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   3  * Find all rows in the categorylinks table whose collation is out-of-date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 use MediaWiki\MediaWikiServices;
  30 use Wikimedia\Rdbms\IDatabase;
  31
  32 /**
  33  * Maintenance script that will find all rows in the categorylinks table
  34  * whose collation is out-of-date.
  35  *
  36  * @ingroup Maintenance
  37  */
  38 class UpdateCollation extends Maintenance {
  39         const BATCH_SIZE = 100; // Number of rows to process in one batch
  40         const SYNC_INTERVAL = 5; // Wait for replica DBs after this many batches
  41
  42         public $sizeHistogram = [];
  43
  44         public function __construct() {
  45                 parent::__construct();
  46
  47                 $categoryCollation = $this->getConfig()->get( 'CategoryCollation' );
  48                 $this->addDescription( <<<TEXT
  49 This script will find all rows in the categorylinks table whose collation is
  50 out-of-date (cl_collation != '$categoryCollation') and repopulate cl_sortkey
  51 using the page title and cl_sortkey_prefix.  If all collations are
  52 up-to-date, it will do nothing.
  53 TEXT
  54                 );
  55
  56                 $this->addOption( 'force', 'Run on all rows, even if the collation is ' .
  57                         'supposed to be up-to-date.', false, false, 'f' );
  58                 $this->addOption( 'previous-collation', 'Set the previous value of ' .
  59                         '$wgCategoryCollation here to speed up this script, especially if your ' .
  60                         'categorylinks table is large. This will only update rows with that ' .
  61                         'collation, though, so it may miss out-of-date rows with a different, ' .
  62                         'even older collation.', false, true );
  63                 $this->addOption( 'target-collation', 'Set this to the new collation type to ' .
  64                         'use instead of $wgCategoryCollation. Usually you should not use this, ' .
  65                         'you should just update $wgCategoryCollation in LocalSettings.php.',
  66                         false, true );
  67                 $this->addOption( 'dry-run', 'Don\'t actually change the collations, just ' .
  68                         'compile statistics.' );
  69                 $this->addOption( 'verbose-stats', 'Show more statistics.' );
  70         }
  71
  72         public function execute() {
  73                 $dbw = $this->getDB( DB_MASTER );
  74                 $dbr = $this->getDB( DB_REPLICA );
  75                 $force = $this->getOption( 'force' );
  76                 $dryRun = $this->getOption( 'dry-run' );
  77                 $verboseStats = $this->getOption( 'verbose-stats' );
  78                 if ( $this->hasOption( 'target-collation' ) ) {
  79                         $collationName = $this->getOption( 'target-collation' );
  80                         $collation = Collation::factory( $collationName );
  81                 } else {
  82                         $collationName = $this->getConfig()->get( 'CategoryCollation' );
  83                         $collation = Collation::singleton();
  84                 }
  85
  86                 // Collation sanity check: in some cases the constructor will work,
  87                 // but this will raise an exception, breaking all category pages
  88                 $collation->getFirstLetter( 'MediaWiki' );
  89
  90                 // Locally at least, (my local is a rather old version of mysql)
  91                 // mysql seems to filesort if there is both an equality
  92                 // (but not for an inequality) condition on cl_collation in the
  93                 // WHERE and it is also the first item in the ORDER BY.
  94                 if ( $this->hasOption( 'previous-collation' ) ) {
  95                         $orderBy = 'cl_to, cl_type, cl_from';
  96                 } else {
  97                         $orderBy = 'cl_collation, cl_to, cl_type, cl_from';
  98                 }
  99                 $options = [
 100                         'LIMIT' => self::BATCH_SIZE,
 101                         'ORDER BY' => $orderBy,
 102                         'STRAIGHT_JOIN' // per T58041
 103                 ];
 104
 105                 $collationConds = [];
 106                 if ( !$force ) {
 107                         if ( $this->hasOption( 'previous-collation' ) ) {
 108                                 $collationConds['cl_collation'] = $this->getOption( 'previous-collation' );
 109                         } else {
 110                                 $collationConds = [ 0 =>
 111                                         'cl_collation != ' . $dbw->addQuotes( $collationName )
 112                                 ];
 113                         }
 114
 115                         $count = $dbr->estimateRowCount(
 116                                 'categorylinks',
 117                                 '*',
 118                                 $collationConds,
 119                                 __METHOD__
 120                         );
 121                         // Improve estimate if feasible
 122                         if ( $count < 1000000 ) {
 123                                 $count = $dbr->selectField(
 124                                         'categorylinks',
 125                                         'COUNT(*)',
 126                                         $collationConds,
 127                                         __METHOD__
 128                                 );
 129                         }
 130                         if ( $count == 0 ) {
 131                                 $this->output( "Collations up-to-date.\n" );
 132
 133                                 return;
 134                         }
 135                         if ( $dryRun ) {
 136                                 $this->output( "$count rows would be updated.\n" );
 137                         } else {
 138                                 $this->output( "Fixing collation for $count rows.\n" );
 139                         }
 140                         wfWaitForSlaves();
 141                 }
 142                 $count = 0;
 143                 $batchConds = [];
 144                 do {
 145                         $this->output( "Selecting next " . self::BATCH_SIZE . " rows..." );
 146
 147                         // cl_type must be selected as a number for proper paging because
 148                         // enums suck.
 149                         if ( $dbw->getType() === 'mysql' ) {
 150                                 $clType = 'cl_type+0 AS "cl_type_numeric"';
 151                         } else {
 152                                 $clType = 'cl_type';
 153                         }
 154                         $res = $dbw->select(
 155                                 [ 'categorylinks', 'page' ],
 156                                 [ 'cl_from', 'cl_to', 'cl_sortkey_prefix', 'cl_collation',
 157                                         'cl_sortkey', $clType,
 158                                         'page_namespace', 'page_title'
 159                                 ],
 160                                 array_merge( $collationConds, $batchConds, [ 'cl_from = page_id' ] ),
 161                                 __METHOD__,
 162                                 $options
 163                         );
 164                         $this->output( " processing..." );
 165
 166                         if ( !$dryRun ) {
 167                                 $this->beginTransaction( $dbw, __METHOD__ );
 168                         }
 169                         foreach ( $res as $row ) {
 170                                 $title = Title::newFromRow( $row );
 171                                 if ( !$row->cl_collation ) {
 172                                         # This is an old-style row, so the sortkey needs to be
 173                                         # converted.
 174                                         if ( $row->cl_sortkey == $title->getText()
 175                                                 || $row->cl_sortkey == $title->getPrefixedText()
 176                                         ) {
 177                                                 $prefix = '';
 178                                         } else {
 179                                                 # Custom sortkey, use it as a prefix
 180                                                 $prefix = $row->cl_sortkey;
 181                                         }
 182                                 } else {
 183                                         $prefix = $row->cl_sortkey_prefix;
 184                                 }
 185                                 # cl_type will be wrong for lots of pages if cl_collation is 0,
 186                                 # so let's update it while we're here.
 187                                 $type = MediaWikiServices::getInstance()->getNamespaceInfo()->
 188                                         getCategoryLinkType( $title->getNamespace() );
 189                                 $newSortKey = $collation->getSortKey(
 190                                         $title->getCategorySortkey( $prefix ) );
 191                                 if ( $verboseStats ) {
 192                                         $this->updateSortKeySizeHistogram( $newSortKey );
 193                                 }
 194
 195                                 if ( $dryRun ) {
 196                                         // Add 1 to the count if the sortkey was changed. (Note that this doesn't count changes in
 197                                         // other fields, if any, those usually only happen when upgrading old MediaWikis.)
 198                                         $count += ( $row->cl_sortkey !== $newSortKey );
 199                                 } else {
 200                                         $dbw->update(
 201                                                 'categorylinks',
 202                                                 [
 203                                                         'cl_sortkey' => $newSortKey,
 204                                                         'cl_sortkey_prefix' => $prefix,
 205                                                         'cl_collation' => $collationName,
 206                                                         'cl_type' => $type,
 207                                                         'cl_timestamp = cl_timestamp',
 208                                                 ],
 209                                                 [ 'cl_from' => $row->cl_from, 'cl_to' => $row->cl_to ],
 210                                                 __METHOD__
 211                                         );
 212                                         $count++;
 213                                 }
 214                                 if ( $row ) {
 215                                         $batchConds = [ $this->getBatchCondition( $row, $dbw ) ];
 216                                 }
 217                         }
 218                         if ( !$dryRun ) {
 219                                 $this->commitTransaction( $dbw, __METHOD__ );
 220                         }
 221
 222                         if ( $dryRun ) {
 223                                 $this->output( "$count rows would be updated so far.\n" );
 224                         } else {
 225                                 $this->output( "$count done.\n" );
 226                         }
 227                 } while ( $res->numRows() == self::BATCH_SIZE );
 228
 229                 if ( !$dryRun ) {
 230                         $this->output( "$count rows processed\n" );
 231                 }
 232
 233                 if ( $verboseStats ) {
 234                         $this->output( "\n" );
 235                         $this->showSortKeySizeHistogram();
 236                 }
 237         }
 238
 239         /**
 240          * Return an SQL expression selecting rows which sort above the given row,
 241          * assuming an ordering of cl_collation, cl_to, cl_type, cl_from
 242          * @param stdClass $row
 243          * @param IDatabase $dbw
 244          * @return string
 245          */
 246         function getBatchCondition( $row, $dbw ) {
 247                 if ( $this->hasOption( 'previous-collation' ) ) {
 248                         $fields = [ 'cl_to', 'cl_type', 'cl_from' ];
 249                 } else {
 250                         $fields = [ 'cl_collation', 'cl_to', 'cl_type', 'cl_from' ];
 251                 }
 252                 $first = true;
 253                 $cond = false;
 254                 $prefix = false;
 255                 foreach ( $fields as $field ) {
 256                         if ( $dbw->getType() === 'mysql' && $field === 'cl_type' ) {
 257                                 // Range conditions with enums are weird in mysql
 258                                 // This must be a numeric literal, or it won't work.
 259                                 $encValue = intval( $row->cl_type_numeric );
 260                         } else {
 261                                 $encValue = $dbw->addQuotes( $row->$field );
 262                         }
 263                         $inequality = "$field > $encValue";
 264                         $equality = "$field = $encValue";
 265                         if ( $first ) {
 266                                 $cond = $inequality;
 267                                 $prefix = $equality;
 268                                 $first = false;
 269                         } else {
 270                                 $cond .= " OR ($prefix AND $inequality)";
 271                                 $prefix .= " AND $equality";
 272                         }
 273                 }
 274
 275                 return $cond;
 276         }
 277
 278         function updateSortKeySizeHistogram( $key ) {
 279                 $length = strlen( $key );
 280                 if ( !isset( $this->sizeHistogram[$length] ) ) {
 281                         $this->sizeHistogram[$length] = 0;
 282                 }
 283                 $this->sizeHistogram[$length]++;
 284         }
 285
 286         function showSortKeySizeHistogram() {
 287                 $maxLength = max( array_keys( $this->sizeHistogram ) );
 288                 if ( $maxLength == 0 ) {
 289                         return;
 290                 }
 291                 $numBins = 20;
 292                 $coarseHistogram = array_fill( 0, $numBins, 0 );
 293                 $coarseBoundaries = [];
 294                 $boundary = 0;
 295                 for ( $i = 0; $i < $numBins - 1; $i++ ) {
 296                         $boundary += $maxLength / $numBins;
 297                         $coarseBoundaries[$i] = round( $boundary );
 298                 }
 299                 $coarseBoundaries[$numBins - 1] = $maxLength + 1;
 300                 $raw = '';
 301                 for ( $i = 0; $i <= $maxLength; $i++ ) {
 302                         if ( $raw !== '' ) {
 303                                 $raw .= ', ';
 304                         }
 305                         $val = $this->sizeHistogram[$i] ?? 0;
 306                         for ( $coarseIndex = 0; $coarseIndex < $numBins - 1; $coarseIndex++ ) {
 307                                 if ( $coarseBoundaries[$coarseIndex] > $i ) {
 308                                         $coarseHistogram[$coarseIndex] += $val;
 309                                         break;
 310                                 }
 311                         }
 312                         if ( $coarseIndex == $numBins - 1 ) {
 313                                 $coarseHistogram[$coarseIndex] += $val;
 314                         }
 315                         $raw .= $val;
 316                 }
 317
 318                 $this->output( "Sort key size histogram\nRaw data: $raw\n\n" );
 319
 320                 $maxBinVal = max( $coarseHistogram );
 321                 $scale = 60 / $maxBinVal;
 322                 $prevBoundary = 0;
 323                 for ( $coarseIndex = 0; $coarseIndex < $numBins; $coarseIndex++ ) {
 324                         $val = $coarseHistogram[$coarseIndex] ?? 0;
 325                         $boundary = $coarseBoundaries[$coarseIndex];
 326                         $this->output( sprintf( "%-10s %-10d |%s\n",
 327                                 $prevBoundary . '-' . ( $boundary - 1 ) . ': ',
 328                                 $val,
 329                                 str_repeat( '*', $scale * $val ) ) );
 330                         $prevBoundary = $boundary;
 331                 }
 332         }
 333 }
 334
 335 $maintClass = UpdateCollation::class;
 336 require_once RUN_MAINTENANCE_IF_MAIN;