maintenance/updateCollation.php

   1 <?php
   2 /**
   3  * Find all rows in the categorylinks table whose collation is out-of-date
   4  * (cl_collation != $wgCategoryCollation) and repopulate cl_sortkey
   5  * using the page title and cl_sortkey_prefix.
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License along
  18  * with this program; if not, write to the Free Software Foundation, Inc.,
  19  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @file
  23  * @ingroup Maintenance
  24  * @author Aryeh Gregor (Simetrical)
  25  */
  26
  27 require_once __DIR__ . '/Maintenance.php';
  28
  29 /**
  30  * Maintenance script that will find all rows in the categorylinks table
  31  * whose collation is out-of-date.
  32  *
  33  * @ingroup Maintenance
  34  */
  35 class UpdateCollation extends Maintenance {
  36         const BATCH_SIZE = 10000; // Number of rows to process in one batch
  37         const SYNC_INTERVAL = 20; // Wait for slaves after this many batches
  38
  39         public $sizeHistogram = array();
  40
  41         public function __construct() {
  42                 parent::__construct();
  43
  44                 global $wgCategoryCollation;
  45                 $this->mDescription = <<<TEXT
  46 This script will find all rows in the categorylinks table whose collation is
  47 out-of-date (cl_collation != '$wgCategoryCollation') and repopulate cl_sortkey
  48 using the page title and cl_sortkey_prefix.  If all collations are
  49 up-to-date, it will do nothing.
  50 TEXT;
  51
  52                 $this->addOption( 'force', 'Run on all rows, even if the collation is ' .
  53                         'supposed to be up-to-date.' );
  54                 $this->addOption( 'previous-collation', 'Set the previous value of ' .
  55                         '$wgCategoryCollation here to speed up this script, especially if your ' .
  56                         'categorylinks table is large. This will only update rows with that ' .
  57                         'collation, though, so it may miss out-of-date rows with a different, ' .
  58                         'even older collation.', false, true );
  59                 $this->addOption( 'target-collation', 'Set this to the new collation type to ' .
  60                         'use instead of $wgCategoryCollation. Usually you should not use this, ' .
  61                         'you should just update $wgCategoryCollation in LocalSettings.php.',
  62                         false, true );
  63                 $this->addOption( 'dry-run', 'Don\'t actually change the collations, just ' .
  64                         'compile statistics.' );
  65                 $this->addOption( 'verbose-stats', 'Show more statistics.' );
  66         }
  67
  68         public function execute() {
  69                 global $wgCategoryCollation;
  70
  71                 $dbw = $this->getDB( DB_MASTER );
  72                 $force = $this->getOption( 'force' );
  73                 $dryRun = $this->getOption( 'dry-run' );
  74                 $verboseStats = $this->getOption( 'verbose-stats' );
  75                 if ( $this->hasOption( 'target-collation' ) ) {
  76                         $collationName = $this->getOption( 'target-collation' );
  77                         $collation = Collation::factory( $collationName );
  78                 } else {
  79                         $collationName = $wgCategoryCollation;
  80                         $collation = Collation::singleton();
  81                 }
  82
  83                 // Collation sanity check: in some cases the constructor will work,
  84                 // but this will raise an exception, breaking all category pages
  85                 $collation->getFirstLetter( 'MediaWiki' );
  86
  87                 $options = array(
  88                         'LIMIT' => self::BATCH_SIZE,
  89                         'ORDER BY' => 'cl_to, cl_type, cl_from',
  90                         'STRAIGHT_JOIN',
  91                 );
  92
  93                 if ( $force || $dryRun ) {
  94                         $collationConds = array();
  95                 } else {
  96                         if ( $this->hasOption( 'previous-collation' ) ) {
  97                                 $collationConds['cl_collation'] = $this->getOption( 'previous-collation' );
  98                         } else {
  99                                 $collationConds = array( 0 =>
 100                                         'cl_collation != ' . $dbw->addQuotes( $collationName )
 101                                 );
 102                         }
 103
 104                         $count = $dbw->estimateRowCount(
 105                                 'categorylinks',
 106                                 '*',
 107                                 $collationConds,
 108                                 __METHOD__
 109                         );
 110                         // Improve estimate if feasible
 111                         if ( $count < 1000000 ) {
 112                                 $count = $dbw->selectField(
 113                                         'categorylinks',
 114                                         'COUNT(*)',
 115                                         $collationConds,
 116                                         __METHOD__
 117                                 );
 118                         }
 119                         if ( $count == 0 ) {
 120                                 $this->output( "Collations up-to-date.\n" );
 121
 122                                 return;
 123                         }
 124                         $this->output( "Fixing collation for $count rows.\n" );
 125                 }
 126
 127                 $count = 0;
 128                 $batchCount = 0;
 129                 $batchConds = array();
 130                 do {
 131                         $this->output( "Selecting next " . self::BATCH_SIZE . " rows..." );
 132                         $res = $dbw->select(
 133                                 array( 'categorylinks', 'page' ),
 134                                 array( 'cl_from', 'cl_to', 'cl_sortkey_prefix', 'cl_collation',
 135                                         'cl_sortkey', 'cl_type', 'page_namespace', 'page_title'
 136                                 ),
 137                                 array_merge( $collationConds, $batchConds, array( 'cl_from = page_id' ) ),
 138                                 __METHOD__,
 139                                 $options
 140                         );
 141                         $this->output( " processing..." );
 142
 143                         if ( !$dryRun ) {
 144                                 $dbw->begin( __METHOD__ );
 145                         }
 146                         foreach ( $res as $row ) {
 147                                 $title = Title::newFromRow( $row );
 148                                 if ( !$row->cl_collation ) {
 149                                         # This is an old-style row, so the sortkey needs to be
 150                                         # converted.
 151                                         if ( $row->cl_sortkey == $title->getText()
 152                                                 || $row->cl_sortkey == $title->getPrefixedText()
 153                                         ) {
 154                                                 $prefix = '';
 155                                         } else {
 156                                                 # Custom sortkey, use it as a prefix
 157                                                 $prefix = $row->cl_sortkey;
 158                                         }
 159                                 } else {
 160                                         $prefix = $row->cl_sortkey_prefix;
 161                                 }
 162                                 # cl_type will be wrong for lots of pages if cl_collation is 0,
 163                                 # so let's update it while we're here.
 164                                 if ( $title->getNamespace() == NS_CATEGORY ) {
 165                                         $type = 'subcat';
 166                                 } elseif ( $title->getNamespace() == NS_FILE ) {
 167                                         $type = 'file';
 168                                 } else {
 169                                         $type = 'page';
 170                                 }
 171                                 $newSortKey = $collation->getSortKey(
 172                                         $title->getCategorySortkey( $prefix ) );
 173                                 if ( $verboseStats ) {
 174                                         $this->updateSortKeySizeHistogram( $newSortKey );
 175                                 }
 176
 177                                 if ( !$dryRun ) {
 178                                         $dbw->update(
 179                                                 'categorylinks',
 180                                                 array(
 181                                                         'cl_sortkey' => $newSortKey,
 182                                                         'cl_sortkey_prefix' => $prefix,
 183                                                         'cl_collation' => $collationName,
 184                                                         'cl_type' => $type,
 185                                                         'cl_timestamp = cl_timestamp',
 186                                                 ),
 187                                                 array( 'cl_from' => $row->cl_from, 'cl_to' => $row->cl_to ),
 188                                                 __METHOD__
 189                                         );
 190                                 }
 191                                 if ( $row ) {
 192                                         $batchConds = array( $this->getBatchCondition( $row, $dbw ) );
 193                                 }
 194                         }
 195                         if ( !$dryRun ) {
 196                                 $dbw->commit( __METHOD__ );
 197                         }
 198
 199                         $count += $res->numRows();
 200                         $this->output( "$count done.\n" );
 201
 202                         if ( !$dryRun && ++$batchCount % self::SYNC_INTERVAL == 0 ) {
 203                                 $this->output( "Waiting for slaves ... " );
 204                                 wfWaitForSlaves();
 205                                 $this->output( "done\n" );
 206                         }
 207                 } while ( $res->numRows() == self::BATCH_SIZE );
 208
 209                 $this->output( "$count rows processed\n" );
 210
 211                 if ( $verboseStats ) {
 212                         $this->output( "\n" );
 213                         $this->showSortKeySizeHistogram();
 214                 }
 215         }
 216
 217         /**
 218          * Return an SQL expression selecting rows which sort above the given row,
 219          * assuming an ordering of cl_to, cl_type, cl_from
 220          * @param stdClass $row
 221          * @param DatabaseBase $dbw
 222          * @return string
 223          */
 224         function getBatchCondition( $row, $dbw ) {
 225                 $fields = array( 'cl_to', 'cl_type', 'cl_from' );
 226                 $first = true;
 227                 $cond = false;
 228                 $prefix = false;
 229                 foreach ( $fields as $field ) {
 230                         $encValue = $dbw->addQuotes( $row->$field );
 231                         $inequality = "$field > $encValue";
 232                         $equality = "$field = $encValue";
 233                         if ( $first ) {
 234                                 $cond = $inequality;
 235                                 $prefix = $equality;
 236                                 $first = false;
 237                         } else {
 238                                 $cond .= " OR ($prefix AND $inequality)";
 239                                 $prefix .= " AND $equality";
 240                         }
 241                 }
 242
 243                 return $cond;
 244         }
 245
 246         function updateSortKeySizeHistogram( $key ) {
 247                 $length = strlen( $key );
 248                 if ( !isset( $this->sizeHistogram[$length] ) ) {
 249                         $this->sizeHistogram[$length] = 0;
 250                 }
 251                 $this->sizeHistogram[$length]++;
 252         }
 253
 254         function showSortKeySizeHistogram() {
 255                 $maxLength = max( array_keys( $this->sizeHistogram ) );
 256                 if ( $maxLength == 0 ) {
 257                         return;
 258                 }
 259                 $numBins = 20;
 260                 $coarseHistogram = array_fill( 0, $numBins, 0 );
 261                 $coarseBoundaries = array();
 262                 $boundary = 0;
 263                 for ( $i = 0; $i < $numBins - 1; $i++ ) {
 264                         $boundary += $maxLength / $numBins;
 265                         $coarseBoundaries[$i] = round( $boundary );
 266                 }
 267                 $coarseBoundaries[$numBins - 1] = $maxLength + 1;
 268                 $raw = '';
 269                 for ( $i = 0; $i <= $maxLength; $i++ ) {
 270                         if ( $raw !== '' ) {
 271                                 $raw .= ', ';
 272                         }
 273                         if ( !isset( $this->sizeHistogram[$i] ) ) {
 274                                 $val = 0;
 275                         } else {
 276                                 $val = $this->sizeHistogram[$i];
 277                         }
 278                         for ( $coarseIndex = 0; $coarseIndex < $numBins - 1; $coarseIndex++ ) {
 279                                 if ( $coarseBoundaries[$coarseIndex] > $i ) {
 280                                         $coarseHistogram[$coarseIndex] += $val;
 281                                         break;
 282                                 }
 283                         }
 284                         if ( $coarseIndex == $numBins - 1 ) {
 285                                 $coarseHistogram[$coarseIndex] += $val;
 286                         }
 287                         $raw .= $val;
 288                 }
 289
 290                 $this->output( "Sort key size histogram\nRaw data: $raw\n\n" );
 291
 292                 $maxBinVal = max( $coarseHistogram );
 293                 $scale = 60 / $maxBinVal;
 294                 $prevBoundary = 0;
 295                 for ( $coarseIndex = 0; $coarseIndex < $numBins; $coarseIndex++ ) {
 296                         if ( !isset( $coarseHistogram[$coarseIndex] ) ) {
 297                                 $val = 0;
 298                         } else {
 299                                 $val = $coarseHistogram[$coarseIndex];
 300                         }
 301                         $boundary = $coarseBoundaries[$coarseIndex];
 302                         $this->output( sprintf( "%-10s %-10d |%s\n",
 303                                 $prevBoundary . '-' . ( $boundary - 1 ) . ': ',
 304                                 $val,
 305                                 str_repeat( '*', $scale * $val ) ) );
 306                         $prevBoundary = $boundary;
 307                 }
 308         }
 309 }
 310
 311 $maintClass = "UpdateCollation";
 312 require_once RUN_MAINTENANCE_IF_MAIN;