mirror of
https://github.com/tesseract-ocr/tesseract.git
synced 2024-12-05 10:49:01 +08:00
426 lines
24 KiB
C++
426 lines
24 KiB
C++
// (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 "absl/strings/str_format.h" // for absl::StrFormat
|
||
#include "include_gunit.h"
|
||
#include "normstrngs.h"
|
||
#include "normstrngs_test.h"
|
||
#include <tesseract/strngs.h>
|
||
#include <tesseract/unichar.h>
|
||
#ifdef INCLUDE_TENSORFLOW
|
||
#include "util/utf8/unilib.h" // for UniLib
|
||
#endif
|
||
|
||
#include "include_gunit.h"
|
||
|
||
namespace tesseract {
|
||
namespace {
|
||
|
||
#if defined(MISSING_CODE)
|
||
static std::string EncodeAsUTF8(const char32 ch32) {
|
||
UNICHAR uni_ch(ch32);
|
||
return std::string(uni_ch.utf8(), uni_ch.utf8_len());
|
||
}
|
||
#endif
|
||
|
||
TEST(NormstrngsTest, BasicText) {
|
||
const char* kBasicText = "AbCd Ef";
|
||
std::string result;
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNormalize,
|
||
GraphemeNorm::kNormalize, kBasicText,
|
||
&result));
|
||
EXPECT_STREQ(kBasicText, result.c_str());
|
||
}
|
||
|
||
TEST(NormstrngsTest, LigatureText) {
|
||
const char* kTwoByteLigText = "ij"; // U+0133 (ij) -> ij
|
||
std::string result;
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNormalize,
|
||
GraphemeNorm::kNormalize, kTwoByteLigText,
|
||
&result));
|
||
EXPECT_STREQ("ij", result.c_str());
|
||
|
||
const char* kThreeByteLigText = "finds"; // U+FB01 (fi) -> fi
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNormalize,
|
||
GraphemeNorm::kNormalize, kThreeByteLigText,
|
||
&result));
|
||
EXPECT_STREQ("finds", result.c_str());
|
||
}
|
||
|
||
TEST(NormstrngsTest, OcrSpecificNormalization) {
|
||
const char* kSingleQuoteText = "‘Hi"; // U+2018 (‘) -> U+027 (')
|
||
std::string result;
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNormalize,
|
||
GraphemeNorm::kNormalize, kSingleQuoteText,
|
||
&result));
|
||
EXPECT_STREQ("'Hi", result.c_str());
|
||
|
||
const char* kDoubleQuoteText = "“Hi"; // U+201C (“) -> U+022 (")
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNormalize,
|
||
GraphemeNorm::kNormalize, kDoubleQuoteText,
|
||
&result));
|
||
EXPECT_STREQ("\"Hi", result.c_str());
|
||
|
||
const char* kEmDash = "Hi—"; // U+2014 (—) -> U+02D (-)
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNormalize,
|
||
GraphemeNorm::kNormalize, kEmDash, &result));
|
||
EXPECT_STREQ("Hi-", result.c_str());
|
||
// Without the ocr normalization, these changes are not made.
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize, kSingleQuoteText,
|
||
&result));
|
||
EXPECT_STREQ(kSingleQuoteText, result.c_str());
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize, kDoubleQuoteText,
|
||
&result));
|
||
EXPECT_STREQ(kDoubleQuoteText, result.c_str());
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize, kEmDash, &result));
|
||
EXPECT_STREQ(kEmDash, result.c_str());
|
||
}
|
||
|
||
// Sample text used in tests.
|
||
const char kEngText[] = "the quick brown fox jumps over the lazy dog";
|
||
const char kHinText[] = "पिताने विवाह की | हो गई उद्विग्न वह सोचा";
|
||
const char kKorText[] = "이는 것으로";
|
||
// Hindi words containing illegal vowel sequences.
|
||
const char* kBadlyFormedHinWords[] = {"उपयोक्ताो", "नहीें", "प्रंात",
|
||
"कहीअे", "पत्रिाका", "छह्णाीस"};
|
||
// Thai illegal sequences.
|
||
const char* kBadlyFormedThaiWords[] = {"ฤิ", "กา้ํ", "กิำ", "นำ้", "เเก"};
|
||
|
||
TEST(NormstrngsTest, DetectsCorrectText) {
|
||
std::string chars;
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize, kEngText, &chars));
|
||
EXPECT_STREQ(kEngText, chars.c_str());
|
||
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize, kHinText, &chars))
|
||
<< "Incorrect text: '" << kHinText << "'";
|
||
EXPECT_STREQ(kHinText, chars.c_str());
|
||
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize, kKorText, &chars));
|
||
EXPECT_STREQ(kKorText, chars.c_str());
|
||
}
|
||
|
||
TEST(NormstrngsTest, DetectsIncorrectText) {
|
||
for (size_t i = 0; i < ARRAYSIZE(kBadlyFormedHinWords); ++i) {
|
||
EXPECT_FALSE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize,
|
||
kBadlyFormedHinWords[i], nullptr))
|
||
<< kBadlyFormedHinWords[i];
|
||
}
|
||
for (size_t i = 0; i < ARRAYSIZE(kBadlyFormedThaiWords); ++i) {
|
||
EXPECT_FALSE(NormalizeUTF8String(UnicodeNormMode::kNFKC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize,
|
||
kBadlyFormedThaiWords[i], nullptr))
|
||
<< kBadlyFormedThaiWords[i];
|
||
}
|
||
}
|
||
|
||
TEST(NormstrngsTest, NonIndicTextDoesntBreakIndicRules) {
|
||
std::string nonindic = "Here's some latin text.";
|
||
std::string dest;
|
||
EXPECT_TRUE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize, nonindic.c_str(),
|
||
&dest))
|
||
<< PrintString32WithUnicodes(nonindic);
|
||
EXPECT_EQ(dest, nonindic);
|
||
}
|
||
|
||
TEST(NormstrngsTest, NoLonelyJoiners) {
|
||
std::string str = "x\u200d\u0d06\u0d34\u0d02";
|
||
std::vector<std::string> glyphs;
|
||
// Returns true, but the joiner is gone.
|
||
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(
|
||
UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNormMode::kCombined, true,
|
||
str.c_str(), &glyphs))
|
||
<< PrintString32WithUnicodes(str);
|
||
EXPECT_EQ(glyphs.size(), 3);
|
||
EXPECT_EQ(glyphs[0], std::string("x"));
|
||
EXPECT_EQ(glyphs[1], std::string("\u0d06"));
|
||
EXPECT_EQ(glyphs[2], std::string("\u0d34\u0d02"));
|
||
}
|
||
|
||
TEST(NormstrngsTest, NoLonelyJoinersPlus) {
|
||
std::string str = "\u0d2a\u200d+\u0d2a\u0d4b";
|
||
std::vector<std::string> glyphs;
|
||
// Returns true, but the joiner is gone.
|
||
EXPECT_TRUE(NormalizeCleanAndSegmentUTF8(
|
||
UnicodeNormMode::kNFC, OCRNorm::kNone, GraphemeNormMode::kCombined, true,
|
||
str.c_str(), &glyphs))
|
||
<< PrintString32WithUnicodes(str);
|
||
EXPECT_EQ(glyphs.size(), 3);
|
||
EXPECT_EQ(glyphs[0], std::string("\u0d2a"));
|
||
EXPECT_EQ(glyphs[1], std::string("+"));
|
||
EXPECT_EQ(glyphs[2], std::string("\u0d2a\u0d4b"));
|
||
}
|
||
|
||
TEST(NormstrngsTest, NoLonelyJoinersNonAlpha) {
|
||
std::string str = "\u200d+\u200c\u200d";
|
||
// Returns true, but the joiners are gone.
|
||
ExpectGraphemeModeResults(str, UnicodeNormMode::kNFC, 1, 1, 1, std::string("+"));
|
||
str = "\u200d\u200c\u200d";
|
||
// Without the plus, the string is invalid.
|
||
std::string result;
|
||
EXPECT_FALSE(NormalizeUTF8String(UnicodeNormMode::kNFC, OCRNorm::kNone,
|
||
GraphemeNorm::kNormalize, str.c_str(),
|
||
&result))
|
||
<< PrintString32WithUnicodes(result);
|
||
}
|
||
|
||
TEST(NormstrngsTest, JoinersStayInArabic) {
|
||
std::string str = "\u0628\u200c\u0628\u200d\u0628";
|
||
// Returns true, string untouched.
|
||
ExpectGraphemeModeResults(str, UnicodeNormMode::kNFC, 5, 5, 2, str);
|
||
}
|
||
|
||
TEST(NormstrngsTest, DigitOK) {
|
||
std::string str = "\u0cea"; // Digit 4.
|
||
ExpectGraphemeModeResults(str, UnicodeNormMode::kNFC, 1, 1, 1, str);
|
||
}
|
||
|
||
TEST(NormstrngsTest, DandaOK) {
|
||
std::string str = "\u0964"; // Single danda.
|
||
ExpectGraphemeModeResults(str, UnicodeNormMode::kNFC, 1, 1, 1, str);
|
||
str = "\u0965"; // Double danda.
|
||
ExpectGraphemeModeResults(str, UnicodeNormMode::kNFC, 1, 1, 1, str);
|
||
}
|
||
|
||
TEST(NormstrngsTest, AllScriptsRegtest) {
|
||
// Tests some valid text in a large number of scripts, some of which were
|
||
// found to be rejected by an earlier version.
|
||
const std::vector<std::pair<std::string, std::string>> kScriptText(
|
||
{{"Arabic",
|
||
" فكان منهم علقمة بن قيس ، و إبراهيم النخعي ، و الأسود بن"
|
||
"توفي بالمدينة في هذه السنة وهو ابن مائة وعشرين سنة "
|
||
"مجموعه هیچ اثری در فنون هنر و ادب و ترجمه، تقدیم پیشگاه ارجمند "
|
||
"سازنده تاریخ نگاه میکرد و به اصطلاح انسان و فطرت انسانی را زیربنای"},
|
||
{"Armenian",
|
||
"անտիկ աշխարհի փիլիսոփաների կենսագրությունը, թե′ նրանց ուս-"
|
||
"պատրաստւում է դալ (բուլամա): Կովկասում կաթից նաև պատ-"
|
||
"Հոգաբարձութեան յղել այդ անձին յիմարութիւնը հաստա-"
|
||
"գծերը եւ միջագծերը կը համրուին վարէն վեր:"},
|
||
{"Bengali",
|
||
"এসে দাঁড়ায় দাও খানি উঁচিয়ে নিয়ে । ঝরনার স্বচ্ছ জলে প্রতিবিম্বিত "
|
||
"পাঠিয়ে, গোবিন্দ স্মরণ করে, নির্ভয়ে রওনা হয়েছিল। তাতে সে "
|
||
"সুলতার। মনে পড়ে বিয়ের সময় বাবা এদের বাড়ি থেকে ঘুরে "
|
||
"কিন্তু তারপর মাতৃহৃদয় কেমন করে আছে? কী"},
|
||
{"Cyrillic",
|
||
"достей, є ще нагороди й почесті, є хай і сумнівна, але слава, "
|
||
"вып., 96б). Параўн. найсвятший у 1 знач., насвятейший у 1 знач., "
|
||
"»Правді«, — гітлерівські окупанти винищували нижчі раси, після дру- "
|
||
"І знов майдан зачорнів од народу. Всередині чоло-"},
|
||
{"Devanagari",
|
||
"डा॰ नै हात्तीमाथि चढेर त्यो भएनेर आइपुगे। राजालाई देखी "
|
||
"बाबतीत लिहिणे ही एक मोठीच जबाबदारी आहे. काकासाहेबांच्या कार्याचा "
|
||
"प्रबंध, आधोगिक प्रबंध तथा बैंकिंग एवम वाणिज्य आदि विषयों में "
|
||
"चित्रकृती दिल्या. शंभराहून अधिक देश आज आपापले चित्रपट निर्माण करीत"},
|
||
{"Greek",
|
||
"Μέσα ένα τετράδιο είχα στριμώξει το πρώτο "
|
||
"νον αξίως τού ευαγγελίου τού χριστού πολιτεύεσθε, ίνα "
|
||
"οὐδεμία ὑπ' αὐτοῦ μνεία γίνεται τῶν οἰκείων χωρίων. "
|
||
"είτα την φάσιν αυτήν ην ούτος εποιήσατο κατά του Μίκω-"},
|
||
{"Gujarati",
|
||
"ઉપહારગૃહે ને નાટ્યસ્થળે આ એ જ તેલ કડકડતું "
|
||
"શકી. ભાવવધારો અટકાવી નથી શકી અને બેકારીને "
|
||
"ત્યાં વાંકુથી પાછે આવ્યો, ચોરીનો માલ સોંપવા ! "
|
||
"કહી. એણે રેશમના કપડામાં વીંટી રાખેલ કુંવરીની છબી"},
|
||
{"Gurmukhi",
|
||
"ਯਾਦ ਰਹੇ ਕਿ ‘ਨਫਰਤ ’ ਦਾ ਵਿਸ਼ਾ ਕ੍ਰਾਤੀ ਨਹੀ ਹੈ ਅਤੇ ਕਵੀ ਦੀ ਇਹ "
|
||
"ਮਹਾਂ ਨੰਦਾ ਕੋਲ ਇਕ ਚੀਜ਼ ਸੀ ਉਹ ਸੀ ਸਚ, ਕੋਰਾ ਸਚ, ਬੇਧਤ੍ਰਕ ਕਹਿੳ "
|
||
"ਭੂਰਾ ਸਾਨੂੰ ਥੜਾ ਚੰਗਾ ਲਗਦਾ ਸੀ । ਉਸ ਦਾ ਇਕ ਪੈਰ ਜਨਮ ਤੋ "
|
||
"ਨੂੰ ਇਹ ਅਧਿਕਾਰ ਦਿੱਤਾ ਕਿ ਉਹ ਸਿੱਖ ਵਿਰੋਧ ਦਾ ਸੰਗਠਨ ਕਰੇ ਅਤੇ 3 ਸਤੰਬਰ,"},
|
||
{"Hangul",
|
||
"로 들어갔다. 이대통령은 아이젠하워 대통령의 뒷모습을 보면서 "
|
||
"그것뿐인 줄 아요? 노름도 했다 캅니다. 빌어묵을 놈이 그러 "
|
||
"의 가장 과학적 태도이며, 우리 역사를 가장 정확하게 학습할 수 있는 "
|
||
"마르크스 레"
|
||
"각하는 그는 그들의 식사보장을 위해 때때로 집에"},
|
||
{"HanS",
|
||
"大凡世界上的先生可 分 三 种： 第一种只会教书， 只会拿一 "
|
||
"书像是探宝一样，在茶叶店里我买过西湖龙井﹑黄山毛峰﹑福建的铁观音﹑大红"
|
||
" "
|
||
"持 “左” 倾冒险主义的干部，便扣上 “富农 "
|
||
"笑说：“我听说了，王总工程师也跟我说过了，只是工作忙，谁"},
|
||
{"HanT",
|
||
"叁、 銀行資產管理的群組分析模式 "
|
||
"民國六十三年，申請就讀台灣大學歷史研究所，並從事著述，"
|
||
"質言之﹐在社會結構中﹐性質﹑特徵﹑地位相類似的一羣人﹐由於 "
|
||
"董橋，一九四二年生，福建晉江人，國立成功大學外"},
|
||
{"Hebrew",
|
||
" אֵ-לִי, אֵ-לִי, כֵּיַצד מְטַפְּסִים בְּקִירוֹת שֶׁל זְכוּכִי"
|
||
" הראשון חוצה אותי שוב. אני בסיבוב הרביעי, הוא בטח מתחיל את"
|
||
" ווערטער געהאט, אבער דער עיקר איז ניט דאָס וואָרט, נאָר"
|
||
" על גחלת היהדות המקורית בעירך, נתת צביון ואופי מיוחד"},
|
||
{"Japanese",
|
||
"は異民族とみなされていた。楚の荘王（前613〜前 "
|
||
"を詳細に吟味する。実際の治療活動の領域は便宜上、(1) 障害者 "
|
||
"困難性は多角企業の場合原則として部門別に判断されている.). "
|
||
"☆ご希望の団体には見本をお送りします"},
|
||
{"Kannada",
|
||
"ಕೂಡ ಯುದ್ಧ ಮಾಡಿ ಜಯಪಡೆ. ನಂತರ ನಗರದೊಳಕ್ಕೆ ನಡೆ ಇದನ್ನು "
|
||
"ಅಸಹ್ಯದೃಶ್ಯ ಯಾರಿಗಾದರೂ ನಾಚಿಕೆತರುವಂತಹದಾಗಿದೆ. ಆರೋಗ್ಯ ದೃಷ್ಟಿ "
|
||
"ಯಾಗಲಿ, ಮೋಹನನಾಗಲಿ ಇಂಥ ಬಿಸಿಲಿನಲ್ಲಿ ಎಂದೂ ಬಹಳ ಹೊತ್ತು "
|
||
"\"ಇದೆ...ಖಂಡಿತಾ ಇದೆ\" ಅಂದ ಮನಸ್ಸಿನಲ್ಲಿಯೇ ವಂದಿಸುತ್ತಾ,"},
|
||
{"Khmer",
|
||
"សិតសក់និងផ្លាស់សម្លៀកបំពាក់ពេលយប់ចេញ។ "
|
||
"និយាយអំពីនគរនេះ ប្រាប់ដល់លោកទាំងមូលឲ្យដឹងច្បាស់លាស់អំពី "
|
||
"កន្លះកាថាសម្រាប់ទន្ទេញឲ្យងាយចាំ បោះពុម្ពនៅក្នុងទ្រង់ទ្រាយបច្ចុប្បន្ន "
|
||
"ឯកសារនេះបានផ្សព្វផ្សាយនៅក្នុងសន្និសីទ"},
|
||
{"Lao",
|
||
"ເອີຍ ! ຟັງສຽງຟ້າມັນຮ້ອງຮ່ວນ ມັນດັງໄກໆ ເອີຍ "
|
||
"ໄດລຽງດູລາວມາດວບຄວາມລາບາກຫລາຍ; "
|
||
"ບາງໄດ້ ເຈົ້າລອງສູ້ບໍ່ໄດ້ຈຶ່ງຫນີລົງມາວຽງຈັນ. "
|
||
"ລົບອອກຈາກ 3 ເຫລືອ 1, ຂ້ອຍຂຽນ 1 (1)"},
|
||
{"Latin",
|
||
"režisoru, palīdzēja to manu domīgo, kluso Dzejas metru ielikt "
|
||
"Ešte nedávno sa chcel mladý Novomeský „liečiť” "
|
||
"tiivisia kysymyksiä, mistä seuraa, että spekula- | don luonteesta "
|
||
"Grabiel Sanchez, yang bertani selama 120 tahun meninggal"},
|
||
{"Malayalam",
|
||
"അമൂർത്തചിത്രമായിരിക്കും. ഛേ! ആ വീട്ടിലേക്ക് അവളൊന്നിച്ച് പോകേണ്ടതാ "
|
||
"മൃഗങ്ങൾക്ക് എന്തെക്കിലും പറ്റിയാൽ മാത്രം ഞാനതു "
|
||
"വെലക്ക് വേണമെങ്കിൽ തരാം. എന്തോ തരും? പറ. "
|
||
"എല്ലാം കഴിഞ്ഞ് സീനിയറിന്റെ അടുത്തു ചെന്ന് കാൽതൊട്ട"},
|
||
{"Tamil",
|
||
"பொருத்தமாகப் பாடினாள் நம் ஔவைப் பாட்டி. காவிரி "
|
||
"உள்ளடக்கி நிற்பது விநோத வார்த்தையின் அஃறிணை "
|
||
"சூரிய கிரஹண சமயத்தில் குருக்ஷேத்திரம் செல்வது "
|
||
"காலங்களில் வெளியே போகும்பொழுது, 'ஸார்', 'ஸார்',"},
|
||
{"Telugu",
|
||
"1892లో ఆమె 10వ సంవత్సరంలో గుంటూరు తాలూకా వేములాపాడు "
|
||
"ఫండ్స్ చట్టము'నందు చేయబడెను. తరువాత క్రీ. శ. "
|
||
"సంచారము చేయును. మీరు ఇప్పుడే కాళకాలయమునకు "
|
||
"ఎంతటి సరళమైన భాషలో వ్రాశాడో విశదమవుతుంది. పైగా ఆనాటి భాష"},
|
||
{"Thai",
|
||
"อ้อ! กับนัง....แม่ยอดพระกลิ่น นั่นเอง ! หรับก็ย่อมจะรู้โดยชัดเจนว่า "
|
||
"ถ้าตราบใดยังมีเรือปืนอยู่ใกล้ ๆ แล้ว ตราบนั้น "
|
||
"พระดำรินี้ ที่มีคตีทำกรวยหมากและธูปเทียน "
|
||
"อันยานมีเรือเปนต้นฃ้ามยาก ฯ เพราะว่าแม่น้ำนั่นมีน้ำใสยิ่ง แม้เพียง"},
|
||
{"Vietnamese",
|
||
"vợ đến tai mụ hung thần Xăng-tô- mê-a. Mụ vô cùng "
|
||
"chiếc xe con gấu chạy qua nhà. Nhưng thỉnh thoảng "
|
||
"hòa hoãn với người Pháp để cho họ được dựng một ngôi nhà thờ nhỏ bằng "
|
||
"Cặp câu đói súc tích mà sâu sắc, là lời chúc lời"}});
|
||
|
||
for (const auto& p : kScriptText) {
|
||
std::string normalized;
|
||
EXPECT_TRUE(tesseract::NormalizeUTF8String(
|
||
tesseract::UnicodeNormMode::kNFKC, tesseract::OCRNorm::kNormalize,
|
||
tesseract::GraphemeNorm::kNormalize, p.second.c_str(), &normalized))
|
||
<< "Script=" << p.first << " text=" << p.second;
|
||
}
|
||
}
|
||
|
||
TEST(NormstrngsTest, IsWhitespace) {
|
||
// U+0020 is whitespace
|
||
EXPECT_TRUE(IsWhitespace(' '));
|
||
EXPECT_TRUE(IsWhitespace('\t'));
|
||
EXPECT_TRUE(IsWhitespace('\r'));
|
||
EXPECT_TRUE(IsWhitespace('\n'));
|
||
// U+2000 through U+200A
|
||
for (char32 ch = 0x2000; ch <= 0x200A; ++ch) {
|
||
SCOPED_TRACE(absl::StrFormat("Failed at U+%x", ch));
|
||
EXPECT_TRUE(IsWhitespace(ch));
|
||
}
|
||
// U+3000 is whitespace
|
||
EXPECT_TRUE(IsWhitespace(0x3000));
|
||
// ZWNBSP is not considered a space.
|
||
EXPECT_FALSE(IsWhitespace(0xFEFF));
|
||
}
|
||
|
||
TEST(NormstrngsTest, SpanUTF8Whitespace) {
|
||
EXPECT_EQ(4, SpanUTF8Whitespace(" \t\r\n"));
|
||
EXPECT_EQ(4, SpanUTF8Whitespace(" \t\r\nabc"));
|
||
EXPECT_EQ(0, SpanUTF8Whitespace("abc \t\r\nabc"));
|
||
EXPECT_EQ(0, SpanUTF8Whitespace(""));
|
||
}
|
||
|
||
TEST(NormstrngsTest, SpanUTF8NotWhitespace) {
|
||
const char kHinText[] = "पिताने विवाह";
|
||
const char kKorText[] = "이는 것으로 다시 넣을";
|
||
const char kMixedText[] = "والفكر 123 والصراع abc";
|
||
|
||
EXPECT_EQ(0, SpanUTF8NotWhitespace(""));
|
||
EXPECT_EQ(0, SpanUTF8NotWhitespace(" abc"));
|
||
EXPECT_EQ(0, SpanUTF8NotWhitespace("\rabc"));
|
||
EXPECT_EQ(0, SpanUTF8NotWhitespace("\tabc"));
|
||
EXPECT_EQ(0, SpanUTF8NotWhitespace("\nabc"));
|
||
EXPECT_EQ(3, SpanUTF8NotWhitespace("abc def"));
|
||
EXPECT_EQ(18, SpanUTF8NotWhitespace(kHinText));
|
||
EXPECT_EQ(6, SpanUTF8NotWhitespace(kKorText));
|
||
EXPECT_EQ(12, SpanUTF8NotWhitespace(kMixedText));
|
||
}
|
||
|
||
// Test that the method clones the util/utf8/unilib definition of
|
||
// interchange validity.
|
||
TEST(NormstrngsTest, IsInterchangeValid) {
|
||
#ifdef INCLUDE_TENSORFLOW
|
||
const int32_t kMinUnicodeValue = 33;
|
||
const int32_t kMaxUnicodeValue = 0x10FFFF;
|
||
for (int32_t ch = kMinUnicodeValue; ch <= kMaxUnicodeValue; ++ch) {
|
||
SCOPED_TRACE(absl::StrFormat("Failed at U+%x", ch));
|
||
EXPECT_EQ(UniLib::IsInterchangeValid(ch), IsInterchangeValid(ch));
|
||
}
|
||
#else
|
||
GTEST_SKIP();
|
||
#endif
|
||
}
|
||
|
||
// Test that the method clones the util/utf8/unilib definition of
|
||
// 7-bit ASCII interchange validity.
|
||
TEST(NormstrngsTest, IsInterchangeValid7BitAscii) {
|
||
#if defined(MISSING_CODE) && defined(INCLUDE_TENSORFLOW)
|
||
const int32_t kMinUnicodeValue = 33;
|
||
const int32_t kMaxUnicodeValue = 0x10FFFF;
|
||
for (int32_t ch = kMinUnicodeValue; ch <= kMaxUnicodeValue; ++ch) {
|
||
SCOPED_TRACE(absl::StrFormat("Failed at U+%x", ch));
|
||
std::string str = EncodeAsUTF8(ch);
|
||
EXPECT_EQ(UniLib::IsInterchangeValid7BitAscii(str),
|
||
IsInterchangeValid7BitAscii(ch));
|
||
}
|
||
#else
|
||
// Skipped because of missing UniLib::IsInterchangeValid7BitAscii.
|
||
GTEST_SKIP();
|
||
#endif
|
||
}
|
||
|
||
// Test that the method clones the util/utf8/unilib definition of
|
||
// fullwidth-halfwidth .
|
||
TEST(NormstrngsTest, FullwidthToHalfwidth) {
|
||
// U+FF21 -> U+0041 (Latin capital letter A)
|
||
EXPECT_EQ('A', FullwidthToHalfwidth(0xFF21));
|
||
// U+FF05 -> U+0025 (percent sign)
|
||
EXPECT_EQ('%', FullwidthToHalfwidth(0xFF05));
|
||
// U+FFE6 -> U+20A9 (won sign)
|
||
EXPECT_EQ(0x20A9, FullwidthToHalfwidth(0xFFE6));
|
||
|
||
#if defined(MISSING_CODE) && defined(INCLUDE_TENSORFLOW)
|
||
// Skipped because of missing UniLib::FullwidthToHalfwidth.
|
||
const int32_t kMinUnicodeValue = 33;
|
||
const int32_t kMaxUnicodeValue = 0x10FFFF;
|
||
for (int32_t ch = kMinUnicodeValue; ch <= kMaxUnicodeValue; ++ch) {
|
||
if (!IsValidCodepoint(ch)) continue;
|
||
SCOPED_TRACE(absl::StrFormat("Failed at U+%x", ch));
|
||
std::string str = EncodeAsUTF8(ch);
|
||
const std::string expected_half_str =
|
||
UniLib::FullwidthToHalfwidth(str.c_str(), str.length(), true);
|
||
EXPECT_EQ(expected_half_str, EncodeAsUTF8(FullwidthToHalfwidth(ch)));
|
||
}
|
||
#endif
|
||
}
|
||
|
||
} // namespace
|
||
} // namespace tesseract
|