includes/parser/Preprocessor.php

   1 <?php
   2 /**
   3  * Interfaces for preprocessors
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Parser
  22  */
  23
  24 use MediaWiki\Logger\LoggerFactory;
  25
  26 /**
  27  * @ingroup Parser
  28  */
  29 abstract class Preprocessor {
  30
  31         const CACHE_VERSION = 1;
  32
  33         /**
  34          * @var array Brace matching rules.
  35          */
  36         protected $rules = [
  37                 '{' => [
  38                         'end' => '}',
  39                         'names' => [
  40                                 2 => 'template',
  41                                 3 => 'tplarg',
  42                         ],
  43                         'min' => 2,
  44                         'max' => 3,
  45                 ],
  46                 '[' => [
  47                         'end' => ']',
  48                         'names' => [ 2 => null ],
  49                         'min' => 2,
  50                         'max' => 2,
  51                 ],
  52                 '-{' => [
  53                         'end' => '}-',
  54                         'names' => [ 2 => null ],
  55                         'min' => 2,
  56                         'max' => 2,
  57                 ],
  58         ];
  59
  60         /**
  61          * Store a document tree in the cache.
  62          *
  63          * @param string $text
  64          * @param int $flags
  65          * @param string $tree
  66          */
  67         protected function cacheSetTree( $text, $flags, $tree ) {
  68                 $config = RequestContext::getMain()->getConfig();
  69
  70                 $length = strlen( $text );
  71                 $threshold = $config->get( 'PreprocessorCacheThreshold' );
  72                 if ( $threshold === false || $length < $threshold || $length > 1e6 ) {
  73                         return;
  74                 }
  75
  76                 $cache = ObjectCache::getLocalClusterInstance();
  77                 $key = $cache->makeKey(
  78                         defined( 'static::CACHE_PREFIX' ) ? static::CACHE_PREFIX : static::class,
  79                         md5( $text ), $flags );
  80                 $value = sprintf( "%08d", static::CACHE_VERSION ) . $tree;
  81
  82                 $cache->set( $key, $value, 86400 );
  83
  84                 LoggerFactory::getInstance( 'Preprocessor' )
  85                         ->info( "Cached preprocessor output (key: $key)" );
  86         }
  87
  88         /**
  89          * Attempt to load a precomputed document tree for some given wikitext
  90          * from the cache.
  91          *
  92          * @param string $text
  93          * @param int $flags
  94          * @return PPNode_Hash_Tree|bool
  95          */
  96         protected function cacheGetTree( $text, $flags ) {
  97                 $config = RequestContext::getMain()->getConfig();
  98
  99                 $length = strlen( $text );
 100                 $threshold = $config->get( 'PreprocessorCacheThreshold' );
 101                 if ( $threshold === false || $length < $threshold || $length > 1e6 ) {
 102                         return false;
 103                 }
 104
 105                 $cache = ObjectCache::getLocalClusterInstance();
 106
 107                 $key = $cache->makeKey(
 108                         defined( 'static::CACHE_PREFIX' ) ? static::CACHE_PREFIX : static::class,
 109                         md5( $text ), $flags );
 110
 111                 $value = $cache->get( $key );
 112                 if ( !$value ) {
 113                         return false;
 114                 }
 115
 116                 $version = intval( substr( $value, 0, 8 ) );
 117                 if ( $version !== static::CACHE_VERSION ) {
 118                         return false;
 119                 }
 120
 121                 LoggerFactory::getInstance( 'Preprocessor' )
 122                         ->info( "Loaded preprocessor output from cache (key: $key)" );
 123
 124                 return substr( $value, 8 );
 125         }
 126
 127         /**
 128          * Create a new top-level frame for expansion of a page
 129          *
 130          * @return PPFrame
 131          */
 132         abstract public function newFrame();
 133
 134         /**
 135          * Create a new custom frame for programmatic use of parameter replacement
 136          * as used in some extensions.
 137          *
 138          * @param array $args
 139          *
 140          * @return PPFrame
 141          */
 142         abstract public function newCustomFrame( $args );
 143
 144         /**
 145          * Create a new custom node for programmatic use of parameter replacement
 146          * as used in some extensions.
 147          *
 148          * @param array $values
 149          */
 150         abstract public function newPartNodeArray( $values );
 151
 152         /**
 153          * Preprocess text to a PPNode
 154          *
 155          * @param string $text
 156          * @param int $flags
 157          *
 158          * @return PPNode
 159          */
 160         abstract public function preprocessToObj( $text, $flags = 0 );
 161 }
 162
 163 /**
 164  * @ingroup Parser
 165  */
 166 interface PPFrame {
 167         const NO_ARGS = 1;
 168         const NO_TEMPLATES = 2;
 169         const STRIP_COMMENTS = 4;
 170         const NO_IGNORE = 8;
 171         const RECOVER_COMMENTS = 16;
 172         const NO_TAGS = 32;
 173
 174         const RECOVER_ORIG = 59; // = 1|2|8|16|32 no constant expression support in PHP yet
 175
 176         /** This constant exists when $indexOffset is supported in newChild() */
 177         const SUPPORTS_INDEX_OFFSET = 1;
 178
 179         /**
 180          * Create a child frame
 181          *
 182          * @param array|bool $args
 183          * @param bool|Title $title
 184          * @param int $indexOffset A number subtracted from the index attributes of the arguments
 185          *
 186          * @return PPFrame
 187          */
 188         public function newChild( $args = false, $title = false, $indexOffset = 0 );
 189
 190         /**
 191          * Expand a document tree node, caching the result on its parent with the given key
 192          * @param string|int $key
 193          * @param string|PPNode $root
 194          * @param int $flags
 195          * @return string
 196          */
 197         public function cachedExpand( $key, $root, $flags = 0 );
 198
 199         /**
 200          * Expand a document tree node
 201          * @param string|PPNode $root
 202          * @param int $flags
 203          * @return string
 204          */
 205         public function expand( $root, $flags = 0 );
 206
 207         /**
 208          * Implode with flags for expand()
 209          * @param string $sep
 210          * @param int $flags
 211          * @param string|PPNode $args,...
 212          * @return string
 213          */
 214         public function implodeWithFlags( $sep, $flags /*, ... */ );
 215
 216         /**
 217          * Implode with no flags specified
 218          * @param string $sep
 219          * @param string|PPNode $args,...
 220          * @return string
 221          */
 222         public function implode( $sep /*, ... */ );
 223
 224         /**
 225          * Makes an object that, when expand()ed, will be the same as one obtained
 226          * with implode()
 227          * @param string $sep
 228          * @param string|PPNode $args,...
 229          * @return PPNode
 230          */
 231         public function virtualImplode( $sep /*, ... */ );
 232
 233         /**
 234          * Virtual implode with brackets
 235          * @param string $start
 236          * @param string $sep
 237          * @param string $end
 238          * @param string|PPNode $args,...
 239          * @return PPNode
 240          */
 241         public function virtualBracketedImplode( $start, $sep, $end /*, ... */ );
 242
 243         /**
 244          * Returns true if there are no arguments in this frame
 245          *
 246          * @return bool
 247          */
 248         public function isEmpty();
 249
 250         /**
 251          * Returns all arguments of this frame
 252          * @return array
 253          */
 254         public function getArguments();
 255
 256         /**
 257          * Returns all numbered arguments of this frame
 258          * @return array
 259          */
 260         public function getNumberedArguments();
 261
 262         /**
 263          * Returns all named arguments of this frame
 264          * @return array
 265          */
 266         public function getNamedArguments();
 267
 268         /**
 269          * Get an argument to this frame by name
 270          * @param int|string $name
 271          * @return string|bool
 272          */
 273         public function getArgument( $name );
 274
 275         /**
 276          * Returns true if the infinite loop check is OK, false if a loop is detected
 277          *
 278          * @param Title $title
 279          * @return bool
 280          */
 281         public function loopCheck( $title );
 282
 283         /**
 284          * Return true if the frame is a template frame
 285          * @return bool
 286          */
 287         public function isTemplate();
 288
 289         /**
 290          * Set the "volatile" flag.
 291          *
 292          * Note that this is somewhat of a "hack" in order to make extensions
 293          * with side effects (such as Cite) work with the PHP parser. New
 294          * extensions should be written in a way that they do not need this
 295          * function, because other parsers (such as Parsoid) are not guaranteed
 296          * to respect it, and it may be removed in the future.
 297          *
 298          * @param bool $flag
 299          */
 300         public function setVolatile( $flag = true );
 301
 302         /**
 303          * Get the "volatile" flag.
 304          *
 305          * Callers should avoid caching the result of an expansion if it has the
 306          * volatile flag set.
 307          *
 308          * @see self::setVolatile()
 309          * @return bool
 310          */
 311         public function isVolatile();
 312
 313         /**
 314          * Get the TTL of the frame's output.
 315          *
 316          * This is the maximum amount of time, in seconds, that this frame's
 317          * output should be cached for. A value of null indicates that no
 318          * maximum has been specified.
 319          *
 320          * Note that this TTL only applies to caching frames as parts of pages.
 321          * It is not relevant to caching the entire rendered output of a page.
 322          *
 323          * @return int|null
 324          */
 325         public function getTTL();
 326
 327         /**
 328          * Set the TTL of the output of this frame and all of its ancestors.
 329          * Has no effect if the new TTL is greater than the one already set.
 330          * Note that it is the caller's responsibility to change the cache
 331          * expiry of the page as a whole, if such behavior is desired.
 332          *
 333          * @see self::getTTL()
 334          * @param int $ttl
 335          */
 336         public function setTTL( $ttl );
 337
 338         /**
 339          * Get a title of frame
 340          *
 341          * @return Title
 342          */
 343         public function getTitle();
 344 }
 345
 346 /**
 347  * There are three types of nodes:
 348  *     * Tree nodes, which have a name and contain other nodes as children
 349  *     * Array nodes, which also contain other nodes but aren't considered part of a tree
 350  *     * Leaf nodes, which contain the actual data
 351  *
 352  * This interface provides access to the tree structure and to the contents of array nodes,
 353  * but it does not provide access to the internal structure of leaf nodes. Access to leaf
 354  * data is provided via two means:
 355  *     * PPFrame::expand(), which provides expanded text
 356  *     * The PPNode::split*() functions, which provide metadata about certain types of tree node
 357  * @ingroup Parser
 358  */
 359 interface PPNode {
 360         /**
 361          * Get an array-type node containing the children of this node.
 362          * Returns false if this is not a tree node.
 363          * @return PPNode
 364          */
 365         public function getChildren();
 366
 367         /**
 368          * Get the first child of a tree node. False if there isn't one.
 369          *
 370          * @return PPNode
 371          */
 372         public function getFirstChild();
 373
 374         /**
 375          * Get the next sibling of any node. False if there isn't one
 376          * @return PPNode
 377          */
 378         public function getNextSibling();
 379
 380         /**
 381          * Get all children of this tree node which have a given name.
 382          * Returns an array-type node, or false if this is not a tree node.
 383          * @param string $type
 384          * @return bool|PPNode
 385          */
 386         public function getChildrenOfType( $type );
 387
 388         /**
 389          * Returns the length of the array, or false if this is not an array-type node
 390          */
 391         public function getLength();
 392
 393         /**
 394          * Returns an item of an array-type node
 395          * @param int $i
 396          * @return bool|PPNode
 397          */
 398         public function item( $i );
 399
 400         /**
 401          * Get the name of this node. The following names are defined here:
 402          *
 403          *    h             A heading node.
 404          *    template      A double-brace node.
 405          *    tplarg        A triple-brace node.
 406          *    title         The first argument to a template or tplarg node.
 407          *    part          Subsequent arguments to a template or tplarg node.
 408          *    #nodelist     An array-type node
 409          *
 410          * The subclass may define various other names for tree and leaf nodes.
 411          * @return string
 412          */
 413         public function getName();
 414
 415         /**
 416          * Split a "<part>" node into an associative array containing:
 417          *    name          PPNode name
 418          *    index         String index
 419          *    value         PPNode value
 420          * @return array
 421          */
 422         public function splitArg();
 423
 424         /**
 425          * Split an "<ext>" node into an associative array containing name, attr, inner and close
 426          * All values in the resulting array are PPNodes. Inner and close are optional.
 427          * @return array
 428          */
 429         public function splitExt();
 430
 431         /**
 432          * Split an "<h>" node
 433          * @return array
 434          */
 435         public function splitHeading();
 436 }