matecat /
xliff-parser
| Total Complexity | 63 |
| Total Lines | 357 |
| Duplicated Lines | 0 % |
| Changes | 3 | ||
| Bugs | 1 | Features | 0 |
Complex classes like XliffParserV1 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 XliffParserV1, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 14 | class XliffParserV1 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( $file ); |
||
| 26 | |||
| 27 | // reference |
||
| 28 | if ( !empty( $this->extractReference( $file ) ) ) { |
||
| 29 | $output[ 'files' ][ $i ][ 'reference' ] = $this->extractReference( $file ); |
||
| 30 | } |
||
| 31 | |||
| 32 | // trans-units |
||
| 33 | $transUnitIdArrayForUniquenessCheck = []; |
||
| 34 | $j = 1; |
||
| 35 | foreach ( $file->childNodes as $body ) { |
||
| 36 | |||
| 37 | // external-file |
||
| 38 | if ( $body->nodeName === 'header' ) { |
||
| 39 | foreach ( $body->childNodes as $header ) { |
||
| 40 | $this->extractExternalFile( $header, $i, $output ); |
||
| 41 | } |
||
| 42 | } |
||
| 43 | |||
| 44 | if ( $body->nodeName === 'body' ) { |
||
| 45 | foreach ( $body->childNodes as $childNode ) { |
||
| 46 | $this->extractTuFromNode( $childNode, $transUnitIdArrayForUniquenessCheck, $dom, $output, $i, $j ); |
||
| 47 | } |
||
| 48 | |||
| 49 | // trans-unit re-count check |
||
| 50 | $totalTransUnitsId = count( $transUnitIdArrayForUniquenessCheck ); |
||
| 51 | $transUnitsUniqueId = count( array_unique( $transUnitIdArrayForUniquenessCheck ) ); |
||
| 52 | if ( $totalTransUnitsId != $transUnitsUniqueId ) { |
||
| 53 | throw new DuplicateTransUnitIdInXliff( "Invalid trans-unit id, duplicate found.", 400 ); |
||
| 54 | } |
||
| 55 | |||
| 56 | $i++; |
||
| 57 | } |
||
| 58 | } |
||
| 59 | } |
||
| 60 | |||
| 61 | return $output; |
||
| 62 | } |
||
| 63 | |||
| 64 | /** |
||
| 65 | * @param DOMNode $header |
||
| 66 | * @param $i |
||
| 67 | * @param $output |
||
| 68 | */ |
||
| 69 | private function extractExternalFile( DOMNode $header, $i, &$output ) { |
||
| 70 | |||
| 71 | if ( $header->nodeName === "skl" ) { |
||
| 72 | foreach ( $header->childNodes as $referenceNode ) { |
||
| 73 | if ( $referenceNode->nodeName === "reference" ) { |
||
| 74 | foreach ( $referenceNode->childNodes as $childNode ) { |
||
| 75 | if ( $childNode->nodeName === "external-file" ) { |
||
| 76 | $href = $childNode->getAttribute( "href" ); |
||
| 77 | $output[ 'files' ][ $i ][ 'attr' ][ 'external-file' ] = $href; |
||
| 78 | } |
||
| 79 | } |
||
| 80 | } elseif ( $referenceNode->nodeName === "external-file" ) { |
||
| 81 | $href = $referenceNode->getAttribute( "href" ); |
||
| 82 | $output[ 'files' ][ $i ][ 'attr' ][ 'external-file' ] = $href; |
||
| 83 | } |
||
| 84 | } |
||
| 85 | } elseif ( $header->nodeName === "reference" ) { |
||
| 86 | foreach ( $header->childNodes as $referenceNode ) { |
||
| 87 | if ( $referenceNode->nodeName === "external-file" ) { |
||
| 88 | $href = $referenceNode->getAttribute( "href" ); |
||
| 89 | $output[ 'files' ][ $i ][ 'attr' ][ 'external-file' ] = $href; |
||
| 90 | } |
||
| 91 | } |
||
| 92 | } |
||
| 93 | } |
||
| 94 | |||
| 95 | /** |
||
| 96 | * @param DOMElement $file |
||
| 97 | * |
||
| 98 | * @return array |
||
| 99 | */ |
||
| 100 | private function extractMetadata( DOMElement $file ): array { |
||
| 101 | $metadata = []; |
||
| 102 | $customAttr = []; |
||
| 103 | |||
| 104 | /** @var DOMAttr $attribute */ |
||
| 105 | foreach ( $file->attributes as $attribute ) { |
||
| 106 | switch ( $attribute->localName ) { |
||
| 107 | // original |
||
| 108 | case 'original': |
||
| 109 | $metadata[ 'original' ] = $attribute->value; |
||
| 110 | break; |
||
| 111 | |||
| 112 | // source-language |
||
| 113 | case 'source-language': |
||
| 114 | $metadata[ 'source-language' ] = $attribute->value; |
||
| 115 | break; |
||
| 116 | |||
| 117 | // data-type |
||
| 118 | case 'datatype': |
||
| 119 | $metadata[ 'data-type' ] = $attribute->value; |
||
| 120 | break; |
||
| 121 | |||
| 122 | // target-language |
||
| 123 | case 'target-language': |
||
| 124 | $metadata[ 'target-language' ] = $attribute->value; |
||
| 125 | break; |
||
| 126 | } |
||
| 127 | |||
| 128 | // Custom MateCat x-Attribute |
||
| 129 | preg_match( '|x-(.*?)|si', $attribute->localName, $temp ); |
||
| 130 | if ( isset( $temp[ 1 ] ) ) { |
||
| 131 | $customAttr[ $attribute->localName ] = $attribute->value; |
||
| 132 | } |
||
| 133 | unset( $temp ); |
||
| 134 | |||
| 135 | // Custom MateCat namespace Attribute mtc: |
||
| 136 | preg_match( '|mtc:(.*?)|si', $attribute->nodeName, $temp ); |
||
| 137 | if ( isset( $temp[ 1 ] ) ) { |
||
| 138 | $customAttr[ $attribute->nodeName ] = $attribute->value; |
||
| 139 | } |
||
| 140 | unset( $temp ); |
||
| 141 | |||
| 142 | if ( !empty( $customAttr ) ) { |
||
| 143 | $metadata[ 'custom' ] = $customAttr; |
||
| 144 | } |
||
| 145 | } |
||
| 146 | |||
| 147 | return $metadata; |
||
| 148 | } |
||
| 149 | |||
| 150 | /** |
||
| 151 | * @param DOMElement $file |
||
| 152 | * |
||
| 153 | * @return array |
||
| 154 | */ |
||
| 155 | private function extractReference( DOMElement $file ): array { |
||
| 156 | $reference = []; |
||
| 157 | |||
| 158 | $order = 0; |
||
| 159 | foreach ( $file->getElementsByTagName( 'reference' ) as $ref ) { |
||
| 160 | /** @var DOMNode $childNode */ |
||
| 161 | foreach ( $ref->childNodes as $childNode ) { |
||
| 162 | if ( $childNode->nodeName === 'internal-file' ) { |
||
| 163 | $reference[ $order ][ 'form-type' ] = $childNode->attributes->getNamedItem( 'form' )->nodeValue; |
||
| 164 | $reference[ $order ][ 'base64' ] = trim( $childNode->nodeValue ); |
||
| 165 | $order++; |
||
| 166 | } |
||
| 167 | } |
||
| 168 | } |
||
| 169 | |||
| 170 | return $reference; |
||
| 171 | } |
||
| 172 | |||
| 173 | /** |
||
| 174 | * Extract and populate 'trans-units' array |
||
| 175 | * |
||
| 176 | * @param DOMElement $transUnit |
||
| 177 | * @param array $transUnitIdArrayForUniquenessCheck |
||
| 178 | * @param DOMDocument $dom |
||
| 179 | * @param array $output |
||
| 180 | * @param int $i |
||
| 181 | * @param int $j |
||
| 182 | * @param array|null $contextGroups |
||
| 183 | * |
||
| 184 | * @throws Exception |
||
| 185 | */ |
||
| 186 | protected function extractTransUnit( DOMElement $transUnit, array &$transUnitIdArrayForUniquenessCheck, DomDocument $dom, array &$output, int &$i, int &$j, ?array $contextGroups = [] ) { |
||
| 187 | // metadata |
||
| 188 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'attr' ] = $this->extractTransUnitMetadata( $transUnit, $transUnitIdArrayForUniquenessCheck ); |
||
| 189 | |||
| 190 | // notes |
||
| 191 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'notes' ] = $this->extractTransUnitNotes( $dom, $transUnit ); |
||
| 192 | |||
| 193 | // content |
||
| 194 | /** @var DOMElement $childNode */ |
||
| 195 | foreach ( $transUnit->childNodes as $childNode ) { |
||
| 196 | // source |
||
| 197 | if ( $childNode->nodeName === 'source' ) { |
||
| 198 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'source' ] = $this->extractContent( $dom, $childNode ); |
||
| 199 | } |
||
| 200 | |||
| 201 | // seg-source |
||
| 202 | if ( $childNode->nodeName === 'seg-source' ) { |
||
| 203 | $rawSegment = $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'source' ][ 'raw-content' ]; |
||
| 204 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'seg-source' ] = $this->extractContentWithMarksAndExtTags( $dom, $childNode, $rawSegment ); |
||
|
|
|||
| 205 | } |
||
| 206 | |||
| 207 | // target |
||
| 208 | if ( $childNode->nodeName === 'target' ) { |
||
| 209 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'target' ] = $this->extractContent( $dom, $childNode ); |
||
| 210 | |||
| 211 | // seg-target |
||
| 212 | $targetRawContent = $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'target' ][ 'raw-content' ] ?? null; |
||
| 213 | $segSource = $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'seg-source' ] ?? null; |
||
| 214 | |||
| 215 | if ( !empty( $targetRawContent ) and isset( $segSource ) and count( $segSource ) > 0 ) { |
||
| 216 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'seg-target' ] = $this->extractContentWithMarksAndExtTags( $dom, $childNode ); |
||
| 217 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'seg-target' ][ 0 ][ 'attr' ] = $this->extractTagAttributes( $childNode ); |
||
| 218 | } |
||
| 219 | } |
||
| 220 | |||
| 221 | // locked |
||
| 222 | if ( $childNode->nodeName === 'sdl:seg' ) { |
||
| 223 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'locked' ] = $this->extractLocked( $childNode ); |
||
| 224 | } |
||
| 225 | } |
||
| 226 | |||
| 227 | // context-group |
||
| 228 | if ( !empty( $contextGroups ) ) { |
||
| 229 | foreach ( $contextGroups as $contextGroup ) { |
||
| 230 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'context-group' ][] = $this->extractTransUnitContextGroup( $dom, $contextGroup ); |
||
| 231 | } |
||
| 232 | } |
||
| 233 | |||
| 234 | foreach ( $transUnit->getElementsByTagName( 'context-group' ) as $contextGroup ) { |
||
| 235 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'context-group' ][] = $this->extractTransUnitContextGroup( $dom, $contextGroup ); |
||
| 236 | } |
||
| 237 | |||
| 238 | // alt-trans |
||
| 239 | foreach ( $transUnit->getElementsByTagName( 'alt-trans' ) as $altTrans ) { |
||
| 240 | $output[ 'files' ][ $i ][ 'trans-units' ][ $j ][ 'alt-trans' ][] = $this->extractTransUnitAltTrans( $altTrans ); |
||
| 241 | } |
||
| 242 | |||
| 243 | $j++; |
||
| 244 | } |
||
| 245 | |||
| 246 | /** |
||
| 247 | * @param DOMElement $transUnit |
||
| 248 | * @param array $transUnitIdArrayForUniquenessCheck |
||
| 249 | * |
||
| 250 | * @return array |
||
| 251 | * @throws Exception |
||
| 252 | */ |
||
| 253 | private function extractTransUnitMetadata( DOMElement $transUnit, array &$transUnitIdArrayForUniquenessCheck ): array { |
||
| 293 | } |
||
| 294 | |||
| 295 | /** |
||
| 296 | * @param DOMDocument $dom |
||
| 297 | * @param DOMElement $transUnit |
||
| 298 | * |
||
| 299 | * @return array |
||
| 300 | * @throws Exception |
||
| 301 | */ |
||
| 302 | private function extractTransUnitNotes( DOMDocument $dom, DOMElement $transUnit ): array { |
||
| 303 | $notes = []; |
||
| 304 | foreach ( $transUnit->getElementsByTagName( 'note' ) as $note ) { |
||
| 305 | |||
| 306 | $noteValue = $this->extractTagContent( $dom, $note ); |
||
| 307 | |||
| 308 | if ( '' !== $noteValue ) { |
||
| 309 | |||
| 310 | $extractedNote = $this->JSONOrRawContentArray( $noteValue ); |
||
| 311 | |||
| 312 | // extract all the attributes |
||
| 313 | foreach ( $note->attributes as $attribute ) { |
||
| 314 | $extractedNote[ $attribute->name ] = $attribute->value; |
||
| 315 | } |
||
| 316 | |||
| 317 | $notes[] = $extractedNote; |
||
| 318 | } |
||
| 319 | } |
||
| 320 | |||
| 321 | return $notes; |
||
| 322 | } |
||
| 323 | |||
| 324 | /** |
||
| 325 | * @param DOMDocument $dom |
||
| 326 | * @param DOMElement $contextGroup |
||
| 327 | * |
||
| 328 | * @return array |
||
| 329 | */ |
||
| 330 | private function extractTransUnitContextGroup( DOMDocument $dom, DOMElement $contextGroup ): array { |
||
| 331 | $cg = []; |
||
| 332 | $cg[ 'attr' ] = $this->extractTagAttributes( $contextGroup ); |
||
| 333 | |||
| 334 | /** @var DOMNode $context */ |
||
| 335 | foreach ( $contextGroup->childNodes as $context ) { |
||
| 336 | if ( $context->nodeName === 'context' ) { |
||
| 337 | $cg[ 'contexts' ][] = $this->extractContent( $dom, $context ); |
||
| 338 | } |
||
| 339 | } |
||
| 340 | |||
| 341 | return $cg; |
||
| 342 | } |
||
| 343 | |||
| 344 | /** |
||
| 345 | * @param DOMElement $altTrans |
||
| 346 | * |
||
| 347 | * @return array |
||
| 348 | */ |
||
| 349 | private function extractTransUnitAltTrans( DOMElement $altTrans ) { |
||
| 362 | } |
||
| 363 | |||
| 364 | /** |
||
| 365 | * @param DOMElement $locked |
||
| 366 | * |
||
| 367 | * @return bool |
||
| 368 | */ |
||
| 369 | private function extractLocked( DOMElement $locked ) { |
||
| 371 | } |
||
| 372 | } |
||
| 373 |
Loading history...
This check compares calls to functions or methods with their respective definitions. If the call has more arguments than are defined, it raises an issue.
If a function is defined several times with a different number of parameters, the check may pick up the wrong definition and report false positives. One codebase where this has been known to happen is Wordpress. Please note the @ignore annotation hint above.