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  * @addtogroup SpecialPage
  22  */
  23
  24 $originalDir = getcwd();
  25
  26 require_once( 'commandLine.inc' );
  27 require_once( 'backup.inc' );
  28
  29 /**
  30  * Stream wrapper around 7za filter program.
  31  * Required since we can't pass an open file resource to XMLReader->open()
  32  * which is used for the text prefetch.
  33  */
  34 class SevenZipStream {
  35         var $stream;
  36
  37         private function stripPath( $path ) {
  38                 $prefix = 'mediawiki.compress.7z://';
  39                 return substr( $path, strlen( $prefix ) );
  40         }
  41
  42         function stream_open( $path, $mode, $options, &$opened_path ) {
  43                 if( $mode{0} == 'r' ) {
  44                         $options = 'e -bd -so';
  45                 } elseif( $mode{0} == 'w' ) {
  46                         $options = 'a -bd -si';
  47                 } else {
  48                         return false;
  49                 }
  50                 $arg = wfEscapeShellArg( $this->stripPath( $path ) );
  51                 $command = "7za $options $arg";
  52                 if( !wfIsWindows() ) {
  53                         // Suppress the stupid messages on stderr
  54                         $command .= ' 2>/dev/null';
  55                 }
  56                 $this->stream = popen( $command, $mode );
  57                 return ($this->stream !== false);
  58         }
  59
  60         function url_stat( $path, $flags ) {
  61                 return stat( $this->stripPath( $path ) );
  62         }
  63
  64         // This is all so lame; there should be a default class we can extend
  65
  66         function stream_close() {
  67                 return fclose( $this->stream );
  68         }
  69
  70         function stream_flush() {
  71                 return fflush( $this->stream );
  72         }
  73
  74         function stream_read( $count ) {
  75                 return fread( $this->stream, $count );
  76         }
  77
  78         function stream_write( $data ) {
  79                 return fwrite( $this->stream, $data );
  80         }
  81
  82         function stream_tell() {
  83                 return ftell( $this->stream );
  84         }
  85
  86         function stream_eof() {
  87                 return feof( $this->stream );
  88         }
  89
  90         function stream_seek( $offset, $whence ) {
  91                 return fseek( $this->stream, $offset, $whence );
  92         }
  93 }
  94 stream_wrapper_register( 'mediawiki.compress.7z', 'SevenZipStream' );
  95
  96
  97 class TextPassDumper extends BackupDumper {
  98         var $prefetch = null;
  99         var $input = "php://stdin";
 100         var $history = WikiExporter::FULL;
 101         var $fetchCount = 0;
 102         var $prefetchCount = 0;
 103
 104         var $failures = 0;
 105         var $maxFailures = 200;
 106         var $failureTimeout = 5; // Seconds to sleep after db failure
 107
 108         var $php = "php";
 109         var $spawn = false;
 110         var $spawnProc = false;
 111         var $spawnWrite = false;
 112         var $spawnRead = false;
 113         var $spawnErr = false;
 114
 115         function dump() {
 116                 # This shouldn't happen if on console... ;)
 117                 header( 'Content-type: text/html; charset=UTF-8' );
 118
 119                 # Notice messages will foul up your XML output even if they're
 120                 # relatively harmless.
 121                 if( ini_get( 'display_errors' ) )
 122                         ini_set( 'display_errors', 'stderr' );
 123
 124                 $this->initProgress( $this->history );
 125
 126                 $this->db = $this->backupDb();
 127
 128                 $this->egress = new ExportProgressFilter( $this->sink, $this );
 129
 130                 $input = fopen( $this->input, "rt" );
 131                 $result = $this->readDump( $input );
 132
 133                 if( WikiError::isError( $result ) ) {
 134                         wfDie( $result->getMessage() );
 135                 }
 136
 137                 if( $this->spawnProc ) {
 138                         $this->closeSpawn();
 139                 }
 140
 141                 $this->report( true );
 142         }
 143
 144         function processOption( $opt, $val, $param ) {
 145                 $url = $this->processFileOpt( $val, $param );
 146
 147                 switch( $opt ) {
 148                 case 'prefetch':
 149                         require_once 'maintenance/backupPrefetch.inc';
 150                         $this->prefetch = new BaseDump( $url );
 151                         break;
 152                 case 'stub':
 153                         $this->input = $url;
 154                         break;
 155                 case 'current':
 156                         $this->history = WikiExporter::CURRENT;
 157                         break;
 158                 case 'full':
 159                         $this->history = WikiExporter::FULL;
 160                         break;
 161                 case 'spawn':
 162                         $this->spawn = true;
 163                         if( $val ) {
 164                                 $this->php = $val;
 165                         }
 166                         break;
 167                 }
 168         }
 169
 170         function processFileOpt( $val, $param ) {
 171                 switch( $val ) {
 172                 case "file":
 173                         return $param;
 174                 case "gzip":
 175                         return "compress.zlib://$param";
 176                 case "bzip2":
 177                         return "compress.bzip2://$param";
 178                 case "7zip":
 179                         return "mediawiki.compress.7z://$param";
 180                 default:
 181                         return $val;
 182                 }
 183         }
 184
 185         /**
 186          * Overridden to include prefetch ratio if enabled.
 187          */
 188         function showReport() {
 189                 if( !$this->prefetch ) {
 190                         return parent::showReport();
 191                 }
 192
 193                 if( $this->reporting ) {
 194                         $delta = wfTime() - $this->startTime;
 195                         $now = wfTimestamp( TS_DB );
 196                         if( $delta ) {
 197                                 $rate = $this->pageCount / $delta;
 198                                 $revrate = $this->revCount / $delta;
 199                                 $portion = $this->revCount / $this->maxCount;
 200                                 $eta = $this->startTime + $delta / $portion;
 201                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 202                                 $fetchrate = 100.0 * $this->prefetchCount / $this->fetchCount;
 203                         } else {
 204                                 $rate = '-';
 205                                 $revrate = '-';
 206                                 $etats = '-';
 207                                 $fetchrate = '-';
 208                         }
 209                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), %0.1f%% prefetched, ETA %s [max %d]",
 210                                 $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $fetchrate, $etats, $this->maxCount ) );
 211                 }
 212         }
 213
 214         function readDump( $input ) {
 215                 $this->buffer = "";
 216                 $this->openElement = false;
 217                 $this->atStart = true;
 218                 $this->state = "";
 219                 $this->lastName = "";
 220                 $this->thisPage = 0;
 221                 $this->thisRev = 0;
 222
 223                 $parser = xml_parser_create( "UTF-8" );
 224                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 225
 226                 xml_set_element_handler( $parser, array( &$this, 'startElement' ), array( &$this, 'endElement' ) );
 227                 xml_set_character_data_handler( $parser, array( &$this, 'characterData' ) );
 228
 229                 $offset = 0; // for context extraction on error reporting
 230                 $bufferSize = 512 * 1024;
 231                 do {
 232                         $chunk = fread( $input, $bufferSize );
 233                         if( !xml_parse( $parser, $chunk, feof( $input ) ) ) {
 234                                 wfDebug( "TextDumpPass::readDump encountered XML parsing error\n" );
 235                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 236                         }
 237                         $offset += strlen( $chunk );
 238                 } while( $chunk !== false && !feof( $input ) );
 239                 xml_parser_free( $parser );
 240
 241                 return true;
 242         }
 243
 244         function getText( $id ) {
 245                 $this->fetchCount++;
 246                 if( isset( $this->prefetch ) ) {
 247                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 248                         if( $text === null ) {
 249                                 // Entry missing from prefetch dump
 250                         } elseif( $text === "" ) {
 251                                 // Blank entries may indicate that the prior dump was broken.
 252                                 // To be safe, reload it.
 253                         } else {
 254                                 $this->prefetchCount++;
 255                                 return $text;
 256                         }
 257                 }
 258                 return $this->doGetText( $id );
 259         }
 260
 261         private function doGetText( $id ) {
 262                 if( $this->spawn ) {
 263                         return $this->getTextSpawned( $id );
 264                 } else {
 265                         return $this->getTextDbSafe( $id );
 266                 }
 267         }
 268
 269         /**
 270          * Fetch a text revision from the database, retrying in case of failure.
 271          * This may survive some transitory errors by reconnecting, but
 272          * may not survive a long-term server outage.
 273          */
 274         private function getTextDbSafe( $id ) {
 275                 while( true ) {
 276                         try {
 277                                 $text = $this->getTextDb( $id );
 278                                 $ex = new MWException("Graceful storage failure");
 279                         } catch (DBQueryError $ex) {
 280                                 $text = false;
 281                         }
 282                         if( $text === false ) {
 283                                 $this->failures++;
 284                                 if( $this->failures > $this->maxFailures ) {
 285                                         throw $ex;
 286                                 } else {
 287                                         $this->progress( "Database failure $this->failures " .
 288                                                 "of allowed $this->maxFailures for revision $id! " .
 289                                                 "Pausing $this->failureTimeout seconds..." );
 290                                         sleep( $this->failureTimeout );
 291                                 }
 292                         } else {
 293                                 return $text;
 294                         }
 295                 }
 296         }
 297
 298         /**
 299          * May throw a database error if, say, the server dies during query.
 300          */
 301         private function getTextDb( $id ) {
 302                 $id = intval( $id );
 303                 $row = $this->db->selectRow( 'text',
 304                         array( 'old_text', 'old_flags' ),
 305                         array( 'old_id' => $id ),
 306                         'TextPassDumper::getText' );
 307                 $text = Revision::getRevisionText( $row );
 308                 if( $text === false ) {
 309                         return false;
 310                 }
 311                 $stripped = str_replace( "\r", "", $text );
 312                 $normalized = UtfNormal::cleanUp( $stripped );
 313                 return $normalized;
 314         }
 315
 316         private function getTextSpawned( $id ) {
 317                 wfSuppressWarnings();
 318                 if( !$this->spawnProc ) {
 319                         // First time?
 320                         $this->openSpawn();
 321                 }
 322                 while( true ) {
 323
 324                         $text = $this->getTextSpawnedOnce( $id );
 325                         if( !is_string( $text ) ) {
 326                                 $this->progress("Database subprocess failed. Respawning...");
 327
 328                                 $this->closeSpawn();
 329                                 sleep( $this->failureTimeout );
 330                                 $this->openSpawn();
 331
 332                                 continue;
 333                         }
 334                         wfRestoreWarnings();
 335                         return $text;
 336                 }
 337         }
 338
 339         function openSpawn() {
 340                 global $IP, $wgDBname;
 341
 342                 $cmd = implode( " ",
 343                         array_map( 'wfEscapeShellArg',
 344                                 array(
 345                                         $this->php,
 346                                         "$IP/maintenance/fetchText.php",
 347                                         $wgDBname ) ) );
 348                 $spec = array(
 349                         0 => array( "pipe", "r" ),
 350                         1 => array( "pipe", "w" ),
 351                         2 => array( "file", "/dev/null", "a" ) );
 352                 $pipes = array();
 353
 354                 $this->progress( "Spawning database subprocess: $cmd" );
 355                 $this->spawnProc = proc_open( $cmd, $spec, $pipes );
 356                 if( !$this->spawnProc ) {
 357                         // shit
 358                         $this->progress( "Subprocess spawn failed." );
 359                         return false;
 360                 }
 361                 list(
 362                         $this->spawnWrite, // -> stdin
 363                         $this->spawnRead,  // <- stdout
 364                 ) = $pipes;
 365
 366                 return true;
 367         }
 368
 369         private function closeSpawn() {
 370                 wfSuppressWarnings();
 371                 if( $this->spawnRead )
 372                         fclose( $this->spawnRead );
 373                 $this->spawnRead = false;
 374                 if( $this->spawnWrite )
 375                         fclose( $this->spawnWrite );
 376                 $this->spawnWrite = false;
 377                 if( $this->spawnErr )
 378                         fclose( $this->spawnErr );
 379                 $this->spawnErr = false;
 380                 if( $this->spawnProc )
 381                         pclose( $this->spawnProc );
 382                 $this->spawnProc = false;
 383                 wfRestoreWarnings();
 384         }
 385
 386         private function getTextSpawnedOnce( $id ) {
 387                 $ok = fwrite( $this->spawnWrite, "$id\n" );
 388                 //$this->progress( ">> $id" );
 389                 if( !$ok ) return false;
 390
 391                 $ok = fflush( $this->spawnWrite );
 392                 //$this->progress( ">> [flush]" );
 393                 if( !$ok ) return false;
 394
 395                 $len = fgets( $this->spawnRead );
 396                 //$this->progress( "<< " . trim( $len ) );
 397                 if( $len === false ) return false;
 398
 399                 $nbytes = intval( $len );
 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( $text === 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 = UtfNormal::cleanUp( $stripped );
 418                 return $normalized;
 419         }
 420
 421         function startElement( $parser, $name, $attribs ) {
 422                 $this->clearOpenElement( null );
 423                 $this->lastName = $name;
 424
 425                 if( $name == 'revision' ) {
 426                         $this->state = $name;
 427                         $this->egress->writeOpenPage( null, $this->buffer );
 428                         $this->buffer = "";
 429                 } elseif( $name == 'page' ) {
 430                         $this->state = $name;
 431                         if( $this->atStart ) {
 432                                 $this->egress->writeOpenStream( $this->buffer );
 433                                 $this->buffer = "";
 434                                 $this->atStart = false;
 435                         }
 436                 }
 437
 438                 if( $name == "text" && isset( $attribs['id'] ) ) {
 439                         $text = $this->getText( $attribs['id'] );
 440                         $this->openElement = array( $name, array( 'xml:space' => 'preserve' ) );
 441                         if( strlen( $text ) > 0 ) {
 442                                 $this->characterData( $parser, $text );
 443                         }
 444                 } else {
 445                         $this->openElement = array( $name, $attribs );
 446                 }
 447         }
 448
 449         function endElement( $parser, $name ) {
 450                 if( $this->openElement ) {
 451                         $this->clearOpenElement( "" );
 452                 } else {
 453                         $this->buffer .= "</$name>";
 454                 }
 455
 456                 if( $name == 'revision' ) {
 457                         $this->egress->writeRevision( null, $this->buffer );
 458                         $this->buffer = "";
 459                         $this->thisRev = "";
 460                 } elseif( $name == 'page' ) {
 461                         $this->egress->writeClosePage( $this->buffer );
 462                         $this->buffer = "";
 463                         $this->thisPage = "";
 464                 } elseif( $name == 'mediawiki' ) {
 465                         $this->egress->writeCloseStream( $this->buffer );
 466                         $this->buffer = "";
 467                 }
 468         }
 469
 470         function characterData( $parser, $data ) {
 471                 $this->clearOpenElement( null );
 472                 if( $this->lastName == "id" ) {
 473                         if( $this->state == "revision" ) {
 474                                 $this->thisRev .= $data;
 475                         } elseif( $this->state == "page" ) {
 476                                 $this->thisPage .= $data;
 477                         }
 478                 }
 479                 $this->buffer .= htmlspecialchars( $data );
 480         }
 481
 482         function clearOpenElement( $style ) {
 483                 if( $this->openElement ) {
 484                         $this->buffer .= wfElement( $this->openElement[0], $this->openElement[1], $style );
 485                         $this->openElement = false;
 486                 }
 487         }
 488 }
 489
 490
 491 $dumper = new TextPassDumper( $argv );
 492
 493 if( true ) {
 494         $dumper->dump();
 495 } else {
 496         $dumper->progress( <<<END
 497 This script postprocesses XML dumps from dumpBackup.php to add
 498 page text which was stubbed out (using --stub).
 499
 500 XML input is accepted on stdin.
 501 XML output is sent to stdout; progress reports are sent to stderr.
 502
 503 Usage: php dumpTextPass.php [<options>]
 504 Options:
 505   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 506   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 507               pressure on the database.
 508               (Requires PHP 5.0+ and the XMLReader PECL extension)
 509   --quiet     Don't dump status reports to stderr.
 510   --report=n  Report position and speed after every n pages processed.
 511               (Default: 100)
 512   --server=h  Force reading from MySQL server h
 513   --current   Base ETA on number of pages in database instead of all revisions
 514   --spawn     Spawn a subprocess for loading text records
 515 END
 516 );
 517 }
 518
 519