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                 $fileURIs = explode(';',$param);
  99                 foreach ( $fileURIs as $URI ) {
 100                         switch( $val ) {
 101                                 case "file":
 102                                         $newURI = $URI;
 103                                         break;
 104                                 case "gzip":
 105                                         $newURI = "compress.zlib://$URI";
 106                                         break;
 107                                 case "bzip2":
 108                                         $newURI = "compress.bzip2://$URI";
 109                                         break;
 110                                 case "7zip":
 111                                         $newURI = "mediawiki.compress.7z://$URI";
 112                                         break;
 113                                 default:
 114                                         $newURI = $URI;
 115                         }
 116                         $newFileURIs[] = $newURI;
 117                 }
 118                 $val = implode( ';', $newFileURIs );
 119                 return $val;
 120         }
 121
 122         /**
 123          * Overridden to include prefetch ratio if enabled.
 124          */
 125         function showReport() {
 126                 if ( !$this->prefetch ) {
 127                         return parent::showReport();
 128                 }
 129
 130                 if ( $this->reporting ) {
 131                         $delta = wfTime() - $this->startTime;
 132                         $now = wfTimestamp( TS_DB );
 133                         if ( $delta ) {
 134                                 $rate = $this->pageCount / $delta;
 135                                 $revrate = $this->revCount / $delta;
 136                                 $portion = $this->revCount / $this->maxCount;
 137                                 $eta = $this->startTime + $delta / $portion;
 138                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 139                                 $fetchrate = 100.0 * $this->prefetchCount / $this->fetchCount;
 140                         } else {
 141                                 $rate = '-';
 142                                 $revrate = '-';
 143                                 $etats = '-';
 144                                 $fetchrate = '-';
 145                         }
 146                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), %0.1f%% prefetched, ETA %s [max %d]",
 147                                 $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $fetchrate, $etats, $this->maxCount ) );
 148                 }
 149         }
 150
 151         function readDump() {
 152                 $state = '';
 153                 $lastName = '';
 154                 $this->thisPage = 0;
 155                 $this->thisRev = 0;
 156
 157                 $reader = new XMLReader();
 158                 $reader->open( $this->input );
 159                 $writer = new XMLWriter();
 160                 $writer->openMemory();
 161
 162
 163                 while ( $reader->read() ) {
 164                         $tag = $reader->name;
 165                         $type = $reader->nodeType;
 166
 167                         if ( $type == XmlReader::END_ELEMENT ) {
 168                                 $writer->endElement();
 169
 170                                 if ( $tag == 'revision' ) {
 171                                         $this->revCount();
 172                                         $this->thisRev = '';
 173                                 } elseif ( $tag == 'page' ) {
 174                                         $this->reportPage();
 175                                         $this->thisPage = '';
 176                                 }
 177                         } elseif ( $type == XmlReader::ELEMENT ) {
 178                                 $attribs = array();
 179                                 if ( $reader->hasAttributes ) {
 180                                         for ( $i = 0; $reader->moveToAttributeNo( $i ); $i++ ) {
 181                                                 $attribs[$reader->name] = $reader->value;
 182                                         }
 183                                 }
 184
 185                                 if ( $reader->isEmptyElement && $tag == 'text' && isset( $attribs['id'] ) ) {
 186                                         $writer->startElement( 'text' );
 187                                         $writer->writeAttribute( 'xml:space', 'preserve' );
 188                                         $text = $this->getText( $attribs['id'] );
 189                                         if ( strlen( $text ) ) {
 190                                                 $writer->text( $text );
 191                                         }
 192                                         $writer->endElement();
 193                                 } else {
 194                                         $writer->startElement( $tag );
 195                                         foreach( $attribs as $name => $val ) {
 196                                                 $writer->writeAttribute( $name, $val );
 197                                         }
 198                                         if ( $reader->isEmptyElement ) {
 199                                                 $writer->endElement();
 200                                         }
 201                                 }
 202
 203                                 $lastName = $tag;
 204                                 if ( $tag == 'revision' ) {
 205                                         $state = 'revision';
 206                                 } elseif ( $tag == 'page' ) {
 207                                         $state = 'page';
 208                                 }
 209                         } elseif ( $type == XMLReader::SIGNIFICANT_WHITESPACE || $type = XMLReader::TEXT ) {
 210                                 if ( $lastName == 'id' ) {
 211                                         if ( $state == 'revision' ) {
 212                                                 $this->thisRev .= $reader->value;
 213                                         } elseif ( $state == 'page' ) {
 214                                                 $this->thisPage .= $reader->value;
 215                                         }
 216                                 }
 217                                 $writer->text( $reader->value );
 218                         }
 219                         $this->sink->write( $writer->outputMemory() );
 220                 }
 221         }
 222
 223         function getText( $id ) {
 224                 $this->fetchCount++;
 225                 if ( isset( $this->prefetch ) ) {
 226                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 227                         if ( $text !== null ) { // Entry missing from prefetch dump
 228                                 $dbr = wfGetDB( DB_SLAVE );
 229                                 $revID = intval( $this->thisRev );
 230                                 $revLength = $dbr->selectField( 'revision', 'rev_len', array( 'rev_id' => $revID ) );
 231                                 // if length of rev text in file doesn't match length in db, we reload
 232                                 // this avoids carrying forward broken data from previous xml dumps
 233                                 if( strlen( $text ) == $revLength ) {
 234                                         $this->prefetchCount++;
 235                                         return $text;
 236                                 }
 237                         }
 238                 }
 239                 return $this->doGetText( $id );
 240         }
 241
 242         private function doGetText( $id ) {
 243                 $id = intval( $id );
 244                 $this->failures = 0;
 245                 $ex = new MWException( "Graceful storage failure" );
 246                 while (true) {
 247                         if ( $this->spawn ) {
 248                                 if ($this->failures) {
 249                                         // we don't know why it failed, could be the child process
 250                                         // borked, could be db entry busted, could be db server out to lunch,
 251                                         // so cover all bases
 252                                         $this->closeSpawn();
 253                                         $this->openSpawn();
 254                                 }
 255                                 $text =  $this->getTextSpawned( $id );
 256                         } else {
 257                                 $text =  $this->getTextDbSafe( $id );
 258                         }
 259                         if ( $text === false ) {
 260                                 $this->failures++;
 261                                 if ( $this->failures > $this->maxFailures) {
 262                                         $this->progress( "Failed to retrieve revision text for text id ".
 263                                                                          "$id after $this->maxFailures tries, giving up" );
 264                                         // were there so many bad retrievals in a row we want to bail?
 265                                         // at some point we have to declare the dump irretrievably broken
 266                                         $this->failedTextRetrievals++;
 267                                         if ($this->failedTextRetrievals > $this->maxConsecutiveFailedTextRetrievals) {
 268                                                 throw $ex;
 269                                         }
 270                                         else {
 271                                                 // would be nice to return something better to the caller someday,
 272                                                 // log what we know about the failure and about the revision
 273                                                 return("");
 274                                         }
 275                                 } else {
 276                                         $this->progress( "Error $this->failures " .
 277                                                                  "of allowed $this->maxFailures retrieving revision text for text id $id! " .
 278                                                                  "Pausing $this->failureTimeout seconds before retry..." );
 279                                         sleep( $this->failureTimeout );
 280                                 }
 281                         } else {
 282                                 $this->failedTextRetrievals= 0;
 283                                 return( $text );
 284                         }
 285                 }
 286
 287         }
 288
 289         /**
 290          * Fetch a text revision from the database, retrying in case of failure.
 291          * This may survive some transitory errors by reconnecting, but
 292          * may not survive a long-term server outage.
 293          */
 294         private function getTextDbSafe( $id ) {
 295                 while ( true ) {
 296                         try {
 297                                 $text = $this->getTextDb( $id );
 298                         } catch ( DBQueryError $ex ) {
 299                                 $text = false;
 300                         }
 301                         return $text;
 302                 }
 303         }
 304
 305         /**
 306          * May throw a database error if, say, the server dies during query.
 307          */
 308         private function getTextDb( $id ) {
 309                 global $wgContLang;
 310                 $row = $this->db->selectRow( 'text',
 311                         array( 'old_text', 'old_flags' ),
 312                         array( 'old_id' => $id ),
 313                         __METHOD__ );
 314                 $text = Revision::getRevisionText( $row );
 315                 if ( $text === false ) {
 316                         return false;
 317                 }
 318                 $stripped = str_replace( "\r", "", $text );
 319                 $normalized = $wgContLang->normalize( $stripped );
 320                 return $normalized;
 321         }
 322
 323         private function getTextSpawned( $id ) {
 324                 wfSuppressWarnings();
 325                 if ( !$this->spawnProc ) {
 326                         // First time?
 327                         $this->openSpawn();
 328                 }
 329                 $text = $this->getTextSpawnedOnce( $id );
 330                 wfRestoreWarnings();
 331                 return $text;
 332         }
 333
 334         function openSpawn() {
 335                 global $IP;
 336
 337                 $cmd = implode( " ",
 338                         array_map( 'wfEscapeShellArg',
 339                                 array(
 340                                         $this->php,
 341                                         "$IP/maintenance/fetchText.php",
 342                                         '--wiki', wfWikiID() ) ) );
 343                 $spec = array(
 344                         0 => array( "pipe", "r" ),
 345                         1 => array( "pipe", "w" ),
 346                         2 => array( "file", "/dev/null", "a" ) );
 347                 $pipes = array();
 348
 349                 $this->progress( "Spawning database subprocess: $cmd" );
 350                 $this->spawnProc = proc_open( $cmd, $spec, $pipes );
 351                 if ( !$this->spawnProc ) {
 352                         // shit
 353                         $this->progress( "Subprocess spawn failed." );
 354                         return false;
 355                 }
 356                 list(
 357                         $this->spawnWrite, // -> stdin
 358                         $this->spawnRead,  // <- stdout
 359                 ) = $pipes;
 360
 361                 return true;
 362         }
 363
 364         private function closeSpawn() {
 365                 wfSuppressWarnings();
 366                 if ( $this->spawnRead )
 367                         fclose( $this->spawnRead );
 368                 $this->spawnRead = false;
 369                 if ( $this->spawnWrite )
 370                         fclose( $this->spawnWrite );
 371                 $this->spawnWrite = false;
 372                 if ( $this->spawnErr )
 373                         fclose( $this->spawnErr );
 374                 $this->spawnErr = false;
 375                 if ( $this->spawnProc )
 376                         pclose( $this->spawnProc );
 377                 $this->spawnProc = false;
 378                 wfRestoreWarnings();
 379         }
 380
 381         private function getTextSpawnedOnce( $id ) {
 382                 global $wgContLang;
 383
 384                 $ok = fwrite( $this->spawnWrite, "$id\n" );
 385                 // $this->progress( ">> $id" );
 386                 if ( !$ok ) return false;
 387
 388                 $ok = fflush( $this->spawnWrite );
 389                 // $this->progress( ">> [flush]" );
 390                 if ( !$ok ) return false;
 391
 392                 // check that the text id they are sending is the one we asked for
 393                 // this avoids out of sync revision text errors we have encountered in the past
 394                 $newId = fgets( $this->spawnRead );
 395                 if ( $newId === false ) {
 396                         return false;
 397                 }
 398                 if ( $id != intval( $newId ) ) {
 399                         return false;
 400                 }
 401
 402                 $len = fgets( $this->spawnRead );
 403                 // $this->progress( "<< " . trim( $len ) );
 404                 if ( $len === false ) return false;
 405
 406                 $nbytes = intval( $len );
 407                 // actual error, not zero-length text
 408                 if ($nbytes < 0 ) return false;
 409
 410                 $text = "";
 411
 412                 // Subprocess may not send everything at once, we have to loop.
 413                 while ( $nbytes > strlen( $text ) ) {
 414                         $buffer = fread( $this->spawnRead, $nbytes - strlen( $text ) );
 415                         if ( $buffer === false ) break;
 416                         $text .= $buffer;
 417                 }
 418
 419                 $gotbytes = strlen( $text );
 420                 if ( $gotbytes != $nbytes ) {
 421                         $this->progress( "Expected $nbytes bytes from database subprocess, got $gotbytes " );
 422                         return false;
 423                 }
 424
 425                 // Do normalization in the dump thread...
 426                 $stripped = str_replace( "\r", "", $text );
 427                 $normalized = $wgContLang->normalize( $stripped );
 428                 return $normalized;
 429         }
 430 }
 431
 432
 433 $dumper = new TextPassDumper( $argv );
 434
 435 if ( !isset( $options['help'] ) ) {
 436         $dumper->dump( WikiExporter::FULL );
 437 } else {
 438         $dumper->progress( <<<ENDS
 439 This script postprocesses XML dumps from dumpBackup.php to add
 440 page text which was stubbed out (using --stub).
 441
 442 XML input is accepted on stdin.
 443 XML output is sent to stdout; progress reports are sent to stderr.
 444
 445 Usage: php dumpTextPass.php [<options>]
 446 Options:
 447   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 448   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 449                           pressure on the database.
 450   --quiet         Don't dump status reports to stderr.
 451   --report=n  Report position and speed after every n pages processed.
 452                           (Default: 100)
 453   --server=h  Force reading from MySQL server h
 454   --output=<type>:<file> Write to a file instead of stdout
 455               <type>s: file, gzip, bzip2, 7zip
 456   --current       Base ETA on number of pages in database instead of all revisions
 457   --spawn         Spawn a subprocess for loading text records
 458   --help      Display this help message
 459 ENDS
 460 );
 461 }
 462
 463