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