Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
76.35% |
184 / 241 |
|
52.00% |
13 / 25 |
CRAP | |
0.00% |
0 / 1 |
| HtmlOutputRendererHelper | |
76.35% |
184 / 241 |
|
52.00% |
13 / 25 |
164.67 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
1 | |||
| setFlavor | |
80.00% |
4 / 5 |
|
0.00% |
0 / 1 |
3.07 | |||
| getFlavor | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| setOutputProfileVersion | |
0.00% |
0 / 9 |
|
0.00% |
0 / 1 |
12 | |||
| setUseParserCache | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
3 | |||
| setStashingEnabled | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
3 | |||
| setRevision | |
83.33% |
5 / 6 |
|
0.00% |
0 / 1 |
4.07 | |||
| setContent | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
1 | |||
| setContentSource | |
60.00% |
3 / 5 |
|
0.00% |
0 / 1 |
2.26 | |||
| setPageLanguage | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
2 | |||
| init | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
3 | |||
| setVariantConversionLanguage | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
3 | |||
| getAcceptedTargetLanguage | |
66.67% |
2 / 3 |
|
0.00% |
0 / 1 |
2.15 | |||
| getHtml | |
73.81% |
31 / 42 |
|
0.00% |
0 / 1 |
7.88 | |||
| getETag | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
3 | |||
| getLastModified | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getParamSettings | |
0.00% |
0 / 14 |
|
0.00% |
0 / 1 |
2 | |||
| getDefaultPageLanguage | |
75.00% |
3 / 4 |
|
0.00% |
0 / 1 |
2.06 | |||
| getParserOutput | |
92.86% |
26 / 28 |
|
0.00% |
0 / 1 |
10.04 | |||
| getHtmlOutputContentLanguage | |
23.08% |
3 / 13 |
|
0.00% |
0 / 1 |
7.10 | |||
| putHeaders | |
80.00% |
8 / 10 |
|
0.00% |
0 / 1 |
5.20 | |||
| getPageBundle | |
100.00% |
10 / 10 |
|
100.00% |
1 / 1 |
2 | |||
| getRevisionId | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
4 | |||
| stripParsoidSectionTags | |
70.00% |
7 / 10 |
|
0.00% |
0 / 1 |
4.43 | |||
| getParserOutputInternal | |
100.00% |
25 / 25 |
|
100.00% |
1 / 1 |
7 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * This program is free software; you can redistribute it and/or modify |
| 4 | * it under the terms of the GNU General Public License as published by |
| 5 | * the Free Software Foundation; either version 2 of the License, or |
| 6 | * (at your option) any later version. |
| 7 | * |
| 8 | * This program is distributed in the hope that it will be useful, |
| 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 11 | * GNU General Public License for more details. |
| 12 | * |
| 13 | * You should have received a copy of the GNU General Public License along |
| 14 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 15 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 16 | * http://www.gnu.org/copyleft/gpl.html |
| 17 | * |
| 18 | * @file |
| 19 | */ |
| 20 | namespace MediaWiki\Rest\Handler\Helper; |
| 21 | |
| 22 | use Content; |
| 23 | use HttpError; |
| 24 | use IBufferingStatsdDataFactory; |
| 25 | use InvalidArgumentException; |
| 26 | use LanguageCode; |
| 27 | use Liuggio\StatsdClient\Factory\StatsdDataFactoryInterface; |
| 28 | use MediaWiki\Content\IContentHandlerFactory; |
| 29 | use MediaWiki\Edit\ParsoidOutputStash; |
| 30 | use MediaWiki\Edit\SelserContext; |
| 31 | use MediaWiki\Languages\LanguageFactory; |
| 32 | use MediaWiki\Logger\LoggerFactory; |
| 33 | use MediaWiki\MainConfigNames; |
| 34 | use MediaWiki\Page\PageIdentity; |
| 35 | use MediaWiki\Page\ParserOutputAccess; |
| 36 | use MediaWiki\Parser\ParserOutput; |
| 37 | use MediaWiki\Parser\Parsoid\HtmlTransformFactory; |
| 38 | use MediaWiki\Parser\Parsoid\PageBundleParserOutputConverter; |
| 39 | use MediaWiki\Parser\Parsoid\ParsoidOutputAccess; |
| 40 | use MediaWiki\Parser\Parsoid\ParsoidRenderID; |
| 41 | use MediaWiki\Permissions\Authority; |
| 42 | use MediaWiki\Rest\Handler; |
| 43 | use MediaWiki\Rest\HttpException; |
| 44 | use MediaWiki\Rest\LocalizedHttpException; |
| 45 | use MediaWiki\Rest\ResponseInterface; |
| 46 | use MediaWiki\Revision\MutableRevisionRecord; |
| 47 | use MediaWiki\Revision\RevisionAccessException; |
| 48 | use MediaWiki\Revision\RevisionRecord; |
| 49 | use MediaWiki\Revision\SlotRecord; |
| 50 | use MediaWiki\Status\Status; |
| 51 | use MediaWiki\Title\Title; |
| 52 | use MWUnknownContentModelException; |
| 53 | use ParserOptions; |
| 54 | use Wikimedia\Assert\Assert; |
| 55 | use Wikimedia\Bcp47Code\Bcp47Code; |
| 56 | use Wikimedia\Bcp47Code\Bcp47CodeValue; |
| 57 | use Wikimedia\Message\MessageValue; |
| 58 | use Wikimedia\ParamValidator\ParamValidator; |
| 59 | use Wikimedia\Parsoid\Core\PageBundle; |
| 60 | use Wikimedia\Parsoid\DOM\Element; |
| 61 | use Wikimedia\Parsoid\Parsoid; |
| 62 | use Wikimedia\Parsoid\Utils\ContentUtils; |
| 63 | use Wikimedia\Parsoid\Utils\DOMCompat; |
| 64 | use Wikimedia\Parsoid\Utils\DOMUtils; |
| 65 | use Wikimedia\Parsoid\Utils\WTUtils; |
| 66 | |
| 67 | /** |
| 68 | * Helper for getting output of a given wikitext page rendered by parsoid. |
| 69 | * |
| 70 | * @since 1.36 |
| 71 | * |
| 72 | * @unstable Pending consolidation of the Parsoid extension with core code. |
| 73 | */ |
| 74 | class HtmlOutputRendererHelper implements HtmlOutputHelper { |
| 75 | use RestAuthorizeTrait; |
| 76 | use RestStatusTrait; |
| 77 | |
| 78 | /** |
| 79 | * @internal |
| 80 | * @var string[] |
| 81 | */ |
| 82 | public const CONSTRUCTOR_OPTIONS = [ |
| 83 | MainConfigNames::ParsoidCacheConfig |
| 84 | ]; |
| 85 | |
| 86 | /** @var string[] */ |
| 87 | private const OUTPUT_FLAVORS = [ 'view', 'stash', 'fragment', 'edit' ]; |
| 88 | |
| 89 | /** @var ParsoidOutputStash */ |
| 90 | private $parsoidOutputStash; |
| 91 | |
| 92 | /** @var PageIdentity|null */ |
| 93 | private $page = null; |
| 94 | |
| 95 | /** @var RevisionRecord|int|null */ |
| 96 | private $revisionOrId = null; |
| 97 | |
| 98 | /** @var Bcp47Code|null */ |
| 99 | private $pageLanguage = null; |
| 100 | |
| 101 | /** @var ?string One of the flavors from OUTPUT_FLAVORS */ |
| 102 | private $flavor = null; |
| 103 | |
| 104 | /** @var bool */ |
| 105 | private $stash = false; |
| 106 | |
| 107 | /** @var IBufferingStatsdDataFactory */ |
| 108 | private $stats; |
| 109 | |
| 110 | /** @var Authority */ |
| 111 | private $authority; |
| 112 | |
| 113 | /** @var ParsoidOutputAccess */ |
| 114 | private $parsoidOutputAccess; |
| 115 | |
| 116 | /** @var ParserOutput */ |
| 117 | private $parserOutput; |
| 118 | |
| 119 | /** @var ParserOutput */ |
| 120 | private $processedParserOutput; |
| 121 | |
| 122 | /** @var HtmlTransformFactory */ |
| 123 | private $htmlTransformFactory; |
| 124 | |
| 125 | /** @var IContentHandlerFactory */ |
| 126 | private $contentHandlerFactory; |
| 127 | |
| 128 | /** @var LanguageFactory */ |
| 129 | private $languageFactory; |
| 130 | |
| 131 | /** @var ?Bcp47Code */ |
| 132 | private $sourceLanguage = null; |
| 133 | |
| 134 | /** @var ?Bcp47Code */ |
| 135 | private $targetLanguage = null; |
| 136 | |
| 137 | /** Should we ignore mismatched $page and $revisionOrId values? */ |
| 138 | private bool $lenientRevHandling = false; |
| 139 | |
| 140 | /** |
| 141 | * Flags to be passed as $options to ParsoidOutputAccess::getParserOutput, |
| 142 | * to control parser cache access. |
| 143 | * |
| 144 | * @var int Use ParsoidOutputAccess::OPT_* |
| 145 | */ |
| 146 | private $parsoidOutputAccessOptions = 0; |
| 147 | |
| 148 | /** |
| 149 | * @see the $options parameter on Parsoid::wikitext2html |
| 150 | * @var array |
| 151 | */ |
| 152 | private $parsoidOptions = []; |
| 153 | |
| 154 | /** |
| 155 | * Whether the result can be cached in the parser cache and the web cache. |
| 156 | * Set to false when bespoke options are set. |
| 157 | * |
| 158 | * @var bool |
| 159 | */ |
| 160 | private $isCacheable = true; |
| 161 | |
| 162 | /** |
| 163 | * @param ParsoidOutputStash $parsoidOutputStash |
| 164 | * @param StatsdDataFactoryInterface $statsDataFactory |
| 165 | * @param ParsoidOutputAccess $parsoidOutputAccess |
| 166 | * @param HtmlTransformFactory $htmlTransformFactory |
| 167 | * @param IContentHandlerFactory $contentHandlerFactory |
| 168 | * @param LanguageFactory $languageFactory |
| 169 | * @param bool $lenientRevHandling Should we ignore mismatches |
| 170 | * $page and the page that $revision belongs to? Usually happens |
| 171 | * because of page moves. This should be set to true only for |
| 172 | * internal API calls. |
| 173 | */ |
| 174 | public function __construct( |
| 175 | ParsoidOutputStash $parsoidOutputStash, |
| 176 | StatsdDataFactoryInterface $statsDataFactory, |
| 177 | ParsoidOutputAccess $parsoidOutputAccess, |
| 178 | HtmlTransformFactory $htmlTransformFactory, |
| 179 | IContentHandlerFactory $contentHandlerFactory, |
| 180 | LanguageFactory $languageFactory, |
| 181 | bool $lenientRevHandling = false |
| 182 | ) { |
| 183 | $this->parsoidOutputStash = $parsoidOutputStash; |
| 184 | $this->stats = $statsDataFactory; |
| 185 | $this->parsoidOutputAccess = $parsoidOutputAccess; |
| 186 | $this->htmlTransformFactory = $htmlTransformFactory; |
| 187 | $this->contentHandlerFactory = $contentHandlerFactory; |
| 188 | $this->languageFactory = $languageFactory; |
| 189 | $this->lenientRevHandling = $lenientRevHandling; |
| 190 | } |
| 191 | |
| 192 | /** |
| 193 | * Sets the given flavor to use for Wikitext -> HTML transformations. |
| 194 | * |
| 195 | * Flavors may influence parser options, parsoid options, and DOM transformations. |
| 196 | * They will be reflected by the ETag returned by getETag(). |
| 197 | * |
| 198 | * @note This method should not be called if stashing mode is enabled. |
| 199 | * @see setStashingEnabled |
| 200 | * @see getFlavor() |
| 201 | * |
| 202 | * @param string $flavor |
| 203 | * |
| 204 | * @return void |
| 205 | */ |
| 206 | public function setFlavor( string $flavor ): void { |
| 207 | if ( !in_array( $flavor, self::OUTPUT_FLAVORS ) ) { |
| 208 | throw new InvalidArgumentException( 'Invalid flavor supplied' ); |
| 209 | } |
| 210 | |
| 211 | if ( $this->stash ) { |
| 212 | // XXX: throw? |
| 213 | $flavor = 'stash'; |
| 214 | } |
| 215 | |
| 216 | $this->flavor = $flavor; |
| 217 | } |
| 218 | |
| 219 | /** |
| 220 | * Returns the flavor of HTML that will be generated. |
| 221 | * @see setFlavor() |
| 222 | * @return string |
| 223 | */ |
| 224 | public function getFlavor(): string { |
| 225 | return $this->flavor; |
| 226 | } |
| 227 | |
| 228 | /** |
| 229 | * Set the desired Parsoid profile version for the output. |
| 230 | * The actual output version is selected to be compatible with the one given here, |
| 231 | * per the rules of semantic versioning. |
| 232 | * |
| 233 | * @note Will disable caching if the effective output version is different from the default. |
| 234 | * |
| 235 | * @param string $version |
| 236 | * |
| 237 | * @throws HttpException If the given version is not supported (status 406) |
| 238 | */ |
| 239 | public function setOutputProfileVersion( $version ) { |
| 240 | $outputContentVersion = Parsoid::resolveContentVersion( $version ); |
| 241 | |
| 242 | if ( !$outputContentVersion ) { |
| 243 | throw new LocalizedHttpException( |
| 244 | new MessageValue( "rest-unsupported-profile-version", [ $version ] ), 406 |
| 245 | ); |
| 246 | } |
| 247 | |
| 248 | // Only set the option if the value isn't the default! |
| 249 | if ( $outputContentVersion !== Parsoid::defaultHTMLVersion() ) { |
| 250 | throw new LocalizedHttpException( |
| 251 | new MessageValue( "rest-unsupported-profile-version", [ $version ] ), 406 |
| 252 | ); |
| 253 | |
| 254 | // TODO: (T347426) At some later point, we may reintroduce support for |
| 255 | // non-default content versions as part of work on the content |
| 256 | // negotiatiation protocol. |
| 257 | // |
| 258 | // // See Parsoid::wikitext2html |
| 259 | // $this->parsoidOptions['outputContentVersion'] = $outputContentVersion; |
| 260 | // $this->isCacheable = false; |
| 261 | } |
| 262 | } |
| 263 | |
| 264 | /** |
| 265 | * Controls how the parser cache is used. |
| 266 | * |
| 267 | * @param bool $read Whether we should look for cached output before parsing |
| 268 | * @param bool $write Whether we should cache output after parsing |
| 269 | */ |
| 270 | public function setUseParserCache( bool $read, bool $write ) { |
| 271 | $this->parsoidOutputAccessOptions = |
| 272 | ( $read ? 0 : ParserOutputAccess::OPT_FORCE_PARSE ) | |
| 273 | ( $write ? 0 : ParserOutputAccess::OPT_NO_UPDATE_CACHE ); |
| 274 | } |
| 275 | |
| 276 | /** |
| 277 | * Determine whether stashing should be applied. |
| 278 | * |
| 279 | * @param bool $stash |
| 280 | * |
| 281 | * @return void |
| 282 | */ |
| 283 | public function setStashingEnabled( bool $stash ): void { |
| 284 | $this->stash = $stash; |
| 285 | |
| 286 | if ( $stash ) { |
| 287 | $this->setFlavor( 'stash' ); |
| 288 | } elseif ( $this->flavor === 'stash' ) { |
| 289 | $this->setFlavor( 'view' ); |
| 290 | } |
| 291 | } |
| 292 | |
| 293 | /** |
| 294 | * Set the revision to render. |
| 295 | * |
| 296 | * This can take a fake RevisionRecord when rendering for previews |
| 297 | * or when switching the editor from source mode to visual mode. |
| 298 | * |
| 299 | * In that case, $revisionOrId->getId() must return 0 to indicate |
| 300 | * that the ParserCache should be bypassed. Stashing may still apply. |
| 301 | * |
| 302 | * @param RevisionRecord|int $revisionOrId |
| 303 | */ |
| 304 | public function setRevision( $revisionOrId ): void { |
| 305 | Assert::parameterType( [ RevisionRecord::class, 'integer' ], $revisionOrId, '$revision' ); |
| 306 | |
| 307 | if ( is_int( $revisionOrId ) && $revisionOrId <= 0 ) { |
| 308 | throw new HttpError( 400, "Bad revision ID: $revisionOrId" ); |
| 309 | } |
| 310 | |
| 311 | $this->revisionOrId = $revisionOrId; |
| 312 | |
| 313 | if ( $this->getRevisionId() === null ) { |
| 314 | // If we have a RevisionRecord but no revision ID, we are dealing with a fake |
| 315 | // revision used for editor previews or mode switches. The wikitext is coming |
| 316 | // from the request, not the database, so the result is not cacheable for re-use |
| 317 | // by others (though it can be stashed for use by the same client). |
| 318 | $this->isCacheable = false; |
| 319 | } |
| 320 | } |
| 321 | |
| 322 | /** |
| 323 | * Set the content to render. Useful when rendering for previews |
| 324 | * or when switching the editor from source mode to visual mode. |
| 325 | * |
| 326 | * This will create a fake revision for rendering, the revision ID will be 0. |
| 327 | * |
| 328 | * @see setRevision |
| 329 | * @see setContentSource |
| 330 | * |
| 331 | * @param Content $content |
| 332 | */ |
| 333 | public function setContent( Content $content ): void { |
| 334 | $rev = new MutableRevisionRecord( $this->page ); |
| 335 | $rev->setId( 0 ); |
| 336 | $rev->setPageId( $this->page->getId() ); |
| 337 | $rev->setContent( SlotRecord::MAIN, $content ); |
| 338 | $this->setRevision( $rev ); |
| 339 | } |
| 340 | |
| 341 | /** |
| 342 | * Set the content to render. Useful when rendering for previews |
| 343 | * or when switching the editor from source mode to visual mode. |
| 344 | * |
| 345 | * This will create a fake revision for rendering, the revision ID will be 0. |
| 346 | * |
| 347 | * @param string $source The source data, e.g. wikitext |
| 348 | * @param string $model The content model indicating how to interpret $source, e.g. CONTENT_MODEL_WIKITEXT |
| 349 | * |
| 350 | * @see setRevision |
| 351 | * @see setContent |
| 352 | */ |
| 353 | public function setContentSource( string $source, string $model ): void { |
| 354 | try { |
| 355 | $handler = $this->contentHandlerFactory->getContentHandler( $model ); |
| 356 | $content = $handler->unserializeContent( $source ); |
| 357 | $this->setContent( $content ); |
| 358 | } catch ( MWUnknownContentModelException $ex ) { |
| 359 | throw new LocalizedHttpException( new MessageValue( "rest-bad-content-model", [ $model ] ), 400 ); |
| 360 | } |
| 361 | } |
| 362 | |
| 363 | /** |
| 364 | * This is equivalent of 'pageLanguageOverride' in PageConfigFactory |
| 365 | * For example, when clients call the REST API with the 'content-language' |
| 366 | * header to effect language variant conversion. |
| 367 | * |
| 368 | * @param Bcp47Code|string $pageLanguage the page language, as a Bcp47Code |
| 369 | * or a BCP-47 string. |
| 370 | */ |
| 371 | public function setPageLanguage( $pageLanguage ): void { |
| 372 | if ( is_string( $pageLanguage ) ) { |
| 373 | $pageLanguage = new Bcp47CodeValue( $pageLanguage ); |
| 374 | } |
| 375 | $this->pageLanguage = $pageLanguage; |
| 376 | } |
| 377 | |
| 378 | /** |
| 379 | * Initializes the helper with the given parameters like the page |
| 380 | * we're dealing with, parameters gotten from the request inputs, |
| 381 | * and the revision if any is available. |
| 382 | * |
| 383 | * @param PageIdentity $page |
| 384 | * @param array $parameters |
| 385 | * @param Authority $authority |
| 386 | * @param RevisionRecord|int|null $revision |
| 387 | */ |
| 388 | public function init( |
| 389 | PageIdentity $page, |
| 390 | array $parameters, |
| 391 | Authority $authority, |
| 392 | $revision = null |
| 393 | ) { |
| 394 | $this->page = $page; |
| 395 | $this->authority = $authority; |
| 396 | $this->stash = $parameters['stash'] ?? false; |
| 397 | |
| 398 | if ( $revision !== null ) { |
| 399 | $this->setRevision( $revision ); |
| 400 | } |
| 401 | |
| 402 | if ( $this->stash ) { |
| 403 | $this->setFlavor( 'stash' ); |
| 404 | } else { |
| 405 | $this->setFlavor( $parameters['flavor'] ?? 'view' ); |
| 406 | } |
| 407 | } |
| 408 | |
| 409 | /** |
| 410 | * @inheritDoc |
| 411 | */ |
| 412 | public function setVariantConversionLanguage( |
| 413 | $targetLanguage, |
| 414 | $sourceLanguage = null |
| 415 | ): void { |
| 416 | if ( is_string( $targetLanguage ) ) { |
| 417 | $targetLanguage = $this->getAcceptedTargetLanguage( $targetLanguage ); |
| 418 | $targetLanguage = LanguageCode::normalizeNonstandardCodeAndWarn( |
| 419 | $targetLanguage |
| 420 | ); |
| 421 | } |
| 422 | if ( is_string( $sourceLanguage ) ) { |
| 423 | $sourceLanguage = LanguageCode::normalizeNonstandardCodeAndWarn( |
| 424 | $sourceLanguage |
| 425 | ); |
| 426 | } |
| 427 | $this->targetLanguage = $targetLanguage; |
| 428 | $this->sourceLanguage = $sourceLanguage; |
| 429 | } |
| 430 | |
| 431 | /** |
| 432 | * Get a target language from an accept header |
| 433 | */ |
| 434 | private function getAcceptedTargetLanguage( string $targetLanguage ): string { |
| 435 | // We could try to identify the most desirable language here, |
| 436 | // following the rules for Accept-Language headers in RFC9100. |
| 437 | // For now, just take the first language code. |
| 438 | |
| 439 | if ( preg_match( '/^\s*([-\w]+)/', $targetLanguage, $m ) ) { |
| 440 | return $m[1]; |
| 441 | } else { |
| 442 | // "undetermined" per RFC5646 |
| 443 | return 'und'; |
| 444 | } |
| 445 | } |
| 446 | |
| 447 | /** |
| 448 | * @inheritDoc |
| 449 | */ |
| 450 | public function getHtml(): ParserOutput { |
| 451 | if ( $this->processedParserOutput ) { |
| 452 | return $this->processedParserOutput; |
| 453 | } |
| 454 | |
| 455 | $parserOutput = $this->getParserOutput(); |
| 456 | |
| 457 | if ( $this->stash ) { |
| 458 | $this->authorizeWriteOrThrow( $this->authority, 'stashbasehtml', $this->page ); |
| 459 | |
| 460 | $isFakeRevision = $this->getRevisionId() === null; |
| 461 | $parsoidStashKey = ParsoidRenderID::newFromParserOutput( $parserOutput ); |
| 462 | $stashSuccess = $this->parsoidOutputStash->set( |
| 463 | $parsoidStashKey, |
| 464 | new SelserContext( |
| 465 | PageBundleParserOutputConverter::pageBundleFromParserOutput( $parserOutput ), |
| 466 | $parsoidStashKey->getRevisionID(), |
| 467 | $isFakeRevision ? $this->revisionOrId->getContent( SlotRecord::MAIN ) : null |
| 468 | ) |
| 469 | ); |
| 470 | if ( !$stashSuccess ) { |
| 471 | $this->stats->increment( 'htmloutputrendererhelper.stash.fail' ); |
| 472 | |
| 473 | $errorData = [ 'parsoid-stash-key' => $parsoidStashKey ]; |
| 474 | LoggerFactory::getInstance( 'HtmlOutputRendererHelper' )->error( |
| 475 | "Parsoid stash failure", |
| 476 | $errorData |
| 477 | ); |
| 478 | throw new LocalizedHttpException( |
| 479 | MessageValue::new( 'rest-html-stash-failure' ), |
| 480 | 500, |
| 481 | $errorData |
| 482 | ); |
| 483 | } |
| 484 | $this->stats->increment( 'htmloutputrendererhelper.stash.save' ); |
| 485 | } |
| 486 | |
| 487 | if ( $this->flavor === 'edit' ) { |
| 488 | $pb = $this->getPageBundle(); |
| 489 | |
| 490 | // Inject data-parsoid and data-mw attributes. |
| 491 | // XXX: Would be nice if we had a DOM handy. |
| 492 | $doc = DOMUtils::parseHTML( $parserOutput->getRawText() ); |
| 493 | PageBundle::apply( $doc, $pb ); |
| 494 | $parserOutput->setRawText( ContentUtils::toXML( $doc ) ); |
| 495 | } |
| 496 | |
| 497 | // Check if variant conversion has to be performed |
| 498 | // NOTE: Variant conversion is performed on the fly, and kept outside the stash. |
| 499 | if ( $this->targetLanguage ) { |
| 500 | $languageVariantConverter = $this->htmlTransformFactory->getLanguageVariantConverter( $this->page ); |
| 501 | $parserOutput = $languageVariantConverter->convertParserOutputVariant( |
| 502 | $parserOutput, |
| 503 | $this->targetLanguage, |
| 504 | $this->sourceLanguage |
| 505 | ); |
| 506 | } |
| 507 | |
| 508 | $this->processedParserOutput = $parserOutput; |
| 509 | return $parserOutput; |
| 510 | } |
| 511 | |
| 512 | /** |
| 513 | * @inheritDoc |
| 514 | */ |
| 515 | public function getETag( string $suffix = '' ): ?string { |
| 516 | $parserOutput = $this->getParserOutput(); |
| 517 | |
| 518 | $renderID = ParsoidRenderID::newFromParserOutput( $parserOutput )->getKey(); |
| 519 | |
| 520 | if ( $suffix !== '' ) { |
| 521 | $eTag = "$renderID/{$this->flavor}/$suffix"; |
| 522 | } else { |
| 523 | $eTag = "$renderID/{$this->flavor}"; |
| 524 | } |
| 525 | |
| 526 | if ( $this->targetLanguage ) { |
| 527 | $eTag .= "+lang:{$this->targetLanguage->toBcp47Code()}"; |
| 528 | } |
| 529 | |
| 530 | return "\"{$eTag}\""; |
| 531 | } |
| 532 | |
| 533 | /** |
| 534 | * @inheritDoc |
| 535 | */ |
| 536 | public function getLastModified(): ?string { |
| 537 | return $this->getParserOutput()->getCacheTime(); |
| 538 | } |
| 539 | |
| 540 | /** |
| 541 | * @inheritDoc |
| 542 | */ |
| 543 | public function getParamSettings(): array { |
| 544 | return [ |
| 545 | 'stash' => [ |
| 546 | Handler::PARAM_SOURCE => 'query', |
| 547 | ParamValidator::PARAM_TYPE => 'boolean', |
| 548 | ParamValidator::PARAM_DEFAULT => false, |
| 549 | ParamValidator::PARAM_REQUIRED => false, |
| 550 | ], |
| 551 | 'flavor' => [ |
| 552 | Handler::PARAM_SOURCE => 'query', |
| 553 | ParamValidator::PARAM_TYPE => self::OUTPUT_FLAVORS, |
| 554 | ParamValidator::PARAM_DEFAULT => 'view', |
| 555 | ParamValidator::PARAM_REQUIRED => false, |
| 556 | ], |
| 557 | ]; |
| 558 | } |
| 559 | |
| 560 | private function getDefaultPageLanguage( ParserOptions $options ): Bcp47Code { |
| 561 | // NOTE: keep in sync with Parser::getTargetLanguage! |
| 562 | |
| 563 | // XXX: Inject a TitleFactory just for this?! We need a better way to determine the page language... |
| 564 | $title = Title::castFromPageIdentity( $this->page ); |
| 565 | |
| 566 | if ( $options->getInterfaceMessage() ) { |
| 567 | return $options->getUserLangObj(); |
| 568 | } |
| 569 | |
| 570 | return $title->getPageLanguage(); |
| 571 | } |
| 572 | |
| 573 | /** |
| 574 | * @return ParserOutput |
| 575 | */ |
| 576 | private function getParserOutput(): ParserOutput { |
| 577 | if ( !$this->parserOutput ) { |
| 578 | $parserOptions = ParserOptions::newFromAnon(); |
| 579 | $parserOptions->setRenderReason( __METHOD__ ); |
| 580 | |
| 581 | $defaultLanguage = $this->getDefaultPageLanguage( $parserOptions ); |
| 582 | |
| 583 | if ( $this->pageLanguage |
| 584 | && $this->pageLanguage->toBcp47Code() !== $defaultLanguage->toBcp47Code() |
| 585 | ) { |
| 586 | $languageObj = $this->languageFactory->getLanguage( $this->pageLanguage ); |
| 587 | $parserOptions->setTargetLanguage( $languageObj ); |
| 588 | } |
| 589 | |
| 590 | try { |
| 591 | $status = $this->getParserOutputInternal( $parserOptions ); |
| 592 | } catch ( RevisionAccessException $e ) { |
| 593 | throw new LocalizedHttpException( |
| 594 | MessageValue::new( 'rest-nonexistent-title' ), |
| 595 | 404, |
| 596 | [ 'reason' => $e->getMessage() ] |
| 597 | ); |
| 598 | } |
| 599 | |
| 600 | if ( !$status->isOK() ) { |
| 601 | if ( $status->hasMessage( 'parsoid-client-error' ) ) { |
| 602 | $this->throwExceptionForStatus( $status, 'rest-html-backend-error', 400 ); |
| 603 | } elseif ( $status->hasMessage( 'parsoid-resource-limit-exceeded' ) ) { |
| 604 | $this->throwExceptionForStatus( $status, 'rest-resource-limit-exceeded', 413 ); |
| 605 | } elseif ( $status->hasMessage( 'missing-revision-permission' ) ) { |
| 606 | $this->throwExceptionForStatus( $status, 'rest-permission-denied-revision', 403 ); |
| 607 | } elseif ( $status->hasMessage( 'parsoid-revision-access' ) ) { |
| 608 | $this->throwExceptionForStatus( $status, 'rest-specified-revision-unavailable', 404 ); |
| 609 | } else { |
| 610 | $this->logStatusError( $status, 'Parsoid backend error', 'HtmlOutputRendererHelper' ); |
| 611 | $this->throwExceptionForStatus( $status, 'rest-html-backend-error', 500 ); |
| 612 | } |
| 613 | } |
| 614 | |
| 615 | $this->parserOutput = $status->getValue(); |
| 616 | } |
| 617 | |
| 618 | return $this->parserOutput; |
| 619 | } |
| 620 | |
| 621 | /** |
| 622 | * The content language of the HTML output after parsing. |
| 623 | * |
| 624 | * @return Bcp47Code The language, as a BCP-47 code |
| 625 | */ |
| 626 | public function getHtmlOutputContentLanguage(): Bcp47Code { |
| 627 | $contentLanguage = $this->getHtml()->getLanguage(); |
| 628 | |
| 629 | // This shouldn't happen, but don't crash if it does: |
| 630 | if ( !$contentLanguage ) { |
| 631 | if ( $this->pageLanguage ) { |
| 632 | LoggerFactory::getInstance( 'HtmlOutputRendererHelper' )->warning( |
| 633 | "ParserOutput does not specify a language" |
| 634 | ); |
| 635 | |
| 636 | $contentLanguage = $this->pageLanguage; |
| 637 | } else { |
| 638 | LoggerFactory::getInstance( 'HtmlOutputRendererHelper' )->warning( |
| 639 | "ParserOutput does not specify a language and no page language set in helper." |
| 640 | ); |
| 641 | |
| 642 | $title = Title::newFromPageIdentity( $this->page ); |
| 643 | $contentLanguage = $title->getPageLanguage(); |
| 644 | } |
| 645 | } |
| 646 | |
| 647 | return $contentLanguage; |
| 648 | } |
| 649 | |
| 650 | /** |
| 651 | * @inheritDoc |
| 652 | */ |
| 653 | public function putHeaders( ResponseInterface $response, bool $forHtml = true ): void { |
| 654 | if ( $forHtml ) { |
| 655 | // For HTML we want to set the Content-Language. For JSON, we probably don't. |
| 656 | $response->setHeader( 'Content-Language', $this->getHtmlOutputContentLanguage()->toBcp47Code() ); |
| 657 | |
| 658 | $pb = $this->getPageBundle(); |
| 659 | ParsoidFormatHelper::setContentType( $response, ParsoidFormatHelper::FORMAT_HTML, $pb->version ); |
| 660 | } |
| 661 | |
| 662 | if ( $this->targetLanguage ) { |
| 663 | $response->addHeader( 'Vary', 'Accept-Language' ); |
| 664 | } |
| 665 | |
| 666 | // XXX: if Parsoid returns Vary headers, set them here?! |
| 667 | |
| 668 | if ( !$this->isCacheable ) { |
| 669 | $response->setHeader( 'Cache-Control', 'private,no-cache,s-maxage=0' ); |
| 670 | } |
| 671 | |
| 672 | // TODO: cache control for stable HTML? See ContentHelper::setCacheControl |
| 673 | |
| 674 | if ( $this->getRevisionId() ) { |
| 675 | $response->setHeader( 'Content-Revision-Id', (string)$this->getRevisionId() ); |
| 676 | } |
| 677 | } |
| 678 | |
| 679 | /** |
| 680 | * Returns the rendered HTML as a PageBundle object. |
| 681 | * |
| 682 | * @return PageBundle |
| 683 | */ |
| 684 | public function getPageBundle(): PageBundle { |
| 685 | // XXX: converting between PageBundle and ParserOutput is inefficient! |
| 686 | $parserOutput = $this->getParserOutput(); |
| 687 | $pb = PageBundleParserOutputConverter::pageBundleFromParserOutput( $parserOutput ); |
| 688 | |
| 689 | // Check if variant conversion has to be performed |
| 690 | // NOTE: Variant conversion is performed on the fly, and kept outside the stash. |
| 691 | if ( $this->targetLanguage ) { |
| 692 | $languageVariantConverter = $this->htmlTransformFactory->getLanguageVariantConverter( $this->page ); |
| 693 | $pb = $languageVariantConverter->convertPageBundleVariant( |
| 694 | $pb, |
| 695 | $this->targetLanguage, |
| 696 | $this->sourceLanguage |
| 697 | ); |
| 698 | } |
| 699 | |
| 700 | return $pb; |
| 701 | } |
| 702 | |
| 703 | /** |
| 704 | * Returns the ID of the revision that is being rendered. |
| 705 | * |
| 706 | * This will return 0 if no revision has been specified, so the current revision |
| 707 | * will be rendered. |
| 708 | * |
| 709 | * This wil return null if RevisionRecord has been set but that RevisionRecord |
| 710 | * does not have a revision ID, e.g. when rendering a preview. |
| 711 | * |
| 712 | * @return ?int |
| 713 | */ |
| 714 | public function getRevisionId(): ?int { |
| 715 | if ( !$this->revisionOrId ) { |
| 716 | // If we don't have a revision set or it's 0, we are rendering the current revision. |
| 717 | return 0; |
| 718 | } |
| 719 | |
| 720 | if ( is_object( $this->revisionOrId ) ) { |
| 721 | // NOTE: return null even of getId() gave us 0 |
| 722 | return $this->revisionOrId->getId() ?: null; |
| 723 | } |
| 724 | |
| 725 | // It's a revision ID, just return it |
| 726 | return (int)$this->revisionOrId; |
| 727 | } |
| 728 | |
| 729 | /** |
| 730 | * Strip Parsoid's section wrappers |
| 731 | * |
| 732 | * TODO: Should we move this to Parsoid's ContentUtils class? |
| 733 | * There already is a stripUnnecessaryWrappersAndSyntheticNodes but |
| 734 | * it targets html2wt and does a lot more than just section unwrapping. |
| 735 | * |
| 736 | * @param Element $elt |
| 737 | */ |
| 738 | private function stripParsoidSectionTags( Element $elt ): void { |
| 739 | $n = $elt->firstChild; |
| 740 | while ( $n ) { |
| 741 | $next = $n->nextSibling; |
| 742 | if ( $n instanceof Element ) { |
| 743 | // Recurse into subtree before stripping this |
| 744 | $this->stripParsoidSectionTags( $n ); |
| 745 | // Strip <section> tags and synthetic extended-annotation-region wrappers |
| 746 | if ( WTUtils::isParsoidSectionTag( $n ) ) { |
| 747 | $parent = $n->parentNode; |
| 748 | // Help out phan |
| 749 | '@phan-var Element $parent'; |
| 750 | DOMUtils::migrateChildren( $n, $parent, $n ); |
| 751 | $parent->removeChild( $n ); |
| 752 | } |
| 753 | } |
| 754 | $n = $next; |
| 755 | } |
| 756 | } |
| 757 | |
| 758 | /** |
| 759 | * @param ParserOptions $parserOptions |
| 760 | * |
| 761 | * @return Status |
| 762 | */ |
| 763 | private function getParserOutputInternal( ParserOptions $parserOptions ): Status { |
| 764 | // NOTE: ParsoidOutputAccess::getParserOutput() should be used for revisions |
| 765 | // that comes from the database. Either this revision is null to indicate |
| 766 | // the current revision or the revision must have an ID. |
| 767 | // If we have a revision and the ID is 0 or null, then it's a fake revision |
| 768 | // representing a preview. |
| 769 | $parsoidOptions = $this->parsoidOptions; |
| 770 | // NOTE: VisualEditor would set this flavor when transforming from Wikitext to HTML |
| 771 | // for the purpose of editing when doing parsefragment (in body only mode). |
| 772 | if ( $this->flavor === 'fragment' || $this->getRevisionId() === null ) { |
| 773 | $this->isCacheable = false; |
| 774 | } |
| 775 | |
| 776 | // TODO: Decide whether we want to allow stale content for speed for the |
| 777 | // 'view' flavor. In that case, we would want to use PoolCounterWork, |
| 778 | // either directly or through ParserOutputAccess. |
| 779 | |
| 780 | if ( $this->isCacheable ) { |
| 781 | $flags = $this->parsoidOutputAccessOptions; |
| 782 | $status = $this->parsoidOutputAccess->getParserOutput( |
| 783 | $this->page, |
| 784 | $parserOptions, |
| 785 | $this->revisionOrId, |
| 786 | $flags, |
| 787 | $this->lenientRevHandling |
| 788 | ); |
| 789 | } else { |
| 790 | $status = $this->parsoidOutputAccess->parseUncacheable( |
| 791 | $this->page, |
| 792 | $parserOptions, |
| 793 | $this->revisionOrId, |
| 794 | $this->lenientRevHandling |
| 795 | ); |
| 796 | |
| 797 | // @phan-suppress-next-line PhanSuspiciousValueComparison |
| 798 | if ( $status->isOK() && $this->flavor === 'fragment' ) { |
| 799 | // Unwrap sections and return body_only content |
| 800 | // NOTE: This introduces an extra html -> dom -> html roundtrip |
| 801 | // This will get addressed once HtmlHolder work is complete |
| 802 | $parserOutput = $status->getValue(); |
| 803 | $body = DOMCompat::getBody( DOMUtils::parseHTML( $parserOutput->getRawText() ) ); |
| 804 | if ( $body ) { |
| 805 | $this->stripParsoidSectionTags( $body ); |
| 806 | $parserOutput->setText( DOMCompat::getInnerHTML( $body ) ); |
| 807 | } |
| 808 | } |
| 809 | } |
| 810 | |
| 811 | return $status; |
| 812 | } |
| 813 | |
| 814 | } |