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   3  * Find all rows in the categorylinks table whose collation is out-of-date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  34  * whose collation is out-of-date.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  50 out-of-date (cl_collation != '$wgCategoryCollation') and repopulate cl_sortkey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