matecat /
xliff-parser
| Total Complexity | 78 |
| Total Lines | 425 |
| Duplicated Lines | 0 % |
| Changes | 1 | ||
| Bugs | 0 | Features | 0 |
Complex classes like XliffParserV2 often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
While breaking up the class, it is a good idea to analyze how other classes use XliffParserV2, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 14 | class XliffParserV2 extends AbstractXliffParser { |
||
| 15 | /** |
||
| 16 | * @inheritDoc |
||
| 17 | * @throws Exception |
||
| 18 | */ |
||
| 19 | public function parse( DOMDocument $dom, ?array $output = [] ): array { |
||
| 20 | $i = 1; |
||
| 21 | /** @var DOMElement $file */ |
||
| 22 | foreach ( $dom->getElementsByTagName( 'file' ) as $file ) { |
||
| 23 | |||
| 24 | // metadata |
||
| 25 | $output[ 'files' ][ $i ][ 'attr' ] = $this->extractMetadata( $dom ); |
||
| 26 | |||
| 27 | // notes |
||
| 28 | $output[ 'files' ][ $i ][ 'notes' ] = $this->extractNotes( $file ); |
||
| 29 | |||
| 30 | // trans-units |
||
| 31 | $transUnitIdArrayForUniquenessCheck = []; |
||
| 32 | $j = 1; |
||
| 33 | /** @var DOMElement $transUnit */ |
||
| 34 | foreach ( $file->childNodes as $childNode ) { |
||
| 35 | $this->extractTuFromNode( $childNode, $transUnitIdArrayForUniquenessCheck, $dom, $output, $i, $j ); |
||
| 36 | } |
||
| 37 | |||
| 38 | // trans-unit re-count check |
||
| 39 | $totalTransUnitsId = count( $transUnitIdArrayForUniquenessCheck ); |
||
| 40 | $transUnitsUniqueId = count( array_unique( $transUnitIdArrayForUniquenessCheck ) ); |
||
| 41 | if ( $totalTransUnitsId != $transUnitsUniqueId ) { |
||
| 42 | throw new DuplicateTransUnitIdInXliff( "Invalid trans-unit id, duplicate found.", 400 ); |
||
| 43 | } |
||
| 44 | |||
| 45 | $i++; |
||
| 46 | } |
||
| 47 | |||
| 48 | return $output; |
||
| 49 | } |
||
| 50 | |||
| 51 | /** |
||
| 52 | * @param DOMDocument $dom |
||
| 53 | * |
||
| 54 | * @return array |
||
| 55 | */ |
||
| 56 | private function extractMetadata( DOMDocument $dom ): array { |
||
| 57 | $metadata = []; |
||
| 58 | |||
| 59 | $xliffNode = $dom->getElementsByTagName( 'xliff' )->item( 0 ); |
||
| 60 | $fileNode = $dom->getElementsByTagName( 'file' )->item( 0 ); |
||
| 61 | |||
| 62 | // original |
||
| 63 | $metadata[ 'original' ] = ( null !== $fileNode->attributes->getNamedItem( 'original' ) ) ? $fileNode->attributes->getNamedItem( 'original' )->nodeValue : 'no-name'; |
||
|
|
|||
| 64 | |||
| 65 | // source-language |
||
| 66 | $metadata[ 'source-language' ] = ( null !== $xliffNode->attributes->getNamedItem( 'srcLang' ) ) ? $xliffNode->attributes->getNamedItem( 'srcLang' )->nodeValue : 'en-US'; |
||
| 67 | |||
| 68 | // datatype |
||
| 69 | // @TODO to be implemented |
||
| 70 | |||
| 71 | // target-language |
||
| 72 | $metadata[ 'target-language' ] = ( null !== $xliffNode->attributes->getNamedItem( 'trgLang' ) ) ? $xliffNode->attributes->getNamedItem( 'trgLang' )->nodeValue : 'en-US'; |
||
| 73 | |||
| 74 | // custom MateCat x-attribute |
||
| 75 | // @TODO to be implemented |
||
| 76 | |||
| 77 | return $metadata; |
||
| 78 | } |
||
| 79 | |||
| 80 | /** |
||
| 81 | * @param DOMElement $file |
||
| 82 | * |
||
| 83 | * @return array |
||
| 84 | * @throws Exception |
||
| 85 | */ |
||
| 86 | private function extractNotes( DOMElement $file ): array { |
||
| 102 | } |
||
| 103 | |||
| 104 | /** |
||
| 105 | * Extract and populate 'trans-units' array |
||
| 106 | * |
||
| 107 | * @param DOMElement $transUnit |
||
| 108 | * @param array $transUnitIdArrayForUniquenessCheck |
||
| 109 | * @param DOMDocument $dom |
||
| 110 | * @param array $output |
||
| 111 | * @param int $i |
||
| 112 | * @param int $j |
||
| 113 | * @param array|null $contextGroups |
||
| 114 | * |
||
| 115 | * @throws Exception |
||
| 116 | */ |
||
| 117 | protected function extractTransUnit( DOMElement $transUnit, array &$transUnitIdArrayForUniquenessCheck, DomDocument $dom, array &$output, int &$i, int &$j, ?array $contextGroups = [] ) { |
||
| 118 | // metadata |
||
| 119 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'attr' ] = $this->extractTransUnitMetadata( $transUnit, $transUnitIdArrayForUniquenessCheck ); |
||
| 120 | |||
| 121 | // notes |
||
| 122 | // merge <notes> with key and key-note contained in metadata <mda:metaGroup> |
||
| 123 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'notes' ] = $this->extractTransUnitNotes( $transUnit ); |
||
| 124 | |||
| 125 | // uuid |
||
| 126 | foreach ( $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'notes' ] as $note ) { |
||
| 127 | if ( isset( $note[ 'raw-content' ] ) && Strings::isAValidUuid( $note[ 'raw-content' ] ) ) { |
||
| 128 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'attr' ][ 'uuid' ] = $note[ 'raw-content' ]; |
||
| 129 | } |
||
| 130 | } |
||
| 131 | |||
| 132 | // original-data (exclusive for V2) |
||
| 133 | // http://docs.oasis-open.org/xliff/xliff-core/v2.0/xliff-core-v2.0.html#originaldata |
||
| 134 | $originalData = $this->extractTransUnitOriginalData( $transUnit ); |
||
| 135 | if ( !empty( $originalData ) ) { |
||
| 136 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'original-data' ] = $originalData; |
||
| 137 | } |
||
| 138 | |||
| 139 | // additionalTagData (exclusive for V2) |
||
| 140 | $additionalTagData = $this->extractTransUnitAdditionalTagData( $transUnit ); |
||
| 141 | if ( !empty( $additionalTagData ) ) { |
||
| 142 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'additional-tag-data' ] = $additionalTagData; |
||
| 143 | } |
||
| 144 | |||
| 145 | // content |
||
| 146 | |||
| 147 | $source = [ |
||
| 148 | 'attr' => [], |
||
| 149 | 'raw-content' => [], |
||
| 150 | ]; |
||
| 151 | |||
| 152 | $target = [ |
||
| 153 | 'attr' => [], |
||
| 154 | 'raw-content' => [], |
||
| 155 | ]; |
||
| 156 | |||
| 157 | $segSource = []; |
||
| 158 | $segTarget = []; |
||
| 159 | |||
| 160 | /** @var DOMElement $segment */ |
||
| 161 | $c = 0; |
||
| 162 | foreach ( $transUnit->childNodes as $segment ) { |
||
| 163 | if ( $segment->nodeName === 'segment' ) { |
||
| 164 | |||
| 165 | // check segment id consistency |
||
| 166 | $attr = $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'attr' ]; |
||
| 167 | $this->checkSegmentIdConsistency( $segment, $attr ); |
||
| 168 | |||
| 169 | // loop <segment> to get nested <source> and <target> tag |
||
| 170 | foreach ( $segment->childNodes as $childNode ) { |
||
| 171 | if ( $childNode->nodeName === 'source' ) { |
||
| 172 | $extractedSource = $this->extractContent( $dom, $childNode ); |
||
| 173 | $source[ 'raw-content' ][ $c ] = $extractedSource[ 'raw-content' ]; |
||
| 174 | |||
| 175 | if ( !empty( $extractedSource[ 'attr' ] ) ) { |
||
| 176 | $source[ 'attr' ][ $c ] = $extractedSource[ 'attr' ]; |
||
| 177 | } |
||
| 178 | |||
| 179 | // append value to 'seg-source' |
||
| 180 | if ( $this->stringContainsMarks( $extractedSource[ 'raw-content' ] ) ) { |
||
| 181 | $segSource = $this->extractContentWithMarksAndExtTags( $dom, $childNode ); |
||
| 182 | } else { |
||
| 183 | $segSource[] = [ |
||
| 184 | 'attr' => $this->extractTagAttributes( $segment ), |
||
| 185 | 'mid' => count( $segSource ) > 0 ? count( $segSource ) : 0, |
||
| 186 | 'ext-prec-tags' => '', |
||
| 187 | 'raw-content' => $extractedSource[ 'raw-content' ], |
||
| 188 | 'ext-succ-tags' => '', |
||
| 189 | ]; |
||
| 190 | } |
||
| 191 | } |
||
| 192 | |||
| 193 | if ( $childNode->nodeName === 'target' ) { |
||
| 194 | $extractedTarget = $this->extractContent( $dom, $childNode ); |
||
| 195 | $target[ 'raw-content' ][ $c ] = $extractedTarget[ 'raw-content' ]; |
||
| 196 | |||
| 197 | if ( !empty( $extractedTarget[ 'attr' ] ) ) { |
||
| 198 | $target[ 'attr' ][ $c ] = $extractedTarget[ 'attr' ]; |
||
| 199 | } |
||
| 200 | |||
| 201 | // append value to 'seg-target' |
||
| 202 | if ( $this->stringContainsMarks( $extractedTarget[ 'raw-content' ] ) ) { |
||
| 203 | $segTarget = $this->extractContentWithMarksAndExtTags( $dom, $childNode ); |
||
| 204 | } else { |
||
| 205 | $segTarget[] = [ |
||
| 206 | 'attr' => $this->extractTagAttributes( $segment ), |
||
| 207 | 'mid' => count( $segTarget ) > 0 ? count( $segTarget ) : 0, |
||
| 208 | 'ext-prec-tags' => '', |
||
| 209 | 'raw-content' => $extractedTarget[ 'raw-content' ], |
||
| 210 | 'ext-succ-tags' => '', |
||
| 211 | ]; |
||
| 212 | } |
||
| 213 | } |
||
| 214 | } |
||
| 215 | |||
| 216 | $c++; |
||
| 217 | } |
||
| 218 | } |
||
| 219 | |||
| 220 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'source' ] = $source; |
||
| 221 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'target' ] = $target; |
||
| 222 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'seg-source' ] = $segSource; |
||
| 223 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'seg-target' ] = $segTarget; |
||
| 224 | |||
| 225 | $j++; |
||
| 226 | } |
||
| 227 | |||
| 228 | /** |
||
| 229 | * @param DOMElement $transUnit |
||
| 230 | * @param array|null $transUnitIdArrayForUniquenessCheck |
||
| 231 | * |
||
| 232 | * @return array |
||
| 233 | */ |
||
| 234 | private function extractTransUnitMetadata( DOMElement $transUnit, ?array &$transUnitIdArrayForUniquenessCheck = [] ): array { |
||
| 235 | $metadata = []; |
||
| 236 | |||
| 237 | // id |
||
| 238 | if ( null === $transUnit->attributes->getNamedItem( 'id' ) ) { |
||
| 239 | throw new NotFoundIdInTransUnit( 'Invalid trans-unit id found. EMPTY value', 400 ); |
||
| 240 | } |
||
| 241 | |||
| 242 | $id = $transUnit->attributes->getNamedItem( 'id' )->nodeValue; |
||
| 243 | |||
| 244 | if ( strlen( $id ) > 100 ) { |
||
| 245 | throw new SegmentIdTooLongException( 'Segment-id too long. Max 100 characters allowed', 400 ); |
||
| 246 | } |
||
| 247 | |||
| 248 | $transUnitIdArrayForUniquenessCheck[] = $id; |
||
| 249 | $metadata[ 'id' ] = $id; |
||
| 250 | |||
| 251 | // translate |
||
| 252 | if ( null !== $transUnit->attributes->getNamedItem( 'translate' ) ) { |
||
| 253 | $metadata[ 'translate' ] = $transUnit->attributes->getNamedItem( 'translate' )->nodeValue; |
||
| 254 | } |
||
| 255 | |||
| 256 | // tGroupBegin |
||
| 257 | if ( null !== $transUnit->attributes->getNamedItem( 'tGroupBegin' ) ) { |
||
| 258 | $metadata[ 'tGroupBegin' ] = $transUnit->attributes->getNamedItem( 'tGroupBegin' )->nodeValue; |
||
| 259 | } |
||
| 260 | |||
| 261 | // tGroupEnd |
||
| 262 | if ( null !== $transUnit->attributes->getNamedItem( 'tGroupEnd' ) ) { |
||
| 263 | $metadata[ 'tGroupEnd' ] = $transUnit->attributes->getNamedItem( 'tGroupEnd' )->nodeValue; |
||
| 264 | } |
||
| 265 | |||
| 266 | // sizeRestriction |
||
| 267 | if ( null !== $transUnit->attributes->getNamedItem( 'sizeRestriction' ) && '' !== $transUnit->attributes->getNamedItem( 'sizeRestriction' )->nodeValue ) { |
||
| 268 | $metadata[ 'sizeRestriction' ] = (int)$transUnit->attributes->getNamedItem( 'sizeRestriction' )->nodeValue; |
||
| 269 | } |
||
| 270 | |||
| 271 | return $metadata; |
||
| 272 | } |
||
| 273 | |||
| 274 | /** |
||
| 275 | * @param DOMElement $transUnit |
||
| 276 | * |
||
| 277 | * @return array |
||
| 278 | * @throws Exception |
||
| 279 | */ |
||
| 280 | private function extractTransUnitOriginalData( DOMElement $transUnit ): array { |
||
| 281 | $originalData = []; |
||
| 282 | |||
| 283 | // loop <originalData> to get nested content |
||
| 284 | foreach ( $transUnit->childNodes as $childNode ) { |
||
| 285 | if ( $childNode->nodeName === 'originalData' ) { |
||
| 286 | foreach ( $childNode->childNodes as $data ) { |
||
| 287 | if ( null !== $data->attributes && null !== $data->attributes->getNamedItem( 'id' ) ) { |
||
| 288 | $dataId = $data->attributes->getNamedItem( 'id' )->nodeValue; |
||
| 289 | |||
| 290 | $dataValue = str_replace( Placeholder::WHITE_SPACE_PLACEHOLDER, ' ', $data->nodeValue ); |
||
| 291 | $dataValue = str_replace( Placeholder::NEW_LINE_PLACEHOLDER, '\n', $dataValue ); |
||
| 292 | $dataValue = str_replace( Placeholder::TAB_PLACEHOLDER, '\t', $dataValue ); |
||
| 293 | |||
| 294 | if ( '' !== $dataValue ) { |
||
| 295 | |||
| 296 | $jsonOrRawContentArray = $this->JSONOrRawContentArray( $dataValue, false ); |
||
| 297 | |||
| 298 | // restore xliff tags |
||
| 299 | if ( isset( $jsonOrRawContentArray[ 'json' ] ) ) { |
||
| 300 | $jsonOrRawContentArray[ 'json' ] = str_replace( [ Placeholder::LT_PLACEHOLDER, Placeholder::GT_PLACEHOLDER ], [ '<', '>' ], $jsonOrRawContentArray[ 'json' ] ); |
||
| 301 | } |
||
| 302 | |||
| 303 | if ( isset( $jsonOrRawContentArray[ 'raw-content' ] ) ) { |
||
| 304 | $jsonOrRawContentArray[ 'raw-content' ] = str_replace( [ Placeholder::LT_PLACEHOLDER, Placeholder::GT_PLACEHOLDER ], [ '<', '>' ], $jsonOrRawContentArray[ 'raw-content' ] ); |
||
| 305 | } |
||
| 306 | |||
| 307 | $originalData[] = array_merge( |
||
| 308 | $jsonOrRawContentArray, |
||
| 309 | [ |
||
| 310 | 'attr' => [ |
||
| 311 | 'id' => $dataId |
||
| 312 | ] |
||
| 313 | ] |
||
| 314 | ); |
||
| 315 | } |
||
| 316 | } |
||
| 317 | } |
||
| 318 | } |
||
| 319 | } |
||
| 320 | |||
| 321 | return $originalData; |
||
| 322 | } |
||
| 323 | |||
| 324 | /** |
||
| 325 | * @param DOMElement $transUnit |
||
| 326 | * |
||
| 327 | * @return array |
||
| 328 | */ |
||
| 329 | private function extractTransUnitAdditionalTagData( DOMElement $transUnit ): array { |
||
| 330 | $additionalTagData = []; |
||
| 331 | |||
| 332 | // loop <originalData> to get nested content |
||
| 333 | foreach ( $transUnit->childNodes as $childNode ) { |
||
| 334 | if ( $childNode->nodeName === 'memsource:additionalTagData' ) { |
||
| 335 | foreach ( $childNode->childNodes as $data ) { |
||
| 336 | $dataArray = []; |
||
| 337 | |||
| 338 | // id |
||
| 339 | if ( $data->nodeName === 'memsource:tag' ) { |
||
| 340 | if ( null !== $data->attributes && null !== $data->attributes->getNamedItem( 'id' ) ) { |
||
| 341 | $dataId = $data->attributes->getNamedItem( 'id' )->nodeValue; |
||
| 342 | $dataArray[ 'attr' ][ 'id' ] = $dataId; |
||
| 343 | } |
||
| 344 | } |
||
| 345 | |||
| 346 | // in PHP 7.4 $data->childNodes is an empty DomNodeList, it is iterable with size 0 |
||
| 347 | // PHP 5.6 check: in php 5.6 $data->childNodes can be null |
||
| 348 | if ( $data->childNodes != null ) { |
||
| 349 | |||
| 350 | // content |
||
| 351 | foreach ( $data->childNodes as $datum ) { |
||
| 352 | if ( $datum->nodeName === 'memsource:tagId' ) { |
||
| 353 | $dataArray[ 'raw-content' ][ 'tagId' ] = $datum->nodeValue; |
||
| 354 | } |
||
| 355 | |||
| 356 | if ( $datum->nodeName === 'memsource:type' ) { |
||
| 357 | $dataArray[ 'raw-content' ][ 'type' ] = $datum->nodeValue; |
||
| 358 | } |
||
| 359 | } |
||
| 360 | |||
| 361 | } |
||
| 362 | |||
| 363 | if ( !empty( $dataArray ) ) { |
||
| 364 | $additionalTagData[] = $dataArray; |
||
| 365 | } |
||
| 366 | } |
||
| 367 | } |
||
| 368 | } |
||
| 369 | |||
| 370 | return $additionalTagData; |
||
| 371 | } |
||
| 372 | |||
| 373 | /** |
||
| 374 | * Check if segment id is present within tGroupBegin and tGroupEnd attributes |
||
| 375 | * |
||
| 376 | * @param DOMElement $segment |
||
| 377 | * @param array $attr |
||
| 378 | */ |
||
| 379 | private function checkSegmentIdConsistency( DOMElement $segment, array $attr ) { |
||
| 380 | if ( isset( $attr[ 'tGroupBegin' ] ) && isset( $attr[ 'tGroupEnd' ] ) && $segment->attributes->getNamedItem( 'id' ) ) { |
||
| 381 | $id = $segment->attributes->getNamedItem( 'id' )->nodeValue; |
||
| 382 | $min = (int)$attr[ 'tGroupBegin' ]; |
||
| 383 | $max = (int)$attr[ 'tGroupEnd' ]; |
||
| 384 | |||
| 385 | if ( false === ( ( $min <= $id ) && ( $id <= $max ) ) ) { |
||
| 386 | if ( $this->logger ) { |
||
| 387 | $this->logger->warning( 'Segment #' . $id . ' is not included within tGroupBegin and tGroupEnd' ); |
||
| 388 | } |
||
| 389 | } |
||
| 390 | } |
||
| 391 | } |
||
| 392 | |||
| 393 | /** |
||
| 394 | * @param DOMElement $transUnit |
||
| 395 | * |
||
| 396 | * @return array |
||
| 397 | * @throws Exception |
||
| 398 | */ |
||
| 399 | private function extractTransUnitNotes( DOMElement $transUnit ): array { |
||
| 439 | } |
||
| 440 | } |
||
| 441 |
Loading history...
This check looks for calls to methods that do not seem to exist on a given type. It looks for the method on the type itself as well as in inherited classes or implemented interfaces.
This is most likely a typographical error or the method has been renamed.