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 ) { // Entry missing from prefetch dump
 218                                 $dbr = wfGetDB( DB_SLAVE );
 219                                 $revID = intval( $this->thisRev );
 220                                 $revLength = $dbr->selectField( 'revision', 'rev_len', array( 'rev_id' => $revID ) );
 221                                 // if length of rev text in file doesn't match length in db, we reload
 222                                 // this avoids carrying forward broken data from previous xml dumps
 223                                 if( strlen( $text ) == $revLength ) {
 224                                         $this->prefetchCount++;
 225                                         return $text;
 226                                 }
 227                         }
 228                 }
 229                 return $this->doGetText( $id );
 230         }
 231
 232         private function doGetText( $id ) {
 233                 $id = intval( $id );
 234                 $this->failures = 0;
 235                 $ex = new MWException( "Graceful storage failure" );
 236                 while (true) {
 237                         if ( $this->spawn ) {
 238                                 if ($this->failures) {
 239                                         // we don't know why it failed, could be the child process
 240                                         // borked, could be db entry busted, could be db server out to lunch,
 241                                         // so cover all bases
 242                                         $this->closeSpawn();
 243                                         $this->openSpawn();
 244                                 }
 245                                 $text =  $this->getTextSpawned( $id );
 246                         } else {
 247                                 $text =  $this->getTextDbSafe( $id );
 248                         }
 249                         if ( $text === false ) {
 250                                 $this->failures++;
 251                                 if ( $this->failures > $this->maxFailures) {
 252                                         $this->progress( "Failed to retrieve revision text for text id ".
 253                                                                          "$id after $this->maxFailures tries, giving up" );
 254                                         // were there so many bad retrievals in a row we want to bail?
 255                                         // at some point we have to declare the dump irretrievably broken
 256                                         $this->failedTextRetrievals++;
 257                                         if ($this->failedTextRetrievals > $this->maxConsecutiveFailedTextRetrievals) {
 258                                                 throw $ex;
 259                                         }
 260                                         else {
 261                                                 // would be nice to return something better to the caller someday,
 262                                                 // log what we know about the failure and about the revision
 263                                                 return("");
 264                                         }
 265                                 } else {
 266                                         $this->progress( "Error $this->failures " .
 267                                                                  "of allowed $this->maxFailures retrieving revision text for text id $id! " .
 268                                                                  "Pausing $this->failureTimeout seconds before retry..." );
 269                                         sleep( $this->failureTimeout );
 270                                 }
 271                         } else {
 272                                 $this->failedTextRetrievals= 0;
 273                                 return( $text );
 274                         }
 275                 }
 276
 277         }
 278
 279         /**
 280          * Fetch a text revision from the database, retrying in case of failure.
 281          * This may survive some transitory errors by reconnecting, but
 282          * may not survive a long-term server outage.
 283          */
 284         private function getTextDbSafe( $id ) {
 285                 while ( true ) {
 286                         try {
 287                                 $text = $this->getTextDb( $id );
 288                         } catch ( DBQueryError $ex ) {
 289                                 $text = false;
 290                         }
 291                         return $text;
 292                 }
 293         }
 294
 295         /**
 296          * May throw a database error if, say, the server dies during query.
 297          */
 298         private function getTextDb( $id ) {
 299                 global $wgContLang;
 300                 $row = $this->db->selectRow( 'text',
 301                         array( 'old_text', 'old_flags' ),
 302                         array( 'old_id' => $id ),
 303                         __METHOD__ );
 304                 $text = Revision::getRevisionText( $row );
 305                 if ( $text === false ) {
 306                         return false;
 307                 }
 308                 $stripped = str_replace( "\r", "", $text );
 309                 $normalized = $wgContLang->normalize( $stripped );
 310                 return $normalized;
 311         }
 312
 313         private function getTextSpawned( $id ) {
 314                 wfSuppressWarnings();
 315                 if ( !$this->spawnProc ) {
 316                         // First time?
 317                         $this->openSpawn();
 318                 }
 319                 $text = $this->getTextSpawnedOnce( $id );
 320                 wfRestoreWarnings();
 321                 return $text;
 322         }
 323
 324         function openSpawn() {
 325                 global $IP;
 326
 327                 $cmd = implode( " ",
 328                         array_map( 'wfEscapeShellArg',
 329                                 array(
 330                                         $this->php,
 331                                         "$IP/maintenance/fetchText.php",
 332                                         '--wiki', wfWikiID() ) ) );
 333                 $spec = array(
 334                         0 => array( "pipe", "r" ),
 335                         1 => array( "pipe", "w" ),
 336                         2 => array( "file", "/dev/null", "a" ) );
 337                 $pipes = array();
 338
 339                 $this->progress( "Spawning database subprocess: $cmd" );
 340                 $this->spawnProc = proc_open( $cmd, $spec, $pipes );
 341                 if ( !$this->spawnProc ) {
 342                         // shit
 343                         $this->progress( "Subprocess spawn failed." );
 344                         return false;
 345                 }
 346                 list(
 347                         $this->spawnWrite, // -> stdin
 348                         $this->spawnRead,  // <- stdout
 349                 ) = $pipes;
 350
 351                 return true;
 352         }
 353
 354         private function closeSpawn() {
 355                 wfSuppressWarnings();
 356                 if ( $this->spawnRead )
 357                         fclose( $this->spawnRead );
 358                 $this->spawnRead = false;
 359                 if ( $this->spawnWrite )
 360                         fclose( $this->spawnWrite );
 361                 $this->spawnWrite = false;
 362                 if ( $this->spawnErr )
 363                         fclose( $this->spawnErr );
 364                 $this->spawnErr = false;
 365                 if ( $this->spawnProc )
 366                         pclose( $this->spawnProc );
 367                 $this->spawnProc = false;
 368                 wfRestoreWarnings();
 369         }
 370
 371         private function getTextSpawnedOnce( $id ) {
 372                 global $wgContLang;
 373
 374                 $ok = fwrite( $this->spawnWrite, "$id\n" );
 375                 // $this->progress( ">> $id" );
 376                 if ( !$ok ) return false;
 377
 378                 $ok = fflush( $this->spawnWrite );
 379                 // $this->progress( ">> [flush]" );
 380                 if ( !$ok ) return false;
 381
 382                 // check that the text id they are sending is the one we asked for
 383                 // this avoids out of sync revision text errors we have encountered in the past
 384                 $newId = fgets( $this->spawnRead );
 385                 if ( $newId === false ) {
 386                         return false;
 387                 }
 388                 if ( $id != intval( $newId ) ) {
 389                         return false;
 390                 }
 391
 392                 $len = fgets( $this->spawnRead );
 393                 // $this->progress( "<< " . trim( $len ) );
 394                 if ( $len === false ) return false;
 395
 396                 $nbytes = intval( $len );
 397                 // actual error, not zero-length text
 398                 if ($nbytes < 0 ) return false;
 399
 400                 $text = "";
 401
 402                 // Subprocess may not send everything at once, we have to loop.
 403                 while ( $nbytes > strlen( $text ) ) {
 404                         $buffer = fread( $this->spawnRead, $nbytes - strlen( $text ) );
 405                         if ( $buffer === false ) break;
 406                         $text .= $buffer;
 407                 }
 408
 409                 $gotbytes = strlen( $text );
 410                 if ( $gotbytes != $nbytes ) {
 411                         $this->progress( "Expected $nbytes bytes from database subprocess, got $gotbytes " );
 412                         return false;
 413                 }
 414
 415                 // Do normalization in the dump thread...
 416                 $stripped = str_replace( "\r", "", $text );
 417                 $normalized = $wgContLang->normalize( $stripped );
 418                 return $normalized;
 419         }
 420 }
 421
 422
 423 $dumper = new TextPassDumper( $argv );
 424
 425 if ( !isset( $options['help'] ) ) {
 426         $dumper->dump( WikiExporter::FULL );
 427 } else {
 428         $dumper->progress( <<<ENDS
 429 This script postprocesses XML dumps from dumpBackup.php to add
 430 page text which was stubbed out (using --stub).
 431
 432 XML input is accepted on stdin.
 433 XML output is sent to stdout; progress reports are sent to stderr.
 434
 435 Usage: php dumpTextPass.php [<options>]
 436 Options:
 437   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 438   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 439                           pressure on the database.
 440   --quiet         Don't dump status reports to stderr.
 441   --report=n  Report position and speed after every n pages processed.
 442                           (Default: 100)
 443   --server=h  Force reading from MySQL server h
 444   --current       Base ETA on number of pages in database instead of all revisions
 445   --spawn         Spawn a subprocess for loading text records
 446   --help      Display this help message
 447 ENDS
 448 );
 449 }
 450
 451