maintenance/backupPrefetch.inc

   1 <?php
   2 /**
   3  * Helper class for the --prefetch option of dumpTextPass.php
   4  *
   5  * Copyright © 2005 Brion Vibber <brion@pobox.com>
   6  * https://www.mediawiki.org/
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Maintenance
  25  */
  26
  27 /**
  28  * Readahead helper for making large MediaWiki data dumps;
  29  * reads in a previous XML dump to sequentially prefetch text
  30  * records already normalized and decompressed.
  31  *
  32  * This can save load on the external database servers, hopefully.
  33  *
  34  * Assumes that dumps will be recorded in the canonical order:
  35  * - ascending by page_id
  36  * - ascending by rev_id within each page
  37  * - text contents are immutable and should not change once
  38  *   recorded, so the previous dump is a reliable source
  39  *
  40  * @ingroup Maintenance
  41  */
  42 class BaseDump {
  43         protected $reader = null;
  44         protected $atEnd = false;
  45         protected $atPageEnd = false;
  46         protected $lastPage = 0;
  47         protected $lastRev = 0;
  48         protected $infiles = null;
  49
  50         public function __construct( $infile ) {
  51                 $this->infiles = explode( ';', $infile );
  52                 $this->reader = new XMLReader();
  53                 $infile = array_shift( $this->infiles );
  54                 if ( defined( 'LIBXML_PARSEHUGE' ) ) {
  55                         $this->reader->open( $infile, null, LIBXML_PARSEHUGE );
  56                 }
  57                 else {
  58                         $this->reader->open( $infile );
  59                 }
  60         }
  61
  62         /**
  63          * Attempts to fetch the text of a particular page revision
  64          * from the dump stream. May return null if the page is
  65          * unavailable.
  66          *
  67          * @param int $page ID number of page to read
  68          * @param int $rev ID number of revision to read
  69          * @return string|null
  70          */
  71         function prefetch( $page, $rev ) {
  72                 $page = intval( $page );
  73                 $rev = intval( $rev );
  74                 while ( $this->lastPage < $page && !$this->atEnd ) {
  75                         $this->debug( "BaseDump::prefetch at page $this->lastPage, looking for $page" );
  76                         $this->nextPage();
  77                 }
  78                 if ( $this->lastPage > $page || $this->atEnd ) {
  79                         $this->debug( "BaseDump::prefetch already past page $page "
  80                                 . "looking for rev $rev  [$this->lastPage, $this->lastRev]" );
  81                         return null;
  82                 }
  83                 while ( $this->lastRev < $rev && !$this->atEnd && !$this->atPageEnd ) {
  84                         $this->debug( "BaseDump::prefetch at page $this->lastPage, rev $this->lastRev, "
  85                                 . "looking for $page, $rev" );
  86                         $this->nextRev();
  87                 }
  88                 if ( $this->lastRev == $rev && !$this->atEnd ) {
  89                         $this->debug( "BaseDump::prefetch hit on $page, $rev [$this->lastPage, $this->lastRev]" );
  90                         return $this->nextText();
  91                 } else {
  92                         $this->debug( "BaseDump::prefetch already past rev $rev on page $page "
  93                                 . "[$this->lastPage, $this->lastRev]" );
  94                         return null;
  95                 }
  96         }
  97
  98         function debug( $str ) {
  99                 wfDebug( $str . "\n" );
 100                 // global $dumper;
 101                 // $dumper->progress( $str );
 102         }
 103
 104         /**
 105          * @access private
 106          */
 107         function nextPage() {
 108                 if ( $this->skipTo( 'page', 'mediawiki' ) ) {
 109                         if ( $this->skipTo( 'id' ) ) {
 110                                 $this->lastPage = intval( $this->nodeContents() );
 111                                 $this->lastRev = 0;
 112                                 $this->atPageEnd = false;
 113                         }
 114                 } else {
 115                         $this->close();
 116                         if ( count( $this->infiles ) ) {
 117                                 $infile = array_shift( $this->infiles );
 118                                 $this->reader->open( $infile );
 119                                 $this->atEnd = false;
 120                         }
 121                 }
 122         }
 123
 124         /**
 125          * @access private
 126          */
 127         function nextRev() {
 128                 if ( $this->skipTo( 'revision' ) ) {
 129                         if ( $this->skipTo( 'id' ) ) {
 130                                 $this->lastRev = intval( $this->nodeContents() );
 131                         }
 132                 } else {
 133                         $this->atPageEnd = true;
 134                 }
 135         }
 136
 137         /**
 138          * @access private
 139          * @return string
 140          */
 141         function nextText() {
 142                 $this->skipTo( 'text' );
 143                 return strval( $this->nodeContents() );
 144         }
 145
 146         /**
 147          * @access private
 148          * @param string $name
 149          * @param string $parent
 150          * @return bool|null
 151          */
 152         function skipTo( $name, $parent = 'page' ) {
 153                 if ( $this->atEnd ) {
 154                         return false;
 155                 }
 156                 while ( $this->reader->read() ) {
 157                         if ( $this->reader->nodeType == XMLReader::ELEMENT &&
 158                                 $this->reader->name == $name ) {
 159                                 return true;
 160                         }
 161                         if ( $this->reader->nodeType == XMLReader::END_ELEMENT &&
 162                                 $this->reader->name == $parent ) {
 163                                 $this->debug( "BaseDump::skipTo found </$parent> searching for <$name>" );
 164                                 return false;
 165                         }
 166                 }
 167                 return $this->close();
 168         }
 169
 170         /**
 171          * Shouldn't something like this be built-in to XMLReader?
 172          * Fetches text contents of the current element, assuming
 173          * no sub-elements or such scary things.
 174          *
 175          * @return string
 176          * @access private
 177          */
 178         function nodeContents() {
 179                 if ( $this->atEnd ) {
 180                         return null;
 181                 }
 182                 if ( $this->reader->isEmptyElement ) {
 183                         return "";
 184                 }
 185                 $buffer = "";
 186                 while ( $this->reader->read() ) {
 187                         switch ( $this->reader->nodeType ) {
 188                         case XMLReader::TEXT:
 189 //                      case XMLReader::WHITESPACE:
 190                         case XMLReader::SIGNIFICANT_WHITESPACE:
 191                                 $buffer .= $this->reader->value;
 192                                 break;
 193                         case XMLReader::END_ELEMENT:
 194                                 return $buffer;
 195                         }
 196                 }
 197                 return $this->close();
 198         }
 199
 200         /**
 201          * @access private
 202          * @return null
 203          */
 204         function close() {
 205                 $this->reader->close();
 206                 $this->atEnd = true;
 207                 return null;
 208         }
 209 }