maintenance/dumpTextPass.php

   1 <?php
   2 /**
   3  * Copyright (C) 2005 Brion Vibber <brion@pobox.com>
   4  * http://www.mediawiki.org/
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Maintenance
  23  */
  24
  25 $originalDir = getcwd();
  26
  27 require_once( dirname( __FILE__ ) . '/commandLine.inc' );
  28 require_once( 'backup.inc' );
  29
  30 /**
  31  * @ingroup Maintenance
  32  */
  33 class TextPassDumper extends BackupDumper {
  34         var $prefetch = null;
  35         var $input = "php://stdin";
  36         var $history = WikiExporter::FULL;
  37         var $fetchCount = 0;
  38         var $prefetchCount = 0;
  39
  40         var $failures = 0;
  41         var $maxFailures = 200;
  42         var $failureTimeout = 5; // Seconds to sleep after db failure
  43
  44         var $php = "php";
  45         var $spawn = false;
  46         var $spawnProc = false;
  47         var $spawnWrite = false;
  48         var $spawnRead = false;
  49         var $spawnErr = false;
  50
  51         function dump() {
  52                 # This shouldn't happen if on console... ;)
  53                 header( 'Content-type: text/html; charset=UTF-8' );
  54
  55                 # Notice messages will foul up your XML output even if they're
  56                 # relatively harmless.
  57                 if ( ini_get( 'display_errors' ) )
  58                         ini_set( 'display_errors', 'stderr' );
  59
  60                 $this->initProgress( $this->history );
  61
  62                 $this->db = $this->backupDb();
  63
  64                 $this->egress = new ExportProgressFilter( $this->sink, $this );
  65
  66                 $input = fopen( $this->input, "rt" );
  67                 $result = $this->readDump( $input );
  68
  69                 if ( WikiError::isError( $result ) ) {
  70                         wfDie( $result->getMessage() );
  71                 }
  72
  73                 if ( $this->spawnProc ) {
  74                         $this->closeSpawn();
  75                 }
  76
  77                 $this->report( true );
  78         }
  79
  80         function processOption( $opt, $val, $param ) {
  81                 $url = $this->processFileOpt( $val, $param );
  82
  83                 switch( $opt ) {
  84                 case 'prefetch':
  85                         global $IP;
  86                         require_once "$IP/maintenance/backupPrefetch.inc";
  87                         $this->prefetch = new BaseDump( $url );
  88                         break;
  89                 case 'stub':
  90                         $this->input = $url;
  91                         break;
  92                 case 'current':
  93                         $this->history = WikiExporter::CURRENT;
  94                         break;
  95                 case 'full':
  96                         $this->history = WikiExporter::FULL;
  97                         break;
  98                 case 'spawn':
  99                         $this->spawn = true;
 100                         if ( $val ) {
 101                                 $this->php = $val;
 102                         }
 103                         break;
 104                 }
 105         }
 106
 107         function processFileOpt( $val, $param ) {
 108                 switch( $val ) {
 109                 case "file":
 110                         return $param;
 111                 case "gzip":
 112                         return "compress.zlib://$param";
 113                 case "bzip2":
 114                         return "compress.bzip2://$param";
 115                 case "7zip":
 116                         return "mediawiki.compress.7z://$param";
 117                 default:
 118                         return $val;
 119                 }
 120         }
 121
 122         /**
 123          * Overridden to include prefetch ratio if enabled.
 124          */
 125         function showReport() {
 126                 if ( !$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( $input ) {
 152                 $this->buffer = "";
 153                 $this->openElement = false;
 154                 $this->atStart = true;
 155                 $this->state = "";
 156                 $this->lastName = "";
 157                 $this->thisPage = 0;
 158                 $this->thisRev = 0;
 159
 160                 $parser = xml_parser_create( "UTF-8" );
 161                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 162
 163                 xml_set_element_handler( $parser, array( &$this, 'startElement' ), array( &$this, 'endElement' ) );
 164                 xml_set_character_data_handler( $parser, array( &$this, 'characterData' ) );
 165
 166                 $offset = 0; // for context extraction on error reporting
 167                 $bufferSize = 512 * 1024;
 168                 do {
 169                         $chunk = fread( $input, $bufferSize );
 170                         if ( !xml_parse( $parser, $chunk, feof( $input ) ) ) {
 171                                 wfDebug( "TextDumpPass::readDump encountered XML parsing error\n" );
 172                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 173                         }
 174                         $offset += strlen( $chunk );
 175                 } while ( $chunk !== false && !feof( $input ) );
 176                 xml_parser_free( $parser );
 177
 178                 return true;
 179         }
 180
 181         function getText( $id ) {
 182                 $this->fetchCount++;
 183                 if ( isset( $this->prefetch ) ) {
 184                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 185                         if ( $text === null ) {
 186                                 // Entry missing from prefetch dump
 187                         } elseif ( $text === "" ) {
 188                                 // Blank entries may indicate that the prior dump was broken.
 189                                 // To be safe, reload it.
 190                         } else {
 191                                 $dbr = wfGetDB( DB_SLAVE );
 192                                 $revID = intval($this->thisRev);
 193                                 $revLength = $dbr->selectField( 'revision', 'rev_len', array('rev_id' => $revID ) );
 194                                 // if length of rev text in file doesn't match length in db, we reload
 195                                 // this avoids carrying forward broken data from previous xml dumps
 196                                 if( strlen($text) == $revLength ) {
 197                                         $this->prefetchCount++;
 198                                         return $text;
 199                                 }
 200                         }
 201                 }
 202                 return $this->doGetText( $id );
 203         }
 204
 205         private function doGetText( $id ) {
 206                 if ( $this->spawn ) {
 207                         return $this->getTextSpawned( $id );
 208                 } else {
 209                         return $this->getTextDbSafe( $id );
 210                 }
 211         }
 212
 213         /**
 214          * Fetch a text revision from the database, retrying in case of failure.
 215          * This may survive some transitory errors by reconnecting, but
 216          * may not survive a long-term server outage.
 217          */
 218         private function getTextDbSafe( $id ) {
 219                 while ( true ) {
 220                         try {
 221                                 $text = $this->getTextDb( $id );
 222                                 $ex = new MWException( "Graceful storage failure" );
 223                         } catch ( DBQueryError $ex ) {
 224                                 $text = false;
 225                         }
 226                         if ( $text === false ) {
 227                                 $this->failures++;
 228                                 if ( $this->failures > $this->maxFailures ) {
 229                                         throw $ex;
 230                                 } else {
 231                                         $this->progress( "Database failure $this->failures " .
 232                                                 "of allowed $this->maxFailures for revision $id! " .
 233                                                 "Pausing $this->failureTimeout seconds..." );
 234                                         sleep( $this->failureTimeout );
 235                                 }
 236                         } else {
 237                                 return $text;
 238                         }
 239                 }
 240         }
 241
 242         /**
 243          * May throw a database error if, say, the server dies during query.
 244          */
 245         private function getTextDb( $id ) {
 246                 global $wgContLang;
 247                 $id = intval( $id );
 248                 $row = $this->db->selectRow( 'text',
 249                         array( 'old_text', 'old_flags' ),
 250                         array( 'old_id' => $id ),
 251                         'TextPassDumper::getText' );
 252                 $text = Revision::getRevisionText( $row );
 253                 if ( $text === false ) {
 254                         return false;
 255                 }
 256                 $stripped = str_replace( "\r", "", $text );
 257                 $normalized = $wgContLang->normalize( $stripped );
 258                 return $normalized;
 259         }
 260
 261         private function getTextSpawned( $id ) {
 262                 wfSuppressWarnings();
 263                 if ( !$this->spawnProc ) {
 264                         // First time?
 265                         $this->openSpawn();
 266                 }
 267                 while ( true ) {
 268
 269                         $text = $this->getTextSpawnedOnce( $id );
 270                         if ( !is_string( $text ) ) {
 271                                 $this->progress( "Database subprocess failed. Respawning..." );
 272
 273                                 $this->closeSpawn();
 274                                 sleep( $this->failureTimeout );
 275                                 $this->openSpawn();
 276
 277                                 continue;
 278                         }
 279                         wfRestoreWarnings();
 280                         return $text;
 281                 }
 282         }
 283
 284         function openSpawn() {
 285                 global $IP, $wgDBname;
 286
 287                 $cmd = implode( " ",
 288                         array_map( 'wfEscapeShellArg',
 289                                 array(
 290                                         $this->php,
 291                                         "$IP/maintenance/fetchText.php",
 292                                         $wgDBname ) ) );
 293                 $spec = array(
 294                         0 => array( "pipe", "r" ),
 295                         1 => array( "pipe", "w" ),
 296                         2 => array( "file", "/dev/null", "a" ) );
 297                 $pipes = array();
 298
 299                 $this->progress( "Spawning database subprocess: $cmd" );
 300                 $this->spawnProc = proc_open( $cmd, $spec, $pipes );
 301                 if ( !$this->spawnProc ) {
 302                         // shit
 303                         $this->progress( "Subprocess spawn failed." );
 304                         return false;
 305                 }
 306                 list(
 307                         $this->spawnWrite, // -> stdin
 308                         $this->spawnRead,  // <- stdout
 309                 ) = $pipes;
 310
 311                 return true;
 312         }
 313
 314         private function closeSpawn() {
 315                 wfSuppressWarnings();
 316                 if ( $this->spawnRead )
 317                         fclose( $this->spawnRead );
 318                 $this->spawnRead = false;
 319                 if ( $this->spawnWrite )
 320                         fclose( $this->spawnWrite );
 321                 $this->spawnWrite = false;
 322                 if ( $this->spawnErr )
 323                         fclose( $this->spawnErr );
 324                 $this->spawnErr = false;
 325                 if ( $this->spawnProc )
 326                         pclose( $this->spawnProc );
 327                 $this->spawnProc = false;
 328                 wfRestoreWarnings();
 329         }
 330
 331         private function getTextSpawnedOnce( $id ) {
 332                 global $wgContLang;
 333
 334                 $ok = fwrite( $this->spawnWrite, "$id\n" );
 335                 // $this->progress( ">> $id" );
 336                 if ( !$ok ) return false;
 337
 338                 $ok = fflush( $this->spawnWrite );
 339                 // $this->progress( ">> [flush]" );
 340                 if ( !$ok ) return false;
 341
 342                 // check that the text id they are sending is the one we asked for
 343                 // this avoids out of sync revision text errors we have encountered in the past
 344                 $newId = fgets( $this->spawnRead );
 345                 if ( $newId === false ) {
 346                         return false;
 347                 }
 348                 if ( $id != intval( $newId ) ) {
 349                         return false;
 350                 }
 351
 352                 $len = fgets( $this->spawnRead );
 353                 // $this->progress( "<< " . trim( $len ) );
 354                 if ( $len === false ) return false;
 355
 356                 $nbytes = intval( $len );
 357                 $text = "";
 358
 359                 // Subprocess may not send everything at once, we have to loop.
 360                 while ( $nbytes > strlen( $text ) ) {
 361                         $buffer = fread( $this->spawnRead, $nbytes - strlen( $text ) );
 362                         if ( $buffer === false ) break;
 363                         $text .= $buffer;
 364                 }
 365
 366                 $gotbytes = strlen( $text );
 367                 if ( $gotbytes != $nbytes ) {
 368                         $this->progress( "Expected $nbytes bytes from database subprocess, got $gotbytes " );
 369                         return false;
 370                 }
 371
 372                 // Do normalization in the dump thread...
 373                 $stripped = str_replace( "\r", "", $text );
 374                 $normalized = $wgContLang->normalize( $stripped );
 375                 return $normalized;
 376         }
 377
 378         function startElement( $parser, $name, $attribs ) {
 379                 $this->clearOpenElement( null );
 380                 $this->lastName = $name;
 381
 382                 if ( $name == 'revision' ) {
 383                         $this->state = $name;
 384                         $this->egress->writeOpenPage( null, $this->buffer );
 385                         $this->buffer = "";
 386                 } elseif ( $name == 'page' ) {
 387                         $this->state = $name;
 388                         if ( $this->atStart ) {
 389                                 $this->egress->writeOpenStream( $this->buffer );
 390                                 $this->buffer = "";
 391                                 $this->atStart = false;
 392                         }
 393                 }
 394
 395                 if ( $name == "text" && isset( $attribs['id'] ) ) {
 396                         $text = $this->getText( $attribs['id'] );
 397                         $this->openElement = array( $name, array( 'xml:space' => 'preserve' ) );
 398                         if ( strlen( $text ) > 0 ) {
 399                                 $this->characterData( $parser, $text );
 400                         }
 401                 } else {
 402                         $this->openElement = array( $name, $attribs );
 403                 }
 404         }
 405
 406         function endElement( $parser, $name ) {
 407                 if ( $this->openElement ) {
 408                         $this->clearOpenElement( "" );
 409                 } else {
 410                         $this->buffer .= "</$name>";
 411                 }
 412
 413                 if ( $name == 'revision' ) {
 414                         $this->egress->writeRevision( null, $this->buffer );
 415                         $this->buffer = "";
 416                         $this->thisRev = "";
 417                 } elseif ( $name == 'page' ) {
 418                         $this->egress->writeClosePage( $this->buffer );
 419                         $this->buffer = "";
 420                         $this->thisPage = "";
 421                 } elseif ( $name == 'mediawiki' ) {
 422                         $this->egress->writeCloseStream( $this->buffer );
 423                         $this->buffer = "";
 424                 }
 425         }
 426
 427         function characterData( $parser, $data ) {
 428                 $this->clearOpenElement( null );
 429                 if ( $this->lastName == "id" ) {
 430                         if ( $this->state == "revision" ) {
 431                                 $this->thisRev .= $data;
 432                         } elseif ( $this->state == "page" ) {
 433                                 $this->thisPage .= $data;
 434                         }
 435                 }
 436                 $this->buffer .= htmlspecialchars( $data );
 437         }
 438
 439         function clearOpenElement( $style ) {
 440                 if ( $this->openElement ) {
 441                         $this->buffer .= Xml::element( $this->openElement[0], $this->openElement[1], $style );
 442                         $this->openElement = false;
 443                 }
 444         }
 445 }
 446
 447
 448 $dumper = new TextPassDumper( $argv );
 449
 450 if ( true ) {
 451         $dumper->dump();
 452 } else {
 453         $dumper->progress( <<<ENDS
 454 This script postprocesses XML dumps from dumpBackup.php to add
 455 page text which was stubbed out (using --stub).
 456
 457 XML input is accepted on stdin.
 458 XML output is sent to stdout; progress reports are sent to stderr.
 459
 460 Usage: php dumpTextPass.php [<options>]
 461 Options:
 462   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 463   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 464               pressure on the database.
 465               (Requires PHP 5.0+ and the XMLReader PECL extension)
 466   --quiet     Don't dump status reports to stderr.
 467   --report=n  Report position and speed after every n pages processed.
 468               (Default: 100)
 469   --server=h  Force reading from MySQL server h
 470   --current   Base ETA on number of pages in database instead of all revisions
 471   --spawn     Spawn a subprocess for loading text records
 472 ENDS
 473 );
 474 }
 475
 476