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( 'commandLine.inc' );
  28 require_once( 'backup.inc' );
  29
  30 /**
  31  * Stream wrapper around 7za filter program.
  32  * Required since we can't pass an open file resource to XMLReader->open()
  33  * which is used for the text prefetch.
  34  *
  35  * @ingroup Maintenance
  36  */
  37 class SevenZipStream {
  38         var $stream;
  39
  40         private function stripPath( $path ) {
  41                 $prefix = 'mediawiki.compress.7z://';
  42                 return substr( $path, strlen( $prefix ) );
  43         }
  44
  45         function stream_open( $path, $mode, $options, &$opened_path ) {
  46                 if( $mode{0} == 'r' ) {
  47                         $options = 'e -bd -so';
  48                 } elseif( $mode{0} == 'w' ) {
  49                         $options = 'a -bd -si';
  50                 } else {
  51                         return false;
  52                 }
  53                 $arg = wfEscapeShellArg( $this->stripPath( $path ) );
  54                 $command = "7za $options $arg";
  55                 if( !wfIsWindows() ) {
  56                         // Suppress the stupid messages on stderr
  57                         $command .= ' 2>/dev/null';
  58                 }
  59                 $this->stream = popen( $command, $mode );
  60                 return ($this->stream !== false);
  61         }
  62
  63         function url_stat( $path, $flags ) {
  64                 return stat( $this->stripPath( $path ) );
  65         }
  66
  67         // This is all so lame; there should be a default class we can extend
  68
  69         function stream_close() {
  70                 return fclose( $this->stream );
  71         }
  72
  73         function stream_flush() {
  74                 return fflush( $this->stream );
  75         }
  76
  77         function stream_read( $count ) {
  78                 return fread( $this->stream, $count );
  79         }
  80
  81         function stream_write( $data ) {
  82                 return fwrite( $this->stream, $data );
  83         }
  84
  85         function stream_tell() {
  86                 return ftell( $this->stream );
  87         }
  88
  89         function stream_eof() {
  90                 return feof( $this->stream );
  91         }
  92
  93         function stream_seek( $offset, $whence ) {
  94                 return fseek( $this->stream, $offset, $whence );
  95         }
  96 }
  97 stream_wrapper_register( 'mediawiki.compress.7z', 'SevenZipStream' );
  98
  99 /**
 100  * @ingroup Maintenance
 101  */
 102 class TextPassDumper extends BackupDumper {
 103         var $prefetch = null;
 104         var $input = "php://stdin";
 105         var $history = WikiExporter::FULL;
 106         var $fetchCount = 0;
 107         var $prefetchCount = 0;
 108
 109         var $failures = 0;
 110         var $maxFailures = 200;
 111         var $failureTimeout = 5; // Seconds to sleep after db failure
 112
 113         var $php = "php";
 114         var $spawn = false;
 115         var $spawnProc = false;
 116         var $spawnWrite = false;
 117         var $spawnRead = false;
 118         var $spawnErr = false;
 119
 120         function dump() {
 121                 # This shouldn't happen if on console... ;)
 122                 header( 'Content-type: text/html; charset=UTF-8' );
 123
 124                 # Notice messages will foul up your XML output even if they're
 125                 # relatively harmless.
 126                 if( ini_get( 'display_errors' ) )
 127                         ini_set( 'display_errors', 'stderr' );
 128
 129                 $this->initProgress( $this->history );
 130
 131                 $this->db = $this->backupDb();
 132
 133                 $this->egress = new ExportProgressFilter( $this->sink, $this );
 134
 135                 $input = fopen( $this->input, "rt" );
 136                 $result = $this->readDump( $input );
 137
 138                 if( WikiError::isError( $result ) ) {
 139                         wfDie( $result->getMessage() );
 140                 }
 141
 142                 if( $this->spawnProc ) {
 143                         $this->closeSpawn();
 144                 }
 145
 146                 $this->report( true );
 147         }
 148
 149         function processOption( $opt, $val, $param ) {
 150                 $url = $this->processFileOpt( $val, $param );
 151
 152                 switch( $opt ) {
 153                 case 'prefetch':
 154                         global $IP;
 155                         require_once "$IP/maintenance/backupPrefetch.inc";
 156                         $this->prefetch = new BaseDump( $url );
 157                         break;
 158                 case 'stub':
 159                         $this->input = $url;
 160                         break;
 161                 case 'current':
 162                         $this->history = WikiExporter::CURRENT;
 163                         break;
 164                 case 'full':
 165                         $this->history = WikiExporter::FULL;
 166                         break;
 167                 case 'spawn':
 168                         $this->spawn = true;
 169                         if( $val ) {
 170                                 $this->php = $val;
 171                         }
 172                         break;
 173                 }
 174         }
 175
 176         function processFileOpt( $val, $param ) {
 177                 switch( $val ) {
 178                 case "file":
 179                         return $param;
 180                 case "gzip":
 181                         return "compress.zlib://$param";
 182                 case "bzip2":
 183                         return "compress.bzip2://$param";
 184                 case "7zip":
 185                         return "mediawiki.compress.7z://$param";
 186                 default:
 187                         return $val;
 188                 }
 189         }
 190
 191         /**
 192          * Overridden to include prefetch ratio if enabled.
 193          */
 194         function showReport() {
 195                 if( !$this->prefetch ) {
 196                         return parent::showReport();
 197                 }
 198
 199                 if( $this->reporting ) {
 200                         $delta = wfTime() - $this->startTime;
 201                         $now = wfTimestamp( TS_DB );
 202                         if( $delta ) {
 203                                 $rate = $this->pageCount / $delta;
 204                                 $revrate = $this->revCount / $delta;
 205                                 $portion = $this->revCount / $this->maxCount;
 206                                 $eta = $this->startTime + $delta / $portion;
 207                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 208                                 $fetchrate = 100.0 * $this->prefetchCount / $this->fetchCount;
 209                         } else {
 210                                 $rate = '-';
 211                                 $revrate = '-';
 212                                 $etats = '-';
 213                                 $fetchrate = '-';
 214                         }
 215                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), %0.1f%% prefetched, ETA %s [max %d]",
 216                                 $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $fetchrate, $etats, $this->maxCount ) );
 217                 }
 218         }
 219
 220         function readDump( $input ) {
 221                 $this->buffer = "";
 222                 $this->openElement = false;
 223                 $this->atStart = true;
 224                 $this->state = "";
 225                 $this->lastName = "";
 226                 $this->thisPage = 0;
 227                 $this->thisRev = 0;
 228
 229                 $parser = xml_parser_create( "UTF-8" );
 230                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 231
 232                 xml_set_element_handler( $parser, array( &$this, 'startElement' ), array( &$this, 'endElement' ) );
 233                 xml_set_character_data_handler( $parser, array( &$this, 'characterData' ) );
 234
 235                 $offset = 0; // for context extraction on error reporting
 236                 $bufferSize = 512 * 1024;
 237                 do {
 238                         $chunk = fread( $input, $bufferSize );
 239                         if( !xml_parse( $parser, $chunk, feof( $input ) ) ) {
 240                                 wfDebug( "TextDumpPass::readDump encountered XML parsing error\n" );
 241                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 242                         }
 243                         $offset += strlen( $chunk );
 244                 } while( $chunk !== false && !feof( $input ) );
 245                 xml_parser_free( $parser );
 246
 247                 return true;
 248         }
 249
 250         function getText( $id ) {
 251                 $this->fetchCount++;
 252                 if( isset( $this->prefetch ) ) {
 253                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 254                         if( $text === null ) {
 255                                 // Entry missing from prefetch dump
 256                         } elseif( $text === "" ) {
 257                                 // Blank entries may indicate that the prior dump was broken.
 258                                 // To be safe, reload it.
 259                         } else {
 260                                 $this->prefetchCount++;
 261                                 return $text;
 262                         }
 263                 }
 264                 return $this->doGetText( $id );
 265         }
 266
 267         private function doGetText( $id ) {
 268                 if( $this->spawn ) {
 269                         return $this->getTextSpawned( $id );
 270                 } else {
 271                         return $this->getTextDbSafe( $id );
 272                 }
 273         }
 274
 275         /**
 276          * Fetch a text revision from the database, retrying in case of failure.
 277          * This may survive some transitory errors by reconnecting, but
 278          * may not survive a long-term server outage.
 279          */
 280         private function getTextDbSafe( $id ) {
 281                 while( true ) {
 282                         try {
 283                                 $text = $this->getTextDb( $id );
 284                                 $ex = new MWException("Graceful storage failure");
 285                         } catch (DBQueryError $ex) {
 286                                 $text = false;
 287                         }
 288                         if( $text === false ) {
 289                                 $this->failures++;
 290                                 if( $this->failures > $this->maxFailures ) {
 291                                         throw $ex;
 292                                 } else {
 293                                         $this->progress( "Database failure $this->failures " .
 294                                                 "of allowed $this->maxFailures for revision $id! " .
 295                                                 "Pausing $this->failureTimeout seconds..." );
 296                                         sleep( $this->failureTimeout );
 297                                 }
 298                         } else {
 299                                 return $text;
 300                         }
 301                 }
 302         }
 303
 304         /**
 305          * May throw a database error if, say, the server dies during query.
 306          */
 307         private function getTextDb( $id ) {
 308                 $id = intval( $id );
 309                 $row = $this->db->selectRow( 'text',
 310                         array( 'old_text', 'old_flags' ),
 311                         array( 'old_id' => $id ),
 312                         'TextPassDumper::getText' );
 313                 $text = Revision::getRevisionText( $row );
 314                 if( $text === false ) {
 315                         return false;
 316                 }
 317                 $stripped = str_replace( "\r", "", $text );
 318                 $normalized = UtfNormal::cleanUp( $stripped );
 319                 return $normalized;
 320         }
 321
 322         private function getTextSpawned( $id ) {
 323                 wfSuppressWarnings();
 324                 if( !$this->spawnProc ) {
 325                         // First time?
 326                         $this->openSpawn();
 327                 }
 328                 while( true ) {
 329
 330                         $text = $this->getTextSpawnedOnce( $id );
 331                         if( !is_string( $text ) ) {
 332                                 $this->progress("Database subprocess failed. Respawning...");
 333
 334                                 $this->closeSpawn();
 335                                 sleep( $this->failureTimeout );
 336                                 $this->openSpawn();
 337
 338                                 continue;
 339                         }
 340                         wfRestoreWarnings();
 341                         return $text;
 342                 }
 343         }
 344
 345         function openSpawn() {
 346                 global $IP, $wgDBname;
 347
 348                 $cmd = implode( " ",
 349                         array_map( 'wfEscapeShellArg',
 350                                 array(
 351                                         $this->php,
 352                                         "$IP/maintenance/fetchText.php",
 353                                         $wgDBname ) ) );
 354                 $spec = array(
 355                         0 => array( "pipe", "r" ),
 356                         1 => array( "pipe", "w" ),
 357                         2 => array( "file", "/dev/null", "a" ) );
 358                 $pipes = array();
 359
 360                 $this->progress( "Spawning database subprocess: $cmd" );
 361                 $this->spawnProc = proc_open( $cmd, $spec, $pipes );
 362                 if( !$this->spawnProc ) {
 363                         // shit
 364                         $this->progress( "Subprocess spawn failed." );
 365                         return false;
 366                 }
 367                 list(
 368                         $this->spawnWrite, // -> stdin
 369                         $this->spawnRead,  // <- stdout
 370                 ) = $pipes;
 371
 372                 return true;
 373         }
 374
 375         private function closeSpawn() {
 376                 wfSuppressWarnings();
 377                 if( $this->spawnRead )
 378                         fclose( $this->spawnRead );
 379                 $this->spawnRead = false;
 380                 if( $this->spawnWrite )
 381                         fclose( $this->spawnWrite );
 382                 $this->spawnWrite = false;
 383                 if( $this->spawnErr )
 384                         fclose( $this->spawnErr );
 385                 $this->spawnErr = false;
 386                 if( $this->spawnProc )
 387                         pclose( $this->spawnProc );
 388                 $this->spawnProc = false;
 389                 wfRestoreWarnings();
 390         }
 391
 392         private function getTextSpawnedOnce( $id ) {
 393                 $ok = fwrite( $this->spawnWrite, "$id\n" );
 394                 //$this->progress( ">> $id" );
 395                 if( !$ok ) return false;
 396
 397                 $ok = fflush( $this->spawnWrite );
 398                 //$this->progress( ">> [flush]" );
 399                 if( !$ok ) return false;
 400
 401                 $len = fgets( $this->spawnRead );
 402                 //$this->progress( "<< " . trim( $len ) );
 403                 if( $len === false ) return false;
 404
 405                 $nbytes = intval( $len );
 406                 $text = "";
 407
 408                 // Subprocess may not send everything at once, we have to loop.
 409                 while( $nbytes > strlen( $text ) ) {
 410                         $buffer = fread( $this->spawnRead, $nbytes - strlen( $text ) );
 411                         if( $text === false ) break;
 412                         $text .= $buffer;
 413                 }
 414
 415                 $gotbytes = strlen( $text );
 416                 if( $gotbytes != $nbytes ) {
 417                         $this->progress( "Expected $nbytes bytes from database subprocess, got $gotbytes ");
 418                         return false;
 419                 }
 420
 421                 // Do normalization in the dump thread...
 422                 $stripped = str_replace( "\r", "", $text );
 423                 $normalized = UtfNormal::cleanUp( $stripped );
 424                 return $normalized;
 425         }
 426
 427         function startElement( $parser, $name, $attribs ) {
 428                 $this->clearOpenElement( null );
 429                 $this->lastName = $name;
 430
 431                 if( $name == 'revision' ) {
 432                         $this->state = $name;
 433                         $this->egress->writeOpenPage( null, $this->buffer );
 434                         $this->buffer = "";
 435                 } elseif( $name == 'page' ) {
 436                         $this->state = $name;
 437                         if( $this->atStart ) {
 438                                 $this->egress->writeOpenStream( $this->buffer );
 439                                 $this->buffer = "";
 440                                 $this->atStart = false;
 441                         }
 442                 }
 443
 444                 if( $name == "text" && isset( $attribs['id'] ) ) {
 445                         $text = $this->getText( $attribs['id'] );
 446                         $this->openElement = array( $name, array( 'xml:space' => 'preserve' ) );
 447                         if( strlen( $text ) > 0 ) {
 448                                 $this->characterData( $parser, $text );
 449                         }
 450                 } else {
 451                         $this->openElement = array( $name, $attribs );
 452                 }
 453         }
 454
 455         function endElement( $parser, $name ) {
 456                 if( $this->openElement ) {
 457                         $this->clearOpenElement( "" );
 458                 } else {
 459                         $this->buffer .= "</$name>";
 460                 }
 461
 462                 if( $name == 'revision' ) {
 463                         $this->egress->writeRevision( null, $this->buffer );
 464                         $this->buffer = "";
 465                         $this->thisRev = "";
 466                 } elseif( $name == 'page' ) {
 467                         $this->egress->writeClosePage( $this->buffer );
 468                         $this->buffer = "";
 469                         $this->thisPage = "";
 470                 } elseif( $name == 'mediawiki' ) {
 471                         $this->egress->writeCloseStream( $this->buffer );
 472                         $this->buffer = "";
 473                 }
 474         }
 475
 476         function characterData( $parser, $data ) {
 477                 $this->clearOpenElement( null );
 478                 if( $this->lastName == "id" ) {
 479                         if( $this->state == "revision" ) {
 480                                 $this->thisRev .= $data;
 481                         } elseif( $this->state == "page" ) {
 482                                 $this->thisPage .= $data;
 483                         }
 484                 }
 485                 $this->buffer .= htmlspecialchars( $data );
 486         }
 487
 488         function clearOpenElement( $style ) {
 489                 if( $this->openElement ) {
 490                         $this->buffer .= wfElement( $this->openElement[0], $this->openElement[1], $style );
 491                         $this->openElement = false;
 492                 }
 493         }
 494 }
 495
 496
 497 $dumper = new TextPassDumper( $argv );
 498
 499 if( true ) {
 500         $dumper->dump();
 501 } else {
 502         $dumper->progress( <<<END
 503 This script postprocesses XML dumps from dumpBackup.php to add
 504 page text which was stubbed out (using --stub).
 505
 506 XML input is accepted on stdin.
 507 XML output is sent to stdout; progress reports are sent to stderr.
 508
 509 Usage: php dumpTextPass.php [<options>]
 510 Options:
 511   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 512   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 513               pressure on the database.
 514               (Requires PHP 5.0+ and the XMLReader PECL extension)
 515   --quiet     Don't dump status reports to stderr.
 516   --report=n  Report position and speed after every n pages processed.
 517               (Default: 100)
 518   --server=h  Force reading from MySQL server h
 519   --current   Base ETA on number of pages in database instead of all revisions
 520   --spawn     Spawn a subprocess for loading text records
 521 END
 522 );
 523 }
 524
 525