1 // Copyright (C) 2013 Google Inc. 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 #include "rule.h" 16 17 #include <libaddressinput/address_field.h> 18 19 #include <string> 20 #include <utility> 21 #include <vector> 22 23 #include <gtest/gtest.h> 24 25 #include "region_data_constants.h" 26 27 namespace { 28 29 using i18n::addressinput::AddressField; 30 using i18n::addressinput::ADMIN_AREA; 31 using i18n::addressinput::COUNTRY; 32 using i18n::addressinput::FormatElement; 33 using i18n::addressinput::LOCALITY; 34 using i18n::addressinput::ORGANIZATION; 35 using i18n::addressinput::POSTAL_CODE; 36 using i18n::addressinput::RECIPIENT; 37 using i18n::addressinput::RegionDataConstants; 38 using i18n::addressinput::Rule; 39 using i18n::addressinput::STREET_ADDRESS; 40 41 bool IsFormatEmpty(const std::vector<std::vector<FormatElement> >& format) { 42 for (std::vector<std::vector<FormatElement> >::const_iterator 43 it = format.begin(); 44 it != format.end(); 45 ++it) { 46 if (!it->empty()) { 47 return false; 48 } 49 } 50 return true; 51 } 52 53 TEST(RuleTest, CopyOverwritesRule) { 54 Rule rule; 55 ASSERT_TRUE(rule.ParseSerializedRule( 56 "{" 57 "\"fmt\":\"%S%Z\"," 58 "\"require\":\"SZ\"," 59 "\"state_name_type\":\"area\"," 60 "\"zip_name_type\":\"postal\"," 61 "\"sub_keys\":\"CA~NY~TX\"," 62 "\"lang\":\"en\"," 63 "\"languages\":\"en~fr\"," 64 "\"zip\":\"\\\\d{5}([ \\\\-]\\\\d{4})?\"" 65 "}")); 66 67 Rule copy; 68 EXPECT_NE(rule.GetFormat(), copy.GetFormat()); 69 EXPECT_NE(rule.GetRequired(), copy.GetRequired()); 70 EXPECT_NE(rule.GetSubKeys(), copy.GetSubKeys()); 71 EXPECT_NE(rule.GetLanguages(), copy.GetLanguages()); 72 EXPECT_NE(rule.GetLanguage(), copy.GetLanguage()); 73 EXPECT_NE(rule.GetPostalCodeFormat(), copy.GetPostalCodeFormat()); 74 EXPECT_NE(rule.GetAdminAreaNameType(), 75 copy.GetAdminAreaNameType()); 76 EXPECT_NE(rule.GetPostalCodeNameType(), 77 copy.GetPostalCodeNameType()); 78 79 copy.CopyFrom(rule); 80 EXPECT_EQ(rule.GetFormat(), copy.GetFormat()); 81 EXPECT_EQ(rule.GetRequired(), copy.GetRequired()); 82 EXPECT_EQ(rule.GetSubKeys(), copy.GetSubKeys()); 83 EXPECT_EQ(rule.GetLanguages(), copy.GetLanguages()); 84 EXPECT_EQ(rule.GetLanguage(), copy.GetLanguage()); 85 EXPECT_EQ(rule.GetPostalCodeFormat(), copy.GetPostalCodeFormat()); 86 EXPECT_EQ(rule.GetAdminAreaNameType(), 87 copy.GetAdminAreaNameType()); 88 EXPECT_EQ(rule.GetPostalCodeNameType(), 89 copy.GetPostalCodeNameType()); 90 } 91 92 TEST(RuleTest, ParseOverwritesRule) { 93 Rule rule; 94 ASSERT_TRUE(rule.ParseSerializedRule( 95 "{" 96 "\"fmt\":\"%S%Z\"," 97 "\"require\":\"SZ\"," 98 "\"state_name_type\":\"area\"," 99 "\"zip_name_type\":\"postal\"," 100 "\"sub_keys\":\"CA~NY~TX\"," 101 "\"lang\":\"en\"," 102 "\"languages\":\"en~fr\"," 103 "\"zip\":\"\\\\d{5}([ \\\\-]\\\\d{4})?\"" 104 "}")); 105 EXPECT_FALSE(IsFormatEmpty(rule.GetFormat())); 106 EXPECT_FALSE(rule.GetRequired().empty()); 107 EXPECT_FALSE(rule.GetSubKeys().empty()); 108 EXPECT_FALSE(rule.GetLanguages().empty()); 109 EXPECT_FALSE(rule.GetLanguage().empty()); 110 EXPECT_FALSE(rule.GetPostalCodeFormat().empty()); 111 EXPECT_EQ("area", rule.GetAdminAreaNameType()); 112 EXPECT_EQ("postal", rule.GetPostalCodeNameType()); 113 114 ASSERT_TRUE(rule.ParseSerializedRule( 115 "{" 116 "\"fmt\":\"\"," 117 "\"require\":\"\"," 118 "\"state_name_type\":\"do_si\"," 119 "\"zip_name_type\":\"zip\"," 120 "\"sub_keys\":\"\"," 121 "\"lang\":\"\"," 122 "\"languages\":\"\"," 123 "\"zip\":\"\"" 124 "}")); 125 EXPECT_TRUE(IsFormatEmpty(rule.GetFormat())); 126 EXPECT_TRUE(rule.GetRequired().empty()); 127 EXPECT_TRUE(rule.GetSubKeys().empty()); 128 EXPECT_TRUE(rule.GetLanguages().empty()); 129 EXPECT_TRUE(rule.GetLanguage().empty()); 130 EXPECT_TRUE(rule.GetPostalCodeFormat().empty()); 131 EXPECT_EQ("do_si", rule.GetAdminAreaNameType()); 132 EXPECT_EQ("zip", rule.GetPostalCodeNameType()); 133 } 134 135 TEST(RuleTest, ParseEmptyDataDoesNotOverwriteRule) { 136 Rule rule; 137 ASSERT_TRUE(rule.ParseSerializedRule( 138 "{" 139 "\"fmt\":\"%S%Z\"," 140 "\"require\":\"SZ\"," 141 "\"state_name_type\":\"area\"," 142 "\"zip_name_type\":\"postal\"," 143 "\"sub_keys\":\"CA~NY~TX\"," 144 "\"lang\":\"en\"," 145 "\"languages\":\"en~fr\"," 146 "\"zip\":\"\\\\d{5}([ \\\\-]\\\\d{4})?\"" 147 "}")); 148 149 Rule copy; 150 copy.CopyFrom(rule); 151 ASSERT_TRUE(copy.ParseSerializedRule("{}")); 152 153 EXPECT_EQ(rule.GetFormat(), copy.GetFormat()); 154 EXPECT_EQ(rule.GetRequired(), copy.GetRequired()); 155 EXPECT_EQ(rule.GetSubKeys(), copy.GetSubKeys()); 156 EXPECT_EQ(rule.GetLanguages(), copy.GetLanguages()); 157 EXPECT_EQ(rule.GetLanguage(), copy.GetLanguage()); 158 EXPECT_EQ(rule.GetPostalCodeFormat(), copy.GetPostalCodeFormat()); 159 EXPECT_EQ(rule.GetAdminAreaNameType(), 160 copy.GetAdminAreaNameType()); 161 EXPECT_EQ(rule.GetPostalCodeNameType(), 162 copy.GetPostalCodeNameType()); 163 } 164 165 TEST(RuleTest, ParseFormatWithNewLines) { 166 Rule rule; 167 ASSERT_TRUE( 168 rule.ParseSerializedRule("{\"fmt\":\"%O%n%N%n%A%nAX-%Z %C%nLAND\"}")); 169 170 std::vector<std::vector<FormatElement> > expected_format; 171 expected_format.push_back( 172 std::vector<FormatElement>(1, FormatElement(ORGANIZATION))); 173 expected_format.push_back( 174 std::vector<FormatElement>(1, FormatElement(RECIPIENT))); 175 expected_format.push_back( 176 std::vector<FormatElement>(1, FormatElement(STREET_ADDRESS))); 177 expected_format.push_back( 178 std::vector<FormatElement>(1, FormatElement("AX-"))); 179 expected_format.back().push_back(FormatElement(POSTAL_CODE)); 180 expected_format.back().push_back(FormatElement(" ")); 181 expected_format.back().push_back(FormatElement(LOCALITY)); 182 expected_format.push_back( 183 std::vector<FormatElement>(1, FormatElement("LAND"))); 184 185 EXPECT_EQ(expected_format, rule.GetFormat()); 186 } 187 188 TEST(RuleTest, DoubleTokenPrefixDoesNotCrash) { 189 Rule rule; 190 EXPECT_TRUE(rule.ParseSerializedRule("{\"fmt\":\"%%R\"}")); 191 } 192 193 TEST(RuleTest, DoubleNewlineFormatDoesNotCrash) { 194 Rule rule; 195 EXPECT_TRUE(rule.ParseSerializedRule("{\"fmt\":\"%n%n\"}")); 196 } 197 198 TEST(RuleTest, FormatTokenWithoutPrefixDoesNotCrash) { 199 Rule rule; 200 ASSERT_TRUE(rule.ParseSerializedRule("{\"fmt\":\"R\"}")); 201 } 202 203 TEST(RuleTest, ParseDuplicateTokenInFormatDoesNotCrash) { 204 Rule rule; 205 EXPECT_TRUE(rule.ParseSerializedRule("{\"fmt\":\"%R%R\"}")); 206 } 207 208 TEST(RuleTest, ParseInvalidFormatFieldsDoesNotCrash) { 209 Rule rule; 210 EXPECT_TRUE(rule.ParseSerializedRule("{\"fmt\":\"%K%L\"}")); 211 } 212 213 TEST(RuleTest, PrefixWithoutTokenFormatDoesNotCrash) { 214 Rule rule; 215 EXPECT_TRUE(rule.ParseSerializedRule("{\"fmt\":\"%\"}")); 216 } 217 218 TEST(RuleTest, EmptyStringFormatDoesNotCrash) { 219 Rule rule; 220 EXPECT_TRUE(rule.ParseSerializedRule("{\"fmt\":\"\"}")); 221 } 222 223 TEST(RuleTest, ParseRequiredFields) { 224 Rule rule; 225 ASSERT_TRUE(rule.ParseSerializedRule("{\"require\":\"ONAZC\"}")); 226 std::vector<AddressField> expected; 227 expected.push_back(ORGANIZATION); 228 expected.push_back(RECIPIENT); 229 expected.push_back(STREET_ADDRESS); 230 expected.push_back(POSTAL_CODE); 231 expected.push_back(LOCALITY); 232 EXPECT_EQ(expected, rule.GetRequired()); 233 } 234 235 TEST(RuleTest, ParseEmptyStringRequiredFields) { 236 Rule rule; 237 ASSERT_TRUE(rule.ParseSerializedRule("{\"require\":\"\"}")); 238 EXPECT_TRUE(rule.GetRequired().empty()); 239 } 240 241 TEST(RuleTest, ParseInvalidRequiredFields) { 242 Rule rule; 243 ASSERT_TRUE(rule.ParseSerializedRule("{\"require\":\"garbage\"}")); 244 EXPECT_TRUE(rule.GetRequired().empty()); 245 } 246 247 TEST(RuleTest, ParseDuplicateRequiredFields) { 248 Rule rule; 249 ASSERT_TRUE(rule.ParseSerializedRule("{\"require\":\"SSS\"}")); 250 EXPECT_EQ(std::vector<AddressField>(3, ADMIN_AREA), rule.GetRequired()); 251 } 252 253 TEST(RuleTest, ParsesSubKeysCorrectly) { 254 Rule rule; 255 ASSERT_TRUE(rule.ParseSerializedRule("{\"sub_keys\":\"CA~NY~TX\"}")); 256 std::vector<std::string> expected; 257 expected.push_back("CA"); 258 expected.push_back("NY"); 259 expected.push_back("TX"); 260 EXPECT_EQ(expected, rule.GetSubKeys()); 261 } 262 263 TEST(RuleTest, ParsesLanguageCorrectly) { 264 Rule rule; 265 ASSERT_TRUE(rule.ParseSerializedRule("{\"lang\":\"en\"}")); 266 EXPECT_EQ("en", rule.GetLanguage()); 267 } 268 269 TEST(RuleTest, ParsesLanguagesCorrectly) { 270 Rule rule; 271 ASSERT_TRUE(rule.ParseSerializedRule("{\"languages\":\"de~fr~it\"}")); 272 std::vector<std::string> expected; 273 expected.push_back("de"); 274 expected.push_back("fr"); 275 expected.push_back("it"); 276 EXPECT_EQ(expected, rule.GetLanguages()); 277 } 278 279 TEST(RuleTest, ParsesPostalCodeFormatCorrectly) { 280 Rule rule; 281 ASSERT_TRUE(rule.ParseSerializedRule( 282 "{" 283 "\"zip\":\"\\\\d{5}([ \\\\-]\\\\d{4})?\"" 284 "}")); 285 EXPECT_EQ("\\d{5}([ \\-]\\d{4})?", rule.GetPostalCodeFormat()); 286 } 287 288 TEST(RuleTest, EmptyStringIsNotValid) { 289 Rule rule; 290 EXPECT_FALSE(rule.ParseSerializedRule(std::string())); 291 } 292 293 TEST(RuleTest, EmptyDictionaryIsValid) { 294 Rule rule; 295 EXPECT_TRUE(rule.ParseSerializedRule("{}")); 296 } 297 298 TEST(RuleTest, ParseSubKeyTest) { 299 i18n::addressinput::Rule rule; 300 ASSERT_TRUE(rule.ParseSerializedRule( 301 "{ \"sub_keys\": \"FOO~BAR~BAZ\"," 302 " \"sub_names\": \"Foolandia~Bartopolis~Bazmonia\"," 303 " \"sub_lnames\": \"Foolandia2~Bartopolis2~Bazmonia2\" }")); 304 EXPECT_EQ(3U, rule.GetSubKeys().size()); 305 306 std::string sub_key; 307 EXPECT_TRUE(rule.CanonicalizeSubKey("BAR", false, &sub_key)); 308 EXPECT_EQ("BAR", sub_key); 309 sub_key.clear(); 310 311 EXPECT_TRUE(rule.CanonicalizeSubKey("Bartopolis", false, &sub_key)); 312 EXPECT_EQ("BAR", sub_key); 313 sub_key.clear(); 314 315 // Unlatinize. 316 EXPECT_TRUE(rule.CanonicalizeSubKey("Bartopolis2", false, &sub_key)); 317 EXPECT_EQ("BAR", sub_key); 318 sub_key.clear(); 319 320 // Keep input latin. 321 EXPECT_TRUE(rule.CanonicalizeSubKey("Bartopolis2", true, &sub_key)); 322 EXPECT_EQ("Bartopolis2", sub_key); 323 sub_key.clear(); 324 325 EXPECT_FALSE(rule.CanonicalizeSubKey("Beertopia", false, &sub_key)); 326 EXPECT_EQ("", sub_key); 327 } 328 329 struct LabelData { 330 LabelData(const std::string& data, const std::string& name_type) 331 : data(data), name_type(name_type) {} 332 333 ~LabelData() {} 334 335 std::string data; 336 std::string name_type; 337 }; 338 339 // Tests for parsing the postal code name. 340 class PostalCodeNameParseTest : public testing::TestWithParam<LabelData> { 341 protected: 342 Rule rule_; 343 }; 344 345 // Verifies that a postal code name is parsed correctly. 346 TEST_P(PostalCodeNameParseTest, ParsedCorrectly) { 347 ASSERT_TRUE(rule_.ParseSerializedRule(GetParam().data)); 348 EXPECT_EQ(GetParam().name_type, rule_.GetPostalCodeNameType()); 349 } 350 351 // Test parsing all postal code names. 352 INSTANTIATE_TEST_CASE_P( 353 AllPostalCodeNames, PostalCodeNameParseTest, 354 testing::Values( 355 LabelData("{\"zip_name_type\":\"postal\"}", "postal"), 356 LabelData("{\"zip_name_type\":\"zip\"}", "zip"))); 357 358 // Tests for parsing the administrative area name. 359 class AdminAreaNameParseTest : public testing::TestWithParam<LabelData> { 360 protected: 361 Rule rule_; 362 }; 363 364 // Verifies that an administrative area name is parsed correctly. 365 TEST_P(AdminAreaNameParseTest, ParsedCorrectly) { 366 ASSERT_TRUE(rule_.ParseSerializedRule(GetParam().data)); 367 EXPECT_EQ(GetParam().name_type, rule_.GetAdminAreaNameType()); 368 } 369 370 // Test parsing all administrative area names. 371 INSTANTIATE_TEST_CASE_P( 372 AllAdminAreaNames, AdminAreaNameParseTest, 373 testing::Values( 374 LabelData("{\"state_name_type\":\"area\"}", "area"), 375 LabelData("{\"state_name_type\":\"county\"}", "county"), 376 LabelData("{\"state_name_type\":\"department\"}", "department"), 377 LabelData("{\"state_name_type\":\"district\"}", "district"), 378 LabelData("{\"state_name_type\":\"do_si\"}", "do_si"), 379 LabelData("{\"state_name_type\":\"emirate\"}", "emirate"), 380 LabelData("{\"state_name_type\":\"island\"}", "island"), 381 LabelData("{\"state_name_type\":\"parish\"}", "parish"), 382 LabelData("{\"state_name_type\":\"prefecture\"}", "prefecture"), 383 LabelData("{\"state_name_type\":\"province\"}", "province"), 384 LabelData("{\"state_name_type\":\"state\"}", "state"))); 385 386 // Verifies that an address format does not contain consecutive lines with 387 // multiple fields each. Such address format (e.g. {{ELEMENT, ELEMENT}, 388 // {ELEMENT, ELEMENT}}) will result in incorrect behavior of BuildComponents() 389 // public API. 390 TEST(RuleParseTest, ConsecutiveLinesWithMultipleFields) { 391 const std::vector<std::string>& region_codes = 392 RegionDataConstants::GetRegionCodes(); 393 Rule rule; 394 for (size_t i = 0; i < region_codes.size(); ++i) { 395 const std::string& region_data = 396 RegionDataConstants::GetRegionData(region_codes[i]); 397 SCOPED_TRACE(region_codes[i] + ": " + region_data); 398 399 ASSERT_TRUE(rule.ParseSerializedRule(region_data)); 400 bool previous_line_has_single_field = true; 401 for (std::vector<std::vector<FormatElement> >::const_iterator 402 line_it = rule.GetFormat().begin(); 403 line_it != rule.GetFormat().end(); 404 ++line_it) { 405 int num_fields = 0; 406 for (std::vector<FormatElement>::const_iterator 407 element_it = line_it->begin(); 408 element_it != line_it->end(); 409 ++element_it) { 410 if (element_it->IsField()) { 411 ++num_fields; 412 } 413 } 414 if (num_fields == 0) { 415 continue; 416 } 417 ASSERT_TRUE(num_fields == 1 || previous_line_has_single_field); 418 previous_line_has_single_field = num_fields == 1; 419 } 420 } 421 } 422 423 // Verifies that a street line is surrounded by either newlines or spaces. A 424 // different format will result in incorrect behavior in 425 // AddressData::BuildDisplayLines(). 426 TEST(RuleParseTest, StreetAddressSurroundingElements) { 427 const std::vector<std::string>& region_codes = 428 RegionDataConstants::GetRegionCodes(); 429 Rule rule; 430 for (size_t i = 0; i < region_codes.size(); ++i) { 431 const std::string& region_data = 432 RegionDataConstants::GetRegionData(region_codes[i]); 433 SCOPED_TRACE(region_codes[i] + ": " + region_data); 434 435 ASSERT_TRUE(rule.ParseSerializedRule(region_data)); 436 for (std::vector<std::vector<FormatElement> >::const_iterator 437 line_it = rule.GetFormat().begin(); 438 line_it != rule.GetFormat().end(); 439 ++line_it) { 440 for (size_t i = 0; i < line_it->size(); ++i) { 441 const FormatElement& element = line_it->at(i); 442 if (element.IsField() && element.field == STREET_ADDRESS) { 443 bool surrounded_by_newlines = line_it->size() == 1; 444 bool surrounded_by_spaces = 445 i > 0 && 446 i < line_it->size() - 1 && 447 !line_it->at(i - 1).IsField() && 448 line_it->at(i - 1).literal == " " && 449 !line_it->at(i + 1).IsField() && 450 line_it->at(i + 1).literal == " "; 451 EXPECT_TRUE(surrounded_by_newlines || surrounded_by_spaces); 452 } 453 } 454 } 455 } 456 } 457 458 } // namespace 459