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   3  * Find all rows in the categorylinks table whose collation is out-of-date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  31  * whose collation is out-of-date.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 = 100; // Number of rows to process in one batch
  37         const SYNC_INTERVAL = 20; // Wait for slaves after this many batches
  38
  39         public $sizeHistogram = [];
  40
  41         public function __construct() {
  42                 parent::__construct();
  43
  44                 global $wgCategoryCollation;
  45                 $this->addDescription( <<<TEXT
  46 This script will find all rows in the categorylinks table whose collation is
  47 out-of-date (cl_collation != '$wgCategoryCollation') and repopulate cl_sortkey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
  53                 $this->addOption( 'force', 'Run on all rows, even if the collation is ' .
  54                         'supposed to be up-to-date.' );
  55                 $this->addOption( 'previous-collation', 'Set the previous value of ' .
  56                         '$wgCategoryCollation here to speed up this script, especially if your ' .
  57                         'categorylinks table is large. This will only update rows with that ' .
  58                         'collation, though, so it may miss out-of-date rows with a different, ' .
  59                         'even older collation.', false, true );
  60                 $this->addOption( 'target-collation', 'Set this to the new collation type to ' .
  61                         'use instead of $wgCategoryCollation. Usually you should not use this, ' .
  62                         'you should just update $wgCategoryCollation in LocalSettings.php.',
  63                         false, true );
  64                 $this->addOption( 'dry-run', 'Don\'t actually change the collations, just ' .
  65                         'compile statistics.' );
  66                 $this->addOption( 'verbose-stats', 'Show more statistics.' );
  67         }
  68
  69         public function execute() {
  70                 global $wgCategoryCollation;
  71
  72                 $dbw = $this->getDB( DB_MASTER );
  73                 $force = $this->getOption( 'force' );
  74                 $dryRun = $this->getOption( 'dry-run' );
  75                 $verboseStats = $this->getOption( 'verbose-stats' );
  76                 if ( $this->hasOption( 'target-collation' ) ) {
  77                         $collationName = $this->getOption( 'target-collation' );
  78                         $collation = Collation::factory( $collationName );
  79                 } else {
  80                         $collationName = $wgCategoryCollation;
  81                         $collation = Collation::singleton();
  82                 }
  83
  84                 // Collation sanity check: in some cases the constructor will work,
  85                 // but this will raise an exception, breaking all category pages
  86                 $collation->getFirstLetter( 'MediaWiki' );
  87
  88                 // Locally at least, (my local is a rather old version of mysql)
  89                 // mysql seems to filesort if there is both an equality
  90                 // (but not for an inequality) condition on cl_collation in the
  91                 // WHERE and it is also the first item in the ORDER BY.
  92                 if ( $this->hasOption( 'previous-collation' ) ) {
  93                         $orderBy = 'cl_to, cl_type, cl_from';
  94                 } else {
  95                         $orderBy = 'cl_collation, cl_to, cl_type, cl_from';
  96                 }
  97                 $options = [
  98                         'LIMIT' => self::BATCH_SIZE,
  99                         'ORDER BY' => $orderBy,
 100                 ];
 101
 102                 if ( $force || $dryRun ) {
 103                         $collationConds = [];
 104                 } else {
 105                         if ( $this->hasOption( 'previous-collation' ) ) {
 106                                 $collationConds['cl_collation'] = $this->getOption( 'previous-collation' );
 107                         } else {
 108                                 $collationConds = [ 0 =>
 109                                         'cl_collation != ' . $dbw->addQuotes( $collationName )
 110                                 ];
 111                         }
 112
 113                         $count = $dbw->estimateRowCount(
 114                                 'categorylinks',
 115                                 '*',
 116                                 $collationConds,
 117                                 __METHOD__
 118                         );
 119                         // Improve estimate if feasible
 120                         if ( $count < 1000000 ) {
 121                                 $count = $dbw->selectField(
 122                                         'categorylinks',
 123                                         'COUNT(*)',
 124                                         $collationConds,
 125                                         __METHOD__
 126                                 );
 127                         }
 128                         if ( $count == 0 ) {
 129                                 $this->output( "Collations up-to-date.\n" );
 130
 131                                 return;
 132                         }
 133                         $this->output( "Fixing collation for $count rows.\n" );
 134                 }
 135                 $count = 0;
 136                 $batchCount = 0;
 137                 $batchConds = [];
 138                 do {
 139                         $this->output( "Selecting next " . self::BATCH_SIZE . " rows..." );
 140
 141                         // cl_type must be selected as a number for proper paging because
 142                         // enums suck.
 143                         if ( $dbw->getType() === 'mysql' ) {
 144                                 $clType = 'cl_type+0 AS "cl_type_numeric"';
 145                         } else {
 146                                 $clType = 'cl_type';
 147                         }
 148                         $res = $dbw->select(
 149                                 [ 'categorylinks', 'page' ],
 150                                 [ 'cl_from', 'cl_to', 'cl_sortkey_prefix', 'cl_collation',
 151                                         'cl_sortkey', $clType,
 152                                         'page_namespace', 'page_title'
 153                                 ],
 154                                 array_merge( $collationConds, $batchConds, [ 'cl_from = page_id' ] ),
 155                                 __METHOD__,
 156                                 $options
 157                         );
 158                         $this->output( " processing..." );
 159
 160                         if ( !$dryRun ) {
 161                                 $this->beginTransaction( $dbw, __METHOD__ );
 162                         }
 163                         foreach ( $res as $row ) {
 164                                 $title = Title::newFromRow( $row );
 165                                 if ( !$row->cl_collation ) {
 166                                         # This is an old-style row, so the sortkey needs to be
 167                                         # converted.
 168                                         if ( $row->cl_sortkey == $title->getText()
 169                                                 || $row->cl_sortkey == $title->getPrefixedText()
 170                                         ) {
 171                                                 $prefix = '';
 172                                         } else {
 173                                                 # Custom sortkey, use it as a prefix
 174                                                 $prefix = $row->cl_sortkey;
 175                                         }
 176                                 } else {
 177                                         $prefix = $row->cl_sortkey_prefix;
 178                                 }
 179                                 # cl_type will be wrong for lots of pages if cl_collation is 0,
 180                                 # so let's update it while we're here.
 181                                 if ( $title->getNamespace() == NS_CATEGORY ) {
 182                                         $type = 'subcat';
 183                                 } elseif ( $title->getNamespace() == NS_FILE ) {
 184                                         $type = 'file';
 185                                 } else {
 186                                         $type = 'page';
 187                                 }
 188                                 $newSortKey = $collation->getSortKey(
 189                                         $title->getCategorySortkey( $prefix ) );
 190                                 if ( $verboseStats ) {
 191                                         $this->updateSortKeySizeHistogram( $newSortKey );
 192                                 }
 193
 194                                 if ( !$dryRun ) {
 195                                         $dbw->update(
 196                                                 'categorylinks',
 197                                                 [
 198                                                         'cl_sortkey' => $newSortKey,
 199                                                         'cl_sortkey_prefix' => $prefix,
 200                                                         'cl_collation' => $collationName,
 201                                                         'cl_type' => $type,
 202                                                         'cl_timestamp = cl_timestamp',
 203                                                 ],
 204                                                 [ 'cl_from' => $row->cl_from, 'cl_to' => $row->cl_to ],
 205                                                 __METHOD__
 206                                         );
 207                                 }
 208                                 if ( $row ) {
 209                                         $batchConds = [ $this->getBatchCondition( $row, $dbw ) ];
 210                                 }
 211                         }
 212                         if ( !$dryRun ) {
 213                                 $this->commitTransaction( $dbw, __METHOD__ );
 214                         }
 215
 216                         $count += $res->numRows();
 217                         $this->output( "$count done.\n" );
 218
 219                         if ( !$dryRun && ++$batchCount % self::SYNC_INTERVAL == 0 ) {
 220                                 $this->output( "Waiting for slaves ... " );
 221                                 wfWaitForSlaves();
 222                                 $this->output( "done\n" );
 223                         }
 224                 } while ( $res->numRows() == self::BATCH_SIZE );
 225
 226                 $this->output( "$count rows processed\n" );
 227
 228                 if ( $verboseStats ) {
 229                         $this->output( "\n" );
 230                         $this->showSortKeySizeHistogram();
 231                 }
 232         }
 233
 234         /**
 235          * Return an SQL expression selecting rows which sort above the given row,
 236          * assuming an ordering of cl_collation, cl_to, cl_type, cl_from
 237          * @param stdClass $row
 238          * @param DatabaseBase $dbw
 239          * @return string
 240          */
 241         function getBatchCondition( $row, $dbw ) {
 242                 if ( $this->hasOption( 'previous-collation' ) ) {
 243                         $fields = [ 'cl_to', 'cl_type', 'cl_from' ];
 244                 } else {
 245                         $fields = [ 'cl_collation', 'cl_to', 'cl_type', 'cl_from' ];
 246                 }
 247                 $first = true;
 248                 $cond = false;
 249                 $prefix = false;
 250                 foreach ( $fields as $field ) {
 251                         if ( $dbw->getType() === 'mysql' && $field === 'cl_type' ) {
 252                                 // Range conditions with enums are weird in mysql
 253                                 // This must be a numeric literal, or it won't work.
 254                                 $encValue = intval( $row->cl_type_numeric );
 255                         } else {
 256                                 $encValue = $dbw->addQuotes( $row->$field );
 257                         }
 258                         $inequality = "$field > $encValue";
 259                         $equality = "$field = $encValue";
 260                         if ( $first ) {
 261                                 $cond = $inequality;
 262                                 $prefix = $equality;
 263                                 $first = false;
 264                         } else {
 265                                 $cond .= " OR ($prefix AND $inequality)";
 266                                 $prefix .= " AND $equality";
 267                         }
 268                 }
 269
 270                 return $cond;
 271         }
 272
 273         function updateSortKeySizeHistogram( $key ) {
 274                 $length = strlen( $key );
 275                 if ( !isset( $this->sizeHistogram[$length] ) ) {
 276                         $this->sizeHistogram[$length] = 0;
 277                 }
 278                 $this->sizeHistogram[$length]++;
 279         }
 280
 281         function showSortKeySizeHistogram() {
 282                 $maxLength = max( array_keys( $this->sizeHistogram ) );
 283                 if ( $maxLength == 0 ) {
 284                         return;
 285                 }
 286                 $numBins = 20;
 287                 $coarseHistogram = array_fill( 0, $numBins, 0 );
 288                 $coarseBoundaries = [];
 289                 $boundary = 0;
 290                 for ( $i = 0; $i < $numBins - 1; $i++ ) {
 291                         $boundary += $maxLength / $numBins;
 292                         $coarseBoundaries[$i] = round( $boundary );
 293                 }
 294                 $coarseBoundaries[$numBins - 1] = $maxLength + 1;
 295                 $raw = '';
 296                 for ( $i = 0; $i <= $maxLength; $i++ ) {
 297                         if ( $raw !== '' ) {
 298                                 $raw .= ', ';
 299                         }
 300                         if ( !isset( $this->sizeHistogram[$i] ) ) {
 301                                 $val = 0;
 302                         } else {
 303                                 $val = $this->sizeHistogram[$i];
 304                         }
 305                         for ( $coarseIndex = 0; $coarseIndex < $numBins - 1; $coarseIndex++ ) {
 306                                 if ( $coarseBoundaries[$coarseIndex] > $i ) {
 307                                         $coarseHistogram[$coarseIndex] += $val;
 308                                         break;
 309                                 }
 310                         }
 311                         if ( $coarseIndex == $numBins - 1 ) {
 312                                 $coarseHistogram[$coarseIndex] += $val;
 313                         }
 314                         $raw .= $val;
 315                 }
 316
 317                 $this->output( "Sort key size histogram\nRaw data: $raw\n\n" );
 318
 319                 $maxBinVal = max( $coarseHistogram );
 320                 $scale = 60 / $maxBinVal;
 321                 $prevBoundary = 0;
 322                 for ( $coarseIndex = 0; $coarseIndex < $numBins; $coarseIndex++ ) {
 323                         if ( !isset( $coarseHistogram[$coarseIndex] ) ) {
 324                                 $val = 0;
 325                         } else {
 326                                 $val = $coarseHistogram[$coarseIndex];
 327                         }
 328                         $boundary = $coarseBoundaries[$coarseIndex];
 329                         $this->output( sprintf( "%-10s %-10d |%s\n",
 330                                 $prevBoundary . '-' . ( $boundary - 1 ) . ': ',
 331                                 $val,
 332                                 str_repeat( '*', $scale * $val ) ) );
 333                         $prevBoundary = $boundary;
 334                 }
 335         }
 336 }
 337
 338 $maintClass = "UpdateCollation";
 339 require_once RUN_MAINTENANCE_IF_MAIN;