_service.h"
#include "base/strings/utf_string_conversions.h"
#include "testing/gmock/include/gmock/gmock.h"
#include "testing/gtest/include/gtest/gtest.h"
using ::testing::ElementsAreArray;
class OnDeviceTailTokenizerTest : public ::testing::Test {
public:
OnDeviceTailTokenizerTest() { tokenizer_.Reset(); }
protected:
void InitializeTokenizer(const std::string filename) {
base::FilePath file_path;
base::PathService::Get(base::DIR_SRC_TEST_DATA_ROOT, &file_path);
file_path =
file_path.AppendASCII("components/test/data/omnibox/" + filename);
ASSERT_TRUE(base::PathExists(file_path));
tokenizer_.Init(file_path);
EXPECT_TRUE(tokenizer_.IsReady());
}
OnDeviceTailTokenizer tokenizer_;
};
TEST_F(OnDeviceTailTokenizerTest, IsTokenPrintable) {
InitializeTokenizer("vocab_test.txt");
EXPECT_TRUE(tokenizer_.IsTokenPrintable(33));
EXPECT_TRUE(tokenizer_.IsTokenPrintable(260));
EXPECT_FALSE(tokenizer_.IsTokenPrintable(1));
EXPECT_FALSE(tokenizer_.IsTokenPrintable(257));
EXPECT_FALSE(tokenizer_.IsTokenPrintable(600));
}
TEST_F(OnDeviceTailTokenizerTest, CreatePrefixTokenization) {
{
SCOPED_TRACE("Test for ASCII vocab #1.");
InitializeTokenizer("vocab_test.txt");
OnDeviceTailTokenizer::Tokenization tokenization;
// Expect tokens ["n", "j", " ", "do", "c"].
// See OnDeviceTailTokenizer::EncodeRawString for details and simplified
// examples about how ID sequences are determined.
tokenizer_.CreatePrefixTokenization("nj doc", &tokenization);
EXPECT_THAT(tokenization.unambiguous_ids,
testing::ElementsAreArray({257, 110, 106, 32, 297}));
EXPECT_EQ("c", tokenization.constraint_prefix);
EXPECT_EQ("nj do", tokenization.unambiguous_prefix);
}
{
SCOPED_TRACE("Test for ASCII vocab #2.");
InitializeTokenizer("vocab_test.txt");
OnDeviceTailTokenizer::Tokenization tokenization;
// Expect tokens ["re", "mi", "t", "ly", " ", "log", "in"].
tokenizer_.CreatePrefixTokenization("remitly login", &tokenization);
EXPECT_THAT(tokenization.unambiguous_ids,
testing::ElementsAreArray({257, 414, 366, 116, 363, 32, 521}));
EXPECT_EQ("in", tokenization.constraint_prefix);
EXPECT_EQ("remitly log", tokenization.unambiguous_prefix);
}
{
SCOPED_TRACE("Test for ASCII vocab #3.");
InitializeTokenizer("vocab_test.txt");
OnDeviceTailTokenizer::Tokenization tokenization;
// Expect tokens
// ["us", " ", "pa", "ss", "po", "rt", " ", "ap", "pl", "ica", "tio", "n"]
tokenizer_.CreatePrefixTokenization("us passport application",
&tokenization);
EXPECT_THAT(tokenization.unambiguous_ids,
testing::ElementsAreArray({257, 456, 32, 402, 434, 407, 424, 32,
270, 406, 507, 549}));
EXPECT_EQ("n", tokenization.constraint_prefix);
EXPECT_EQ("us passport applicatio", tokenization.unambiguous_prefix);
}
{
SCOPED_TRACE("Test for i18n languages.");
InitializeTokenizer("vocab_i18n_test.txt");
OnDeviceTailTokenizer::Tokenization tokenization;
// Expect tokens
// ["us", "ल्", "वावि", "てる", "a", "वा"]
tokenizer_.CreatePrefixTokenization("usल्वाविてるaवा", &tokenization);
EXPECT_THAT(tokenization.unambiguous_ids,
testing::ElementsAreArray({257, 259, 260, 263, 264, 97}));
EXPECT_EQ("वा", tokenization.constraint_prefix);
EXPECT_EQ("usल्वाविてるa", tokenization.unambiguous_prefix);
}
}
TEST_F(OnDeviceTailTokenizerTest, TokenizePrevQuery) {
{
SCOPED_TRACE("Test for ASCII vocab #1.");
InitializeTokenizer("vocab_test.txt");
OnDeviceTailTokenizer::TokenIds token_ids;
tokenizer_.TokenizePrevQuery("facebook", &token_ids);
// Expect tokens: ["fa", "ce", "bo", "ok"]
EXPECT_EQ(4, (int)token_ids.size());
EXPECT_THAT(token_ids, ElementsAreArray({317, 285, 281, 390}));
EXPECT_EQ("fa", tokenizer_.IdToToken(token_ids[0]));
}
{
SCOPED_TRACE("Test for ASCII vocab #2.");
InitializeTokenizer("vocab_test.txt");
OnDeviceTailTokenizer::TokenIds token_ids;
tokenizer_.TokenizePrevQuery("matching gym outfits", &token_ids);
// Expect tokens:
// ["ma", "t", "chi", "ng", " ", "g", "y", "m", " ", "out", "fi", "ts"]
EXPECT_EQ(12, (int)token_ids.size());
EXPECT_THAT(token_ids, ElementsAreArray({364, 116, 488, 375, 32, 103, 121,
109, 32, 533, 320, 443}));
EXPECT_EQ("ma", tokenizer_.IdToToken(token_ids[0]));
}
{
SCOPED_TRACE("Test for i18n languages.");
InitializeTokenizer("vocab_i18n_test.txt");
OnDeviceTailTokenizer::TokenIds token_ids;
tokenizer_.TokenizePrevQuery("usल्वाविてるaवा", &token_ids);
// Expect tokens:
// ["us", "ल्", "वावि", "てる", "a", "वा"]
EXPECT_EQ(6, (int)token_ids.size());
EXPECT_THAT(token_ids, ElementsAreArray({259, 260, 263, 264, 97, 261}));
EXPECT_EQ("us", tokenizer_.IdToToken(token_ids[0]));
}
}
|
