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