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   3  * Find all rows in the categorylinks table whose collation is out-of-date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  31  * whose collation is out-of-date.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  47 out-of-date (cl_collation != '$wgCategoryCollation') and repopulate cl_sortkey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.', false, false, 'f' );
  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                         'STRAIGHT_JOIN' // per T58041
 101                 ];
 102
 103                 if ( $force || $dryRun ) {
 104                         $collationConds = [];
 105                 } else {
 106                         if ( $this->hasOption( 'previous-collation' ) ) {
 107                                 $collationConds['cl_collation'] = $this->getOption( 'previous-collation' );
 108                         } else {
 109                                 $collationConds = [ 0 =>
 110                                         'cl_collation != ' . $dbw->addQuotes( $collationName )
 111                                 ];
 112                         }
 113
 114                         $count = $dbw->estimateRowCount(
 115                                 'categorylinks',
 116                                 '*',
 117                                 $collationConds,
 118                                 __METHOD__
 119                         );
 120                         // Improve estimate if feasible
 121                         if ( $count < 1000000 ) {
 122                                 $count = $dbw->selectField(
 123                                         'categorylinks',
 124                                         'COUNT(*)',
 125                                         $collationConds,
 126                                         __METHOD__
 127                                 );
 128                         }
 129                         if ( $count == 0 ) {
 130                                 $this->output( "Collations up-to-date.\n" );
 131
 132                                 return;
 133                         }
 134                         $this->output( "Fixing collation for $count rows.\n" );
 135                 }
 136                 $count = 0;
 137                 $batchCount = 0;
 138                 $batchConds = [];
 139                 do {
 140                         $this->output( "Selecting next " . self::BATCH_SIZE . " rows..." );
 141
 142                         // cl_type must be selected as a number for proper paging because
 143                         // enums suck.
 144                         if ( $dbw->getType() === 'mysql' ) {
 145                                 $clType = 'cl_type+0 AS "cl_type_numeric"';
 146                         } else {
 147                                 $clType = 'cl_type';
 148                         }
 149                         $res = $dbw->select(
 150                                 [ 'categorylinks', 'page' ],
 151                                 [ 'cl_from', 'cl_to', 'cl_sortkey_prefix', 'cl_collation',
 152                                         'cl_sortkey', $clType,
 153                                         'page_namespace', 'page_title'
 154                                 ],
 155                                 array_merge( $collationConds, $batchConds, [ 'cl_from = page_id' ] ),
 156                                 __METHOD__,
 157                                 $options
 158                         );
 159                         $this->output( " processing..." );
 160
 161                         if ( !$dryRun ) {
 162                                 $this->beginTransaction( $dbw, __METHOD__ );
 163                         }
 164                         foreach ( $res as $row ) {
 165                                 $title = Title::newFromRow( $row );
 166                                 if ( !$row->cl_collation ) {
 167                                         # This is an old-style row, so the sortkey needs to be
 168                                         # converted.
 169                                         if ( $row->cl_sortkey == $title->getText()
 170                                                 || $row->cl_sortkey == $title->getPrefixedText()
 171                                         ) {
 172                                                 $prefix = '';
 173                                         } else {
 174                                                 # Custom sortkey, use it as a prefix
 175                                                 $prefix = $row->cl_sortkey;
 176                                         }
 177                                 } else {
 178                                         $prefix = $row->cl_sortkey_prefix;
 179                                 }
 180                                 # cl_type will be wrong for lots of pages if cl_collation is 0,
 181                                 # so let's update it while we're here.
 182                                 if ( $title->getNamespace() == NS_CATEGORY ) {
 183                                         $type = 'subcat';
 184                                 } elseif ( $title->getNamespace() == NS_FILE ) {
 185                                         $type = 'file';
 186                                 } else {
 187                                         $type = 'page';
 188                                 }
 189                                 $newSortKey = $collation->getSortKey(
 190                                         $title->getCategorySortkey( $prefix ) );
 191                                 if ( $verboseStats ) {
 192                                         $this->updateSortKeySizeHistogram( $newSortKey );
 193                                 }
 194
 195                                 if ( !$dryRun ) {
 196                                         $dbw->update(
 197                                                 'categorylinks',
 198                                                 [
 199                                                         'cl_sortkey' => $newSortKey,
 200                                                         'cl_sortkey_prefix' => $prefix,
 201                                                         'cl_collation' => $collationName,
 202                                                         'cl_type' => $type,
 203                                                         'cl_timestamp = cl_timestamp',
 204                                                 ],
 205                                                 [ 'cl_from' => $row->cl_from, 'cl_to' => $row->cl_to ],
 206                                                 __METHOD__
 207                                         );
 208                                 }
 209                                 if ( $row ) {
 210                                         $batchConds = [ $this->getBatchCondition( $row, $dbw ) ];
 211                                 }
 212                         }
 213                         if ( !$dryRun ) {
 214                                 $this->commitTransaction( $dbw, __METHOD__ );
 215                         }
 216
 217                         $count += $res->numRows();
 218                         $this->output( "$count done.\n" );
 219
 220                         if ( !$dryRun && ++$batchCount % self::SYNC_INTERVAL == 0 ) {
 221                                 $this->output( "Waiting for slaves ... " );
 222                                 wfWaitForSlaves();
 223                                 $this->output( "done\n" );
 224                         }
 225                 } while ( $res->numRows() == self::BATCH_SIZE );
 226
 227                 $this->output( "$count rows processed\n" );
 228
 229                 if ( $verboseStats ) {
 230                         $this->output( "\n" );
 231                         $this->showSortKeySizeHistogram();
 232                 }
 233         }
 234
 235         /**
 236          * Return an SQL expression selecting rows which sort above the given row,
 237          * assuming an ordering of cl_collation, cl_to, cl_type, cl_from
 238          * @param stdClass $row
 239          * @param DatabaseBase $dbw
 240          * @return string
 241          */
 242         function getBatchCondition( $row, $dbw ) {
 243                 if ( $this->hasOption( 'previous-collation' ) ) {
 244                         $fields = [ 'cl_to', 'cl_type', 'cl_from' ];
 245                 } else {
 246                         $fields = [ 'cl_collation', 'cl_to', 'cl_type', 'cl_from' ];
 247                 }
 248                 $first = true;
 249                 $cond = false;
 250                 $prefix = false;
 251                 foreach ( $fields as $field ) {
 252                         if ( $dbw->getType() === 'mysql' && $field === 'cl_type' ) {
 253                                 // Range conditions with enums are weird in mysql
 254                                 // This must be a numeric literal, or it won't work.
 255                                 $encValue = intval( $row->cl_type_numeric );
 256                         } else {
 257                                 $encValue = $dbw->addQuotes( $row->$field );
 258                         }
 259                         $inequality = "$field > $encValue";
 260                         $equality = "$field = $encValue";
 261                         if ( $first ) {
 262                                 $cond = $inequality;
 263                                 $prefix = $equality;
 264                                 $first = false;
 265                         } else {
 266                                 $cond .= " OR ($prefix AND $inequality)";
 267                                 $prefix .= " AND $equality";
 268                         }
 269                 }
 270
 271                 return $cond;
 272         }
 273
 274         function updateSortKeySizeHistogram( $key ) {
 275                 $length = strlen( $key );
 276                 if ( !isset( $this->sizeHistogram[$length] ) ) {
 277                         $this->sizeHistogram[$length] = 0;
 278                 }
 279                 $this->sizeHistogram[$length]++;
 280         }
 281
 282         function showSortKeySizeHistogram() {
 283                 $maxLength = max( array_keys( $this->sizeHistogram ) );
 284                 if ( $maxLength == 0 ) {
 285                         return;
 286                 }
 287                 $numBins = 20;
 288                 $coarseHistogram = array_fill( 0, $numBins, 0 );
 289                 $coarseBoundaries = [];
 290                 $boundary = 0;
 291                 for ( $i = 0; $i < $numBins - 1; $i++ ) {
 292                         $boundary += $maxLength / $numBins;
 293                         $coarseBoundaries[$i] = round( $boundary );
 294                 }
 295                 $coarseBoundaries[$numBins - 1] = $maxLength + 1;
 296                 $raw = '';
 297                 for ( $i = 0; $i <= $maxLength; $i++ ) {
 298                         if ( $raw !== '' ) {
 299                                 $raw .= ', ';
 300                         }
 301                         if ( !isset( $this->sizeHistogram[$i] ) ) {
 302                                 $val = 0;
 303                         } else {
 304                                 $val = $this->sizeHistogram[$i];
 305                         }
 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                         if ( !isset( $coarseHistogram[$coarseIndex] ) ) {
 325                                 $val = 0;
 326                         } else {
 327                                 $val = $coarseHistogram[$coarseIndex];
 328                         }
 329                         $boundary = $coarseBoundaries[$coarseIndex];
 330                         $this->output( sprintf( "%-10s %-10d |%s\n",
 331                                 $prevBoundary . '-' . ( $boundary - 1 ) . ': ',
 332                                 $val,
 333                                 str_repeat( '*', $scale * $val ) ) );
 334                         $prevBoundary = $boundary;
 335                 }
 336         }
 337 }
 338
 339 $maintClass = "UpdateCollation";
 340 require_once RUN_MAINTENANCE_IF_MAIN;