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