Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 226 |
|
0.00% |
0 / 13 |
CRAP | |
0.00% |
0 / 1 |
| Wikimedia\Parsoid\Wt2Html\TT\array_flatten | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
12 | |||
| QuoteTransformer | |
0.00% |
0 / 220 |
|
0.00% |
0 / 12 |
6480 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| reset | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
2 | |||
| startNewChunk | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| onTag | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
20 | |||
| onNewline | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| onEnd | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| onAny | |
0.00% |
0 / 12 |
|
0.00% |
0 / 1 |
6 | |||
| onQuote | |
0.00% |
0 / 16 |
|
0.00% |
0 / 1 |
20 | |||
| processQuotes | |
0.00% |
0 / 62 |
|
0.00% |
0 / 1 |
552 | |||
| convertBold | |
0.00% |
0 / 13 |
|
0.00% |
0 / 1 |
20 | |||
| convertQuotesToTags | |
0.00% |
0 / 77 |
|
0.00% |
0 / 1 |
650 | |||
| quoteToTag | |
0.00% |
0 / 24 |
|
0.00% |
0 / 1 |
182 | |||
| 1 | <?php |
| 2 | declare( strict_types = 1 ); |
| 3 | |
| 4 | namespace Wikimedia\Parsoid\Wt2Html\TT; |
| 5 | |
| 6 | use Wikimedia\Assert\Assert; |
| 7 | use Wikimedia\Parsoid\NodeData\DataParsoid; |
| 8 | use Wikimedia\Parsoid\Tokens\EndTagTk; |
| 9 | use Wikimedia\Parsoid\Tokens\EOFTk; |
| 10 | use Wikimedia\Parsoid\Tokens\NlTk; |
| 11 | use Wikimedia\Parsoid\Tokens\SelfclosingTagTk; |
| 12 | use Wikimedia\Parsoid\Tokens\SourceRange; |
| 13 | use Wikimedia\Parsoid\Tokens\TagTk; |
| 14 | use Wikimedia\Parsoid\Tokens\Token; |
| 15 | use Wikimedia\Parsoid\Utils\PHPUtils; |
| 16 | use Wikimedia\Parsoid\Wt2html\TokenTransformManager; |
| 17 | |
| 18 | /** |
| 19 | * PORT-FIXME: Maybe we need to look at all uses of flatten |
| 20 | * and move it to a real helper in PHPUtils.js |
| 21 | * |
| 22 | * Flattens arrays with nested arrays |
| 23 | * |
| 24 | * @param array $array array |
| 25 | * @return array |
| 26 | */ |
| 27 | function array_flatten( array $array ): array { |
| 28 | $ret = []; |
| 29 | foreach ( $array as $key => $value ) { |
| 30 | if ( is_array( $value ) ) { |
| 31 | PHPUtils::pushArray( $ret, array_flatten( $value ) ); |
| 32 | } else { |
| 33 | $ret[$key] = $value; |
| 34 | } |
| 35 | } |
| 36 | return $ret; |
| 37 | } |
| 38 | |
| 39 | /** |
| 40 | * MediaWiki-compatible italic/bold handling as a token stream transformation. |
| 41 | */ |
| 42 | class QuoteTransformer extends TokenHandler { |
| 43 | /** Chunks alternate between quote tokens and sequences of non-quote |
| 44 | * tokens. The quote tokens are later replaced with the actual tag |
| 45 | * token for italic or bold. The first chunk is a non-quote chunk. |
| 46 | * @var array |
| 47 | */ |
| 48 | private $chunks; |
| 49 | |
| 50 | /** |
| 51 | * The current chunk we're accumulating into. |
| 52 | * @var array |
| 53 | */ |
| 54 | private $currentChunk; |
| 55 | |
| 56 | /** |
| 57 | * Last italic / last bold open tag seen. Used to add autoInserted flags |
| 58 | * where necessary. |
| 59 | * @var array |
| 60 | */ |
| 61 | private $last; |
| 62 | |
| 63 | /** |
| 64 | * @param TokenTransformManager $manager manager environment |
| 65 | * @param array $options options |
| 66 | */ |
| 67 | public function __construct( TokenTransformManager $manager, array $options ) { |
| 68 | parent::__construct( $manager, $options ); |
| 69 | $this->reset(); |
| 70 | } |
| 71 | |
| 72 | /** |
| 73 | * Reset the buffering of chunks |
| 74 | * |
| 75 | */ |
| 76 | private function reset(): void { |
| 77 | // Chunks alternate between quote tokens and sequences of non-quote |
| 78 | // tokens. The quote tokens are later replaced with the actual tag |
| 79 | // token for italic or bold. The first chunk is a non-quote chunk. |
| 80 | $this->chunks = []; |
| 81 | // The current chunk we're accumulating into. |
| 82 | $this->currentChunk = []; |
| 83 | // last italic / last bold open tag seen. Used to add autoInserted flags |
| 84 | // where necessary. |
| 85 | $this->last = []; |
| 86 | |
| 87 | $this->onAnyEnabled = false; |
| 88 | } |
| 89 | |
| 90 | /** |
| 91 | * Make a copy of the token context |
| 92 | * |
| 93 | */ |
| 94 | private function startNewChunk(): void { |
| 95 | $this->chunks[] = $this->currentChunk; |
| 96 | $this->currentChunk = []; |
| 97 | } |
| 98 | |
| 99 | /** |
| 100 | * Handles mw-quote tokens and td/th tokens |
| 101 | * @inheritDoc |
| 102 | */ |
| 103 | public function onTag( Token $token ): ?TokenHandlerResult { |
| 104 | $tkName = $token->getName(); |
| 105 | if ( $tkName === 'mw-quote' ) { |
| 106 | return $this->onQuote( $token ); |
| 107 | } elseif ( $tkName === 'td' || $tkName === 'th' ) { |
| 108 | return $this->processQuotes( $token ); |
| 109 | } else { |
| 110 | return null; |
| 111 | } |
| 112 | } |
| 113 | |
| 114 | /** |
| 115 | * On encountering a NlTk, processes quotes on the current line |
| 116 | * @inheritDoc |
| 117 | */ |
| 118 | public function onNewline( NlTk $token ): ?TokenHandlerResult { |
| 119 | return $this->processQuotes( $token ); |
| 120 | } |
| 121 | |
| 122 | /** |
| 123 | * On encountering an EOFTk, processes quotes on the current line |
| 124 | * @inheritDoc |
| 125 | */ |
| 126 | public function onEnd( EOFTk $token ): ?TokenHandlerResult { |
| 127 | return $this->processQuotes( $token ); |
| 128 | } |
| 129 | |
| 130 | /** |
| 131 | * @inheritDoc |
| 132 | */ |
| 133 | public function onAny( $token ): ?TokenHandlerResult { |
| 134 | $this->env->log( |
| 135 | "trace/quote", |
| 136 | $this->pipelineId, |
| 137 | "ANY |", |
| 138 | static function () use ( $token ) { |
| 139 | return PHPUtils::jsonEncode( $token ); |
| 140 | } |
| 141 | ); |
| 142 | |
| 143 | if ( $this->onAnyEnabled ) { |
| 144 | $this->currentChunk[] = $token; |
| 145 | return new TokenHandlerResult( [] ); |
| 146 | } else { |
| 147 | return null; |
| 148 | } |
| 149 | } |
| 150 | |
| 151 | /** |
| 152 | * Handle QUOTE tags. These are collected in italic/bold lists depending on |
| 153 | * the length of quote string. Actual analysis and conversion to the |
| 154 | * appropriate tag tokens is deferred until the next NEWLINE token triggers |
| 155 | * processQuotes. |
| 156 | * |
| 157 | * @param Token $token token |
| 158 | * @return TokenHandlerResult |
| 159 | */ |
| 160 | private function onQuote( Token $token ): TokenHandlerResult { |
| 161 | $v = $token->getAttributeV( 'value' ); |
| 162 | $qlen = strlen( $v ); |
| 163 | $this->env->log( |
| 164 | "trace/quote", |
| 165 | $this->pipelineId, |
| 166 | "QUOTE |", |
| 167 | static function () use ( $token ) { |
| 168 | return PHPUtils::jsonEncode( $token ); |
| 169 | } |
| 170 | ); |
| 171 | |
| 172 | $this->onAnyEnabled = true; |
| 173 | |
| 174 | if ( $qlen === 2 || $qlen === 3 || $qlen === 5 ) { |
| 175 | $this->startNewChunk(); |
| 176 | $this->currentChunk[] = $token; |
| 177 | $this->startNewChunk(); |
| 178 | } |
| 179 | |
| 180 | return new TokenHandlerResult( [] ); |
| 181 | } |
| 182 | |
| 183 | /** |
| 184 | * Handle NEWLINE tokens, which trigger the actual quote analysis on the |
| 185 | * collected quote tokens so far. |
| 186 | * |
| 187 | * @param Token $token token |
| 188 | * @return TokenHandlerResult|null |
| 189 | */ |
| 190 | private function processQuotes( Token $token ): ?TokenHandlerResult { |
| 191 | if ( !$this->onAnyEnabled ) { |
| 192 | // Nothing to do, quick abort. |
| 193 | return null; |
| 194 | } |
| 195 | |
| 196 | $this->env->log( |
| 197 | "trace/quote", |
| 198 | $this->pipelineId, |
| 199 | "NL |", |
| 200 | static function () use( $token ) { |
| 201 | return PHPUtils::jsonEncode( $token ); |
| 202 | } |
| 203 | ); |
| 204 | |
| 205 | if ( |
| 206 | ( $token->getName() === 'td' || $token->getName() === 'th' ) && |
| 207 | ( $token->dataParsoid->stx ?? '' ) === 'html' |
| 208 | ) { |
| 209 | return null; |
| 210 | } |
| 211 | |
| 212 | // count number of bold and italics |
| 213 | $numbold = 0; |
| 214 | $numitalics = 0; |
| 215 | $chunkCount = count( $this->chunks ); |
| 216 | for ( $i = 1; $i < $chunkCount; $i += 2 ) { |
| 217 | // quote token |
| 218 | Assert::invariant( count( $this->chunks[$i] ) === 1, 'Expected a single token in the chunk' ); |
| 219 | $qlen = strlen( $this->chunks[$i][0]->getAttributeV( "value" ) ); |
| 220 | if ( $qlen === 2 || $qlen === 5 ) { |
| 221 | $numitalics++; |
| 222 | } |
| 223 | if ( $qlen === 3 || $qlen === 5 ) { |
| 224 | $numbold++; |
| 225 | } |
| 226 | } |
| 227 | |
| 228 | // balance out tokens, convert placeholders into tags |
| 229 | if ( ( $numitalics % 2 === 1 ) && ( $numbold % 2 === 1 ) ) { |
| 230 | $firstsingleletterword = -1; |
| 231 | $firstmultiletterword = -1; |
| 232 | $firstspace = -1; |
| 233 | $chunkCount = count( $this->chunks ); |
| 234 | for ( $i = 1; $i < $chunkCount; $i += 2 ) { |
| 235 | // only look at bold tags |
| 236 | if ( strlen( $this->chunks[$i][0]->getAttributeV( "value" ) ) !== 3 ) { |
| 237 | continue; |
| 238 | } |
| 239 | |
| 240 | $tok = $this->chunks[$i][0]; |
| 241 | $lastCharIsSpace = $tok->getAttributeV( 'isSpace_1' ); |
| 242 | $secondLastCharIsSpace = $tok->getAttributeV( 'isSpace_2' ); |
| 243 | if ( $lastCharIsSpace && $firstspace === -1 ) { |
| 244 | $firstspace = $i; |
| 245 | } elseif ( !$lastCharIsSpace ) { |
| 246 | if ( $secondLastCharIsSpace && $firstsingleletterword === -1 ) { |
| 247 | $firstsingleletterword = $i; |
| 248 | // if firstsingleletterword is set, we don't need |
| 249 | // to look at the options options, so we can bail early |
| 250 | break; |
| 251 | } elseif ( $firstmultiletterword === -1 ) { |
| 252 | $firstmultiletterword = $i; |
| 253 | } |
| 254 | } |
| 255 | } |
| 256 | |
| 257 | // now see if we can convert a bold to an italic and an apostrophe |
| 258 | if ( $firstsingleletterword > -1 ) { |
| 259 | $this->convertBold( $firstsingleletterword ); |
| 260 | } elseif ( $firstmultiletterword > -1 ) { |
| 261 | $this->convertBold( $firstmultiletterword ); |
| 262 | } elseif ( $firstspace > -1 ) { |
| 263 | $this->convertBold( $firstspace ); |
| 264 | } else { |
| 265 | // (notice that it is possible for all three to be -1 if, for |
| 266 | // example, there is only one pentuple-apostrophe in the line) |
| 267 | // In this case, do no balancing. |
| 268 | } |
| 269 | } |
| 270 | |
| 271 | // convert the quote tokens into tags |
| 272 | $this->convertQuotesToTags(); |
| 273 | |
| 274 | // return all collected tokens including the newline |
| 275 | $this->currentChunk[] = $token; |
| 276 | $this->startNewChunk(); |
| 277 | // PORT-FIXME: Is there a more efficient way of doing this? |
| 278 | $res = new TokenHandlerResult( array_flatten( $this->chunks ) ); |
| 279 | |
| 280 | $this->env->log( |
| 281 | "trace/quote", |
| 282 | $this->pipelineId, |
| 283 | "----->", |
| 284 | static function () use ( $res ) { |
| 285 | return PHPUtils::jsonEncode( $res->tokens ); |
| 286 | } |
| 287 | ); |
| 288 | |
| 289 | // prepare for next line |
| 290 | $this->reset(); |
| 291 | |
| 292 | return $res; |
| 293 | } |
| 294 | |
| 295 | /** |
| 296 | * Convert a bold token to italic to balance an uneven number of both bold and |
| 297 | * italic tags. In the process, one quote needs to be converted back to text. |
| 298 | * |
| 299 | * @param int $i index into chunks |
| 300 | */ |
| 301 | private function convertBold( int $i ): void { |
| 302 | // this should be a bold tag. |
| 303 | Assert::invariant( $i > 0 && count( $this->chunks[$i] ) === 1 |
| 304 | && strlen( $this->chunks[$i][0]->getAttributeV( "value" ) ) === 3, |
| 305 | 'this should be a bold tag' ); |
| 306 | |
| 307 | // we're going to convert it to a single plain text ' plus an italic tag |
| 308 | $this->chunks[$i - 1][] = "'"; |
| 309 | $oldbold = $this->chunks[$i][0]; |
| 310 | $tsr = $oldbold->dataParsoid->tsr ?? null; |
| 311 | if ( $tsr ) { |
| 312 | $tsr = new SourceRange( $tsr->start + 1, $tsr->end ); |
| 313 | } |
| 314 | $dp = new DataParsoid; |
| 315 | $dp->tsr = $tsr; |
| 316 | $newbold = new SelfclosingTagTk( 'mw-quote', [], $dp ); |
| 317 | $newbold->setAttribute( "value", "''" ); // italic! |
| 318 | $this->chunks[$i] = [ $newbold ]; |
| 319 | } |
| 320 | |
| 321 | /** |
| 322 | * Convert quote tokens to tags, using the same state machine as the |
| 323 | * PHP parser uses. |
| 324 | */ |
| 325 | private function convertQuotesToTags(): void { |
| 326 | $lastboth = -1; |
| 327 | $state = ''; |
| 328 | |
| 329 | $chunkCount = count( $this->chunks ); |
| 330 | for ( $i = 1; $i < $chunkCount; $i += 2 ) { |
| 331 | Assert::invariant( count( $this->chunks[$i] ) === 1, 'expected count chunks[i] == 1' ); |
| 332 | $qlen = strlen( $this->chunks[$i][0]->getAttributeV( "value" ) ); |
| 333 | if ( $qlen === 2 ) { |
| 334 | if ( $state === 'i' ) { |
| 335 | $this->quoteToTag( $i, [ new EndTagTk( 'i' ) ] ); |
| 336 | $state = ''; |
| 337 | } elseif ( $state === 'bi' ) { |
| 338 | $this->quoteToTag( $i, [ new EndTagTk( 'i' ) ] ); |
| 339 | $state = 'b'; |
| 340 | } elseif ( $state === 'ib' ) { |
| 341 | // annoying! |
| 342 | $this->quoteToTag( $i, [ |
| 343 | new EndTagTk( 'b' ), |
| 344 | new EndTagTk( 'i' ), |
| 345 | new TagTk( 'b' ), |
| 346 | ], true ); |
| 347 | $state = 'b'; |
| 348 | } elseif ( $state === 'both' ) { |
| 349 | $this->quoteToTag( $lastboth, [ new TagTk( 'b' ), new TagTk( 'i' ) ] ); |
| 350 | $this->quoteToTag( $i, [ new EndTagTk( 'i' ) ] ); |
| 351 | $state = 'b'; |
| 352 | } else { // state can be 'b' or '' |
| 353 | $this->quoteToTag( $i, [ new TagTk( 'i' ) ] ); |
| 354 | $state .= 'i'; |
| 355 | } |
| 356 | } elseif ( $qlen === 3 ) { |
| 357 | if ( $state === 'b' ) { |
| 358 | $this->quoteToTag( $i, [ new EndTagTk( 'b' ) ] ); |
| 359 | $state = ''; |
| 360 | } elseif ( $state === 'ib' ) { |
| 361 | $this->quoteToTag( $i, [ new EndTagTk( 'b' ) ] ); |
| 362 | $state = 'i'; |
| 363 | } elseif ( $state === 'bi' ) { |
| 364 | // annoying! |
| 365 | $this->quoteToTag( $i, [ |
| 366 | new EndTagTk( 'i' ), |
| 367 | new EndTagTk( 'b' ), |
| 368 | new TagTk( 'i' ), |
| 369 | ], true ); |
| 370 | $state = 'i'; |
| 371 | } elseif ( $state === 'both' ) { |
| 372 | $this->quoteToTag( $lastboth, [ new TagTk( 'i' ), new TagTk( 'b' ) ] ); |
| 373 | $this->quoteToTag( $i, [ new EndTagTk( 'b' ) ] ); |
| 374 | $state = 'i'; |
| 375 | } else { // state can be 'i' or '' |
| 376 | $this->quoteToTag( $i, [ new TagTk( 'b' ) ] ); |
| 377 | $state .= 'b'; |
| 378 | } |
| 379 | } elseif ( $qlen === 5 ) { |
| 380 | if ( $state === 'b' ) { |
| 381 | $this->quoteToTag( $i, [ new EndTagTk( 'b' ), new TagTk( 'i' ) ] ); |
| 382 | $state = 'i'; |
| 383 | } elseif ( $state === 'i' ) { |
| 384 | $this->quoteToTag( $i, [ new EndTagTk( 'i' ), new TagTk( 'b' ) ] ); |
| 385 | $state = 'b'; |
| 386 | } elseif ( $state === 'bi' ) { |
| 387 | $this->quoteToTag( $i, [ new EndTagTk( 'i' ), new EndTagTk( 'b' ) ] ); |
| 388 | $state = ''; |
| 389 | } elseif ( $state === 'ib' ) { |
| 390 | $this->quoteToTag( $i, [ new EndTagTk( 'b' ), new EndTagTk( 'i' ) ] ); |
| 391 | $state = ''; |
| 392 | } elseif ( $state === 'both' ) { |
| 393 | $this->quoteToTag( $lastboth, [ new TagTk( 'i' ), new TagTk( 'b' ) ] ); |
| 394 | $this->quoteToTag( $i, [ new EndTagTk( 'b' ), new EndTagTk( 'i' ) ] ); |
| 395 | $state = ''; |
| 396 | } else { // state == '' |
| 397 | $lastboth = $i; |
| 398 | $state = 'both'; |
| 399 | } |
| 400 | } |
| 401 | } |
| 402 | |
| 403 | // now close all remaining tags. notice that order is important. |
| 404 | if ( $state === 'both' ) { |
| 405 | $this->quoteToTag( $lastboth, [ new TagTk( 'b' ), new TagTk( 'i' ) ] ); |
| 406 | $state = 'bi'; |
| 407 | } |
| 408 | if ( $state === 'b' || $state === 'ib' ) { |
| 409 | $this->currentChunk[] = new EndTagTk( 'b' ); |
| 410 | $this->last["b"]->dataParsoid->autoInsertedEndToken = true; |
| 411 | } |
| 412 | if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) { |
| 413 | $this->currentChunk[] = new EndTagTk( 'i' ); |
| 414 | $this->last["i"]->dataParsoid->autoInsertedEndToken = true; |
| 415 | } |
| 416 | if ( $state === 'bi' ) { |
| 417 | $this->currentChunk[] = new EndTagTk( 'b' ); |
| 418 | $this->last["b"]->dataParsoid->autoInsertedEndToken = true; |
| 419 | } |
| 420 | } |
| 421 | |
| 422 | /** |
| 423 | * Convert italics/bolds into tags. |
| 424 | * |
| 425 | * @param int $chunk chunk buffer |
| 426 | * @param array $tags token |
| 427 | * @param bool $ignoreBogusTwo optional defaults to false |
| 428 | */ |
| 429 | private function quoteToTag( int $chunk, array $tags, bool $ignoreBogusTwo = false ): void { |
| 430 | Assert::invariant( count( $this->chunks[$chunk] ) === 1, 'expected count chunks[i] == 1' ); |
| 431 | $result = []; |
| 432 | $oldtag = $this->chunks[$chunk][0]; |
| 433 | // make tsr |
| 434 | $tsr = $oldtag->dataParsoid->tsr ?? null; |
| 435 | $startpos = $tsr ? $tsr->start : null; |
| 436 | $endpos = $tsr ? $tsr->end : null; |
| 437 | $numTags = count( $tags ); |
| 438 | for ( $i = 0; $i < $numTags; $i++ ) { |
| 439 | if ( $tsr ) { |
| 440 | if ( $i === 0 && $ignoreBogusTwo ) { |
| 441 | $this->last[$tags[$i]->getName()]->dataParsoid->autoInsertedEndToken = true; |
| 442 | } elseif ( $i === 2 && $ignoreBogusTwo ) { |
| 443 | $tags[$i]->dataParsoid->autoInsertedStartToken = true; |
| 444 | } elseif ( $tags[$i]->getName() === 'b' ) { |
| 445 | $tags[$i]->dataParsoid->tsr = new SourceRange( $startpos, $startpos + 3 ); |
| 446 | $startpos = $tags[$i]->dataParsoid->tsr->end; |
| 447 | } elseif ( $tags[$i]->getName() === 'i' ) { |
| 448 | $tags[$i]->dataParsoid->tsr = new SourceRange( $startpos, $startpos + 2 ); |
| 449 | $startpos = $tags[$i]->dataParsoid->tsr->end; |
| 450 | } |
| 451 | } |
| 452 | $this->last[$tags[$i]->getName()] = ( $tags[$i]->getType() === "EndTagTk" ) ? null : $tags[$i]; |
| 453 | $result[] = $tags[$i]; |
| 454 | } |
| 455 | if ( $tsr ) { |
| 456 | Assert::invariant( $startpos === $endpos, 'Start: $startpos !== end: $endpos' ); |
| 457 | } |
| 458 | $this->chunks[$chunk] = $result; |
| 459 | } |
| 460 | } |