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
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
|
// (C) Copyright 2017, Google Inc.
// 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.
#include "include_gunit.h"
#include "normstrngs.h"
#include "normstrngs_test.h"
namespace tesseract {
// Though the unicode example for Telugu in section 12.7:
// http://www.unicode.org/versions/Unicode9.0.0/ch12.pdf
// shows using ZWNJ to force an explicit virama, in practice a ZWNJ is used to
// suppress a conjugate that would otherwise occur. If a consonant is followed
// by a virama and then by a non-Indic character, OpenType will presume that
// the user simply meant to suppress the inherent vowel of the consonant
// and render it as the consonant with an explicit virama, the same as if
// a ZWNJ had followed. Since this is confusing to an OCR engine, the
// normalizer always puts a termninating ZWNJ on the end if not present,
// and accepts the string as valid.
TEST(ValidateIndicTest, AddsJoinerToTerminalVirama) {
std::string str = "\u0c15\u0c4d"; // KA - virama
std::string target_str = "\u0c15\u0c4d\u200c"; // KA - virama - ZWNJ
ExpectGraphemeModeResults(str, UnicodeNormMode::kNFC, 3, 2, 1, target_str);
// Same result if we started with the normalized string.
ExpectGraphemeModeResults(target_str, UnicodeNormMode::kNFC, 3, 2, 1, target_str);
}
// Only one dependent vowel is allowed.
TEST(ValidateIndicTest, OnlyOneDependentVowel) {
std::string str = "\u0d15\u0d3e\u0d42"; // KA AA UU
std::string dest;
EXPECT_FALSE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
}
// [c26][c4d][c01]
// A consonant (DA) followed by the virama followed by a bindu
// Syllable modifiers [c01][c02][c03] all modify the pronunciation of
// the vowel in a syllable, as does the virama [c04]. You can only
// have one of these on a syllable.
//
// References:
// http://www.omniglot.com/writing/telugu.htm
TEST(ValidateIndicTest, OnlyOneVowelModifier) {
std::string str = "\u0c26\u0c4d\u0c01"; // DA virama candrabindu
std::string result;
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &result));
// It made 1 grapheme of 4 chars, by terminating the explicit virama.
EXPECT_EQ(std::string("\u0c26\u0c4d\u200c\u0c01"), result);
str = "\u0995\u0983\u0981"; // KA visarga candrabindu
EXPECT_FALSE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &result));
// Exception: Malayalam allows multiple anusvara.
str = "\u0d15\u0d02\u0d02"; // KA Anusvara Anusvara
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &result));
EXPECT_EQ(str, result);
}
// [c28][c02][c3f]
// A consonant (NA) followed by the Anusvara/sunna and another matra (I).
// The anusvara [c02] is a pronunciation directive
// for a whole syllable and only appears at the end of the syllable
// References:
// + Unicode v9, 12.1 "Modifier Mark Rules R10,"
// and the Microsoft page
// http://www.microsoft.com/typography/otfntdev/teluguot/shaping.aspx
TEST(ValidateIndicTest, VowelModifierMustBeLast) {
std::string str = "\u0c28\u0c02\u0c3f"; // NA Sunna I
std::string dest;
EXPECT_FALSE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
// Swap c02/c3f and all is ok.
str = "\u0c28\u0c3f\u0c02"; // NA I Sunna
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
EXPECT_EQ(dest, str);
}
// [c05][c47]
// A Vowel (A) followed by a combining vowel/matra (EE).
// In Telugu, matras are only put on consonants, not independent
// vowels.
// References:
// + Unicode v9, 12.1:
// Principles of the Devanagari Script: Dependent Vowel Signs (Matras).
// + http://varamozhi.sourceforge.net/iscii91.pdf
TEST(ValidateIndicTest, MatrasFollowConsonantsNotVowels) {
std::string str = "\u0c05\u0c47"; // A EE
std::string dest;
EXPECT_FALSE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
str = "\u0c1e\u0c3e"; // NYA AA
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
EXPECT_EQ(dest, str);
}
// Sub-graphemes are allowed if GraphemeNorm is turned off.
TEST(ValidateIndicTest, SubGraphemes) {
std::string str = "\u0d3e"; // AA
std::string dest;
EXPECT_FALSE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNone,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
EXPECT_EQ(dest, str);
}
TEST(ValidateIndicTest, Nukta) {
std::string str = "\u0c95\u0cbc\u0ccd\u0cb9"; // KA Nukta Virama HA
std::vector<std::string> glyphs;
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(UnicodeNormMode::kNFC, OCRNorm::kNone,
GraphemeNormMode::kGlyphSplit, true, str.c_str(),
&glyphs));
EXPECT_EQ(glyphs.size(), 3);
EXPECT_EQ(glyphs[2], std::string("\u0ccd\u0cb9"));
// Swapped Nukta and Virama are not allowed, but NFC normalization fixes it.
std::string str2 = "\u0c95\u0ccd\u0cbc\u0cb9"; // KA Virama Nukta HA
ExpectGraphemeModeResults(str2, UnicodeNormMode::kNFC, 4, 3, 1, str);
}
// Sinhala has some of its own specific rules. See www.macciato.com/sinhala
TEST(ValidateIndicTest, SinhalaRakaransaya) {
std::string str = "\u0d9a\u0dca\u200d\u0dbb"; // KA Virama ZWJ Rayanna
std::string dest;
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
EXPECT_EQ(dest, str);
std::vector<std::string> glyphs;
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(UnicodeNormMode::kNFC, OCRNorm::kNone,
GraphemeNormMode::kGlyphSplit, true, str.c_str(),
&glyphs));
EXPECT_EQ(glyphs.size(), 2);
EXPECT_EQ(glyphs[1], std::string("\u0dca\u200d\u0dbb"));
// Can be followed by a dependent vowel.
str += "\u0dd9"; // E
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
EXPECT_EQ(dest, str);
}
TEST(ValidateIndicTest, SinhalaYansaya) {
std::string str = "\u0d9a\u0dca\u200d\u0dba"; // KA Virama ZWJ Yayanna
std::string dest;
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
EXPECT_EQ(dest, str);
// Can be followed by a dependent vowel.
str += "\u0ddd"; // OO
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNorm::kNormalize,
str.c_str(), &dest))
<< PrintString32WithUnicodes(str);
EXPECT_EQ(dest, str);
std::vector<std::string> glyphs;
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(UnicodeNormMode::kNFC, OCRNorm::kNone,
GraphemeNormMode::kGlyphSplit, true, str.c_str(),
&glyphs));
EXPECT_EQ(glyphs.size(), 3);
EXPECT_EQ(glyphs[1], std::string("\u0dca\u200d\u0dba"));
}
TEST(ValidateIndicTest, SinhalaRepaya) {
std::string str = "\u0d9a\u0dbb\u0dca\u200d\u0db8"; // KA Rayanna Virama ZWJ MA
std::vector<std::string> glyphs;
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(UnicodeNormMode::kNFC, OCRNorm::kNone,
GraphemeNormMode::kCombined, true, str.c_str(),
&glyphs));
EXPECT_EQ(glyphs.size(), 2);
EXPECT_EQ(glyphs[1], std::string("\u0dbb\u0dca\u200d\u0db8"));
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(UnicodeNormMode::kNFC, OCRNorm::kNone,
GraphemeNormMode::kGlyphSplit, true, str.c_str(),
&glyphs));
EXPECT_EQ(glyphs.size(), 3);
EXPECT_EQ(glyphs[1], std::string("\u0dbb\u0dca\u200d"));
}
TEST(ValidateIndicTest, SinhalaSpecials) {
// Sinhala has some exceptions from the usual rules.
std::string str = "\u0dc0\u0d9c\u0dca\u200d\u0dbb\u0dca\u200d\u0dbb\u0dca\u200d";
std::vector<std::string> glyphs;
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(UnicodeNormMode::kNFC, OCRNorm::kNone,
GraphemeNormMode::kGlyphSplit, true, str.c_str(),
&glyphs));
EXPECT_EQ(glyphs.size(), 5) << PrintStringVectorWithUnicodes(glyphs);
EXPECT_EQ(glyphs[0], std::string("\u0dc0"));
EXPECT_EQ(glyphs[1], std::string("\u0d9c"));
EXPECT_EQ(glyphs[2], std::string("\u0dca\u200d\u0dbb"));
EXPECT_EQ(glyphs[3], std::string("\u0dca\u200d"));
EXPECT_EQ(glyphs[4], std::string("\u0dbb\u0dca\u200d"));
str = "\u0dc3\u0dbb\u0dca\u200d\u0dbb\u0dca\u200d\u0dcf";
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(UnicodeNormMode::kNFC, OCRNorm::kNone,
GraphemeNormMode::kGlyphSplit, true, str.c_str(),
&glyphs));
EXPECT_EQ(glyphs.size(), 4) << PrintStringVectorWithUnicodes(glyphs);
EXPECT_EQ(glyphs[0], std::string("\u0dc3"));
EXPECT_EQ(glyphs[1], std::string("\u0dbb\u0dca\u200d"));
EXPECT_EQ(glyphs[2], std::string("\u0dbb\u0dca\u200d"));
EXPECT_EQ(glyphs[3], std::string("\u0dcf"));
}
} // namespace tesseract
|
