maintenance/dumpTextPass.php

   1 <?php
   2 /**
   3  * Script that postprocesses XML dumps from dumpBackup.php to add page text
   4  *
   5  * Copyright © 2005 Brion Vibber <brion@pobox.com>, 2010 Alexandre Emsenhuber
   6  * http://www.mediawiki.org/
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Maintenance
  25  */
  26
  27 $originalDir = getcwd();
  28
  29 require_once( dirname( __FILE__ ) . '/commandLine.inc' );
  30 require_once( 'backup.inc' );
  31
  32 /**
  33  * @ingroup Maintenance
  34  */
  35 class TextPassDumper extends BackupDumper {
  36         var $prefetch = null;
  37         var $input = "php://stdin";
  38         var $fetchCount = 0;
  39         var $prefetchCount = 0;
  40
  41         var $failures = 0;
  42         var $maxFailures = 5;
  43         var $failedTextRetrievals = 0;
  44         var $maxConsecutiveFailedTextRetrievals = 200;
  45         var $failureTimeout = 5; // Seconds to sleep after db failure
  46
  47         var $php = "php";
  48         var $spawn = false;
  49         var $spawnProc = false;
  50         var $spawnWrite = false;
  51         var $spawnRead = false;
  52         var $spawnErr = false;
  53
  54         function dump( $history, $text = WikiExporter::TEXT ) {
  55                 # This shouldn't happen if on console... ;)
  56                 header( 'Content-type: text/html; charset=UTF-8' );
  57
  58                 # Notice messages will foul up your XML output even if they're
  59                 # relatively harmless.
  60                 if ( ini_get( 'display_errors' ) )
  61                         ini_set( 'display_errors', 'stderr' );
  62
  63                 $this->initProgress( $history );
  64
  65                 $this->db = $this->backupDb();
  66
  67                 $this->readDump();
  68
  69                 if ( $this->spawnProc ) {
  70                         $this->closeSpawn();
  71                 }
  72
  73                 $this->report( true );
  74         }
  75
  76         function processOption( $opt, $val, $param ) {
  77                 global $IP;
  78                 $url = $this->processFileOpt( $val, $param );
  79
  80                 switch( $opt ) {
  81                 case 'prefetch':
  82                         require_once "$IP/maintenance/backupPrefetch.inc";
  83                         $this->prefetch = new BaseDump( $url );
  84                         break;
  85                 case 'stub':
  86                         $this->input = $url;
  87                         break;
  88                 case 'spawn':
  89                         $this->spawn = true;
  90                         if ( $val ) {
  91                                 $this->php = $val;
  92                         }
  93                         break;
  94                 }
  95         }
  96
  97         function processFileOpt( $val, $param ) {
  98                 switch( $val ) {
  99                 case "file":
 100                         return $param;
 101                 case "gzip":
 102                         return "compress.zlib://$param";
 103                 case "bzip2":
 104                         return "compress.bzip2://$param";
 105                 case "7zip":
 106                         return "mediawiki.compress.7z://$param";
 107                 default:
 108                         return $val;
 109                 }
 110         }
 111
 112         /**
 113          * Overridden to include prefetch ratio if enabled.
 114          */
 115         function showReport() {
 116                 if ( !$this->prefetch ) {
 117                         return parent::showReport();
 118                 }
 119
 120                 if ( $this->reporting ) {
 121                         $delta = wfTime() - $this->startTime;
 122                         $now = wfTimestamp( TS_DB );
 123                         if ( $delta ) {
 124                                 $rate = $this->pageCount / $delta;
 125                                 $revrate = $this->revCount / $delta;
 126                                 $portion = $this->revCount / $this->maxCount;
 127                                 $eta = $this->startTime + $delta / $portion;
 128                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 129                                 $fetchrate = 100.0 * $this->prefetchCount / $this->fetchCount;
 130                         } else {
 131                                 $rate = '-';
 132                                 $revrate = '-';
 133                                 $etats = '-';
 134                                 $fetchrate = '-';
 135                         }
 136                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), %0.1f%% prefetched, ETA %s [max %d]",
 137                                 $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $fetchrate, $etats, $this->maxCount ) );
 138                 }
 139         }
 140
 141         function readDump() {
 142                 $state = '';
 143                 $lastName = '';
 144                 $this->thisPage = 0;
 145                 $this->thisRev = 0;
 146
 147                 $reader = new XMLReader();
 148                 $reader->open( $this->input );
 149                 $writer = new XMLWriter();
 150                 $writer->openURI( 'php://stdout' );
 151
 152
 153                 while ( $reader->read() ) {
 154                         $tag = $reader->name;
 155                         $type = $reader->nodeType;
 156
 157                         if ( $type == XmlReader::END_ELEMENT ) {
 158                                 $writer->endElement();
 159
 160                                 if ( $tag == 'revision' ) {
 161                                         $this->revCount();
 162                                         $this->thisRev = '';
 163                                 } elseif ( $tag == 'page' ) {
 164                                         $this->reportPage();
 165                                         $this->thisPage = '';
 166                                 }
 167                         } elseif ( $type == XmlReader::ELEMENT ) {
 168                                 $attribs = array();
 169                                 if ( $reader->hasAttributes ) {
 170                                         for ( $i = 0; $reader->moveToAttributeNo( $i ); $i++ ) {
 171                                                 $attribs[$reader->name] = $reader->value;
 172                                         }
 173                                 }
 174
 175                                 if ( $reader->isEmptyElement && $tag == 'text' && isset( $attribs['id'] ) ) {
 176                                         $writer->startElement( 'text' );
 177                                         $writer->writeAttribute( 'xml:space', 'preserve' );
 178                                         $text = $this->getText( $attribs['id'] );
 179                                         if ( strlen( $text ) ) {
 180                                                 $writer->text( $text );
 181                                         }
 182                                         $writer->endElement();
 183                                 } else {
 184                                         $writer->startElement( $tag );
 185                                         foreach( $attribs as $name => $val ) {
 186                                                 $writer->writeAttribute( $name, $val );
 187                                         }
 188                                         if ( $reader->isEmptyElement ) {
 189                                                 $writer->endElement();
 190                                         }
 191                                 }
 192
 193                                 $lastName = $tag;
 194                                 if ( $tag == 'revision' ) {
 195                                         $state = 'revision';
 196                                 } elseif ( $tag == 'page' ) {
 197                                         $state = 'page';
 198                                 }
 199                         } elseif ( $type == XMLReader::SIGNIFICANT_WHITESPACE || $type = XMLReader::TEXT ) {
 200                                 if ( $lastName == 'id' ) {
 201                                         if ( $state == 'revision' ) {
 202                                                 $this->thisRev .= $reader->value;
 203                                         } elseif ( $state == 'page' ) {
 204                                                 $this->thisPage .= $reader->value;
 205                                         }
 206                                 }
 207                                 $writer->text( $reader->value );
 208                         }
 209                 }
 210                 $writer->flush();
 211         }
 212
 213         function getText( $id ) {
 214                 $this->fetchCount++;
 215                 if ( isset( $this->prefetch ) ) {
 216                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 217                         if ( $text === null ) {
 218                                 // Entry missing from prefetch dump
 219                         } elseif ( $text === "" ) {
 220                                 // Blank entries may indicate that the prior dump was broken.
 221                                 // To be safe, reload it.
 222                         } else {
 223                                 $dbr = wfGetDB( DB_SLAVE );
 224                                 $revID = intval($this->thisRev);
 225                                 $revLength = $dbr->selectField( 'revision', 'rev_len', array('rev_id' => $revID ) );
 226                                 // if length of rev text in file doesn't match length in db, we reload
 227                                 // this avoids carrying forward broken data from previous xml dumps
 228                                 if( strlen($text) == $revLength ) {
 229                                         $this->prefetchCount++;
 230                                         return $text;
 231                                 }
 232                         }
 233                 }
 234                 return $this->doGetText( $id );
 235         }
 236
 237         private function doGetText( $id ) {
 238                 $id = intval( $id );
 239                 $this->failures = 0;
 240                 $ex = new MWException( "Graceful storage failure" );
 241                 while (true) {
 242                         if ( $this->spawn ) {
 243                                 if ($this->failures) {
 244                                         // we don't know why it failed, could be the child process
 245                                         // borked, could be db entry busted, could be db server out to lunch,
 246                                         // so cover all bases
 247                                         $this->closeSpawn();
 248                                         $this->openSpawn();
 249                                 }
 250                                 $text =  $this->getTextSpawned( $id );
 251                         } else {
 252                                 $text =  $this->getTextDbSafe( $id );
 253                         }
 254                         if ( $text === false ) {
 255                                 $this->failures++;
 256                                 if ( $this->failures > $this->maxFailures) {
 257                                         $this->progress( "Failed to retrieve revision text for text id ".
 258                                                                          "$id after $this->maxFailures tries, giving up" );
 259                                         // were there so many bad retrievals in a row we want to bail?
 260                                         // at some point we have to declare the dump irretrievably broken
 261                                         $this->failedTextRetrievals++;
 262                                         if ($this->failedTextRetrievals > $this->maxConsecutiveFailedTextRetrievals) {
 263                                                 throw $ex;
 264                                         }
 265                                         else {
 266                                                 // would be nice to return something better to the caller someday,
 267                                                 // log what we know about the failure and about the revision
 268                                                 return("");
 269                                         }
 270                                 } else {
 271                                         $this->progress( "Error $this->failures " .
 272                                                                  "of allowed $this->maxFailures retrieving revision text for text id $id! " .
 273                                                                  "Pausing $this->failureTimeout seconds before retry..." );
 274                                         sleep( $this->failureTimeout );
 275                                 }
 276                         } else {
 277                                 $this->failedTextRetrievals= 0;
 278                                 return( $text );
 279                         }
 280                 }
 281
 282         }
 283
 284         /**
 285          * Fetch a text revision from the database, retrying in case of failure.
 286          * This may survive some transitory errors by reconnecting, but
 287          * may not survive a long-term server outage.
 288          */
 289         private function getTextDbSafe( $id ) {
 290                 while ( true ) {
 291                         try {
 292                                 $text = $this->getTextDb( $id );
 293                         } catch ( DBQueryError $ex ) {
 294                                 $text = false;
 295                         }
 296                         return $text;
 297                 }
 298         }
 299
 300         /**
 301          * May throw a database error if, say, the server dies during query.
 302          */
 303         private function getTextDb( $id ) {
 304                 global $wgContLang;
 305                 $row = $this->db->selectRow( 'text',
 306                         array( 'old_text', 'old_flags' ),
 307                         array( 'old_id' => $id ),
 308                         __METHOD__ );
 309                 $text = Revision::getRevisionText( $row );
 310                 if ( $text === false ) {
 311                         return false;
 312                 }
 313                 $stripped = str_replace( "\r", "", $text );
 314                 $normalized = $wgContLang->normalize( $stripped );
 315                 return $normalized;
 316         }
 317
 318         private function getTextSpawned( $id ) {
 319                 wfSuppressWarnings();
 320                 if ( !$this->spawnProc ) {
 321                         // First time?
 322                         $this->openSpawn();
 323                 }
 324                 $text = $this->getTextSpawnedOnce( $id );
 325                 wfRestoreWarnings();
 326                 return $text;
 327         }
 328
 329         function openSpawn() {
 330                 global $IP, $wgDBname;
 331
 332                 $cmd = implode( " ",
 333                         array_map( 'wfEscapeShellArg',
 334                                 array(
 335                                         $this->php,
 336                                         "$IP/maintenance/fetchText.php",
 337                                         $wgDBname ) ) );
 338                 $spec = array(
 339                         0 => array( "pipe", "r" ),
 340                         1 => array( "pipe", "w" ),
 341                         2 => array( "file", "/dev/null", "a" ) );
 342                 $pipes = array();
 343
 344                 $this->progress( "Spawning database subprocess: $cmd" );
 345                 $this->spawnProc = proc_open( $cmd, $spec, $pipes );
 346                 if ( !$this->spawnProc ) {
 347                         // shit
 348                         $this->progress( "Subprocess spawn failed." );
 349                         return false;
 350                 }
 351                 list(
 352                         $this->spawnWrite, // -> stdin
 353                         $this->spawnRead,  // <- stdout
 354                 ) = $pipes;
 355
 356                 return true;
 357         }
 358
 359         private function closeSpawn() {
 360                 wfSuppressWarnings();
 361                 if ( $this->spawnRead )
 362                         fclose( $this->spawnRead );
 363                 $this->spawnRead = false;
 364                 if ( $this->spawnWrite )
 365                         fclose( $this->spawnWrite );
 366                 $this->spawnWrite = false;
 367                 if ( $this->spawnErr )
 368                         fclose( $this->spawnErr );
 369                 $this->spawnErr = false;
 370                 if ( $this->spawnProc )
 371                         pclose( $this->spawnProc );
 372                 $this->spawnProc = false;
 373                 wfRestoreWarnings();
 374         }
 375
 376         private function getTextSpawnedOnce( $id ) {
 377                 global $wgContLang;
 378
 379                 $ok = fwrite( $this->spawnWrite, "$id\n" );
 380                 // $this->progress( ">> $id" );
 381                 if ( !$ok ) return false;
 382
 383                 $ok = fflush( $this->spawnWrite );
 384                 // $this->progress( ">> [flush]" );
 385                 if ( !$ok ) return false;
 386
 387                 // check that the text id they are sending is the one we asked for
 388                 // this avoids out of sync revision text errors we have encountered in the past
 389                 $newId = fgets( $this->spawnRead );
 390                 if ( $newId === false ) {
 391                         return false;
 392                 }
 393                 if ( $id != intval( $newId ) ) {
 394                         return false;
 395                 }
 396
 397                 $len = fgets( $this->spawnRead );
 398                 // $this->progress( "<< " . trim( $len ) );
 399                 if ( $len === false ) return false;
 400
 401                 $nbytes = intval( $len );
 402                 // actual error, not zero-length text
 403                 if ($nbytes < 0 ) return false;
 404
 405                 $text = "";
 406
 407                 // Subprocess may not send everything at once, we have to loop.
 408                 while ( $nbytes > strlen( $text ) ) {
 409                         $buffer = fread( $this->spawnRead, $nbytes - strlen( $text ) );
 410                         if ( $buffer === false ) break;
 411                         $text .= $buffer;
 412                 }
 413
 414                 $gotbytes = strlen( $text );
 415                 if ( $gotbytes != $nbytes ) {
 416                         $this->progress( "Expected $nbytes bytes from database subprocess, got $gotbytes " );
 417                         return false;
 418                 }
 419
 420                 // Do normalization in the dump thread...
 421                 $stripped = str_replace( "\r", "", $text );
 422                 $normalized = $wgContLang->normalize( $stripped );
 423                 return $normalized;
 424         }
 425 }
 426
 427
 428 $dumper = new TextPassDumper( $argv );
 429
 430 if ( !isset( $options['help'] ) ) {
 431         $dumper->dump( WikiExporter::FULL );
 432 } else {
 433         $dumper->progress( <<<ENDS
 434 This script postprocesses XML dumps from dumpBackup.php to add
 435 page text which was stubbed out (using --stub).
 436
 437 XML input is accepted on stdin.
 438 XML output is sent to stdout; progress reports are sent to stderr.
 439
 440 Usage: php dumpTextPass.php [<options>]
 441 Options:
 442   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 443   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 444                           pressure on the database.
 445   --quiet         Don't dump status reports to stderr.
 446   --report=n  Report position and speed after every n pages processed.
 447                           (Default: 100)
 448   --server=h  Force reading from MySQL server h
 449   --current       Base ETA on number of pages in database instead of all revisions
 450   --spawn         Spawn a subprocess for loading text records
 451   --help      Display this help message
 452 ENDS
 453 );
 454 }
 455
 456