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  * @package MediaWiki
  22  * @subpackage SpecialPage
  23  */
  24
  25 $originalDir = getcwd();
  26
  27 require_once( 'commandLine.inc' );
  28 require_once( 'SpecialExport.php' );
  29 require_once( 'maintenance/backup.inc' );
  30
  31 /**
  32  * Stream wrapper around 7za filter program.
  33  * Required since we can't pass an open file resource to XMLReader->open()
  34  * which is used for the text prefetch.
  35  */
  36 class SevenZipStream {
  37         var $stream;
  38
  39         private function stripPath( $path ) {
  40                 $prefix = 'mediawiki.compress.7z://';
  41                 return substr( $path, strlen( $prefix ) );
  42         }
  43
  44         function stream_open( $path, $mode, $options, &$opened_path ) {
  45                 if( $mode{0} == 'r' ) {
  46                         $options = 'e -bd -so';
  47                 } elseif( $mode{0} == 'w' ) {
  48                         $options = 'a -bd -si';
  49                 } else {
  50                         return false;
  51                 }
  52                 $arg = wfEscapeShellArg( $this->stripPath( $path ) );
  53                 $command = "7za $options $arg";
  54                 if( !wfIsWindows() ) {
  55                         // Suppress the stupid messages on stderr
  56                         $command .= ' 2>/dev/null';
  57                 }
  58                 $this->stream = popen( $command, $mode );
  59                 return ($this->stream !== false);
  60         }
  61
  62         function url_stat( $path, $flags ) {
  63                 return stat( $this->stripPath( $path ) );
  64         }
  65
  66         // This is all so lame; there should be a default class we can extend
  67
  68         function stream_close() {
  69                 return fclose( $this->stream );
  70         }
  71
  72         function stream_flush() {
  73                 return fflush( $this->stream );
  74         }
  75
  76         function stream_read( $count ) {
  77                 return fread( $this->stream, $count );
  78         }
  79
  80         function stream_write( $data ) {
  81                 return fwrite( $this->stream, $data );
  82         }
  83
  84         function stream_tell() {
  85                 return ftell( $this->stream );
  86         }
  87
  88         function stream_eof() {
  89                 return feof( $this->stream );
  90         }
  91
  92         function stream_seek( $offset, $whence ) {
  93                 return fseek( $this->stream, $offset, $whence );
  94         }
  95 }
  96 stream_wrapper_register( 'mediawiki.compress.7z', 'SevenZipStream' );
  97
  98
  99 class TextPassDumper extends BackupDumper {
 100         var $prefetch = null;
 101         var $input = "php://stdin";
 102         var $history = WikiExporter::FULL;
 103         var $fetchCount = 0;
 104         var $prefetchCount = 0;
 105
 106         var $failures = 0;
 107         var $maxFailures = 200;
 108         var $failureTimeout = 5; // Seconds to sleep after db failure
 109
 110         function dump() {
 111                 # This shouldn't happen if on console... ;)
 112                 header( 'Content-type: text/html; charset=UTF-8' );
 113
 114                 # Notice messages will foul up your XML output even if they're
 115                 # relatively harmless.
 116 //              ini_set( 'display_errors', false );
 117
 118                 $this->initProgress( $this->history );
 119
 120                 $this->db =& $this->backupDb();
 121
 122                 $this->egress = new ExportProgressFilter( $this->sink, $this );
 123
 124                 $input = fopen( $this->input, "rt" );
 125                 $result = $this->readDump( $input );
 126
 127                 if( WikiError::isError( $result ) ) {
 128                         wfDie( $result->getMessage() );
 129                 }
 130
 131                 $this->report( true );
 132         }
 133
 134         function processOption( $opt, $val, $param ) {
 135                 $url = $this->processFileOpt( $val, $param );
 136
 137                 switch( $opt ) {
 138                 case 'prefetch':
 139                         require_once 'maintenance/backupPrefetch.inc';
 140                         $this->prefetch = new BaseDump( $url );
 141                         break;
 142                 case 'stub':
 143                         $this->input = $url;
 144                         break;
 145                 case 'current':
 146                         $this->history = WikiExporter::CURRENT;
 147                         break;
 148                 case 'full':
 149                         $this->history = WikiExporter::FULL;
 150                         break;
 151                 }
 152         }
 153
 154         function processFileOpt( $val, $param ) {
 155                 switch( $val ) {
 156                 case "file":
 157                         return $param;
 158                 case "gzip":
 159                         return "compress.zlib://$param";
 160                 case "bzip2":
 161                         return "compress.bzip2://$param";
 162                 case "7zip":
 163                         return "mediawiki.compress.7z://$param";
 164                 default:
 165                         return $val;
 166                 }
 167         }
 168
 169         /**
 170          * Overridden to include prefetch ratio if enabled.
 171          */
 172         function showReport() {
 173                 if( !$this->prefetch ) {
 174                         return parent::showReport();
 175                 }
 176
 177                 if( $this->reporting ) {
 178                         $delta = wfTime() - $this->startTime;
 179                         $now = wfTimestamp( TS_DB );
 180                         if( $delta ) {
 181                                 $rate = $this->pageCount / $delta;
 182                                 $revrate = $this->revCount / $delta;
 183                                 $portion = $this->revCount / $this->maxCount;
 184                                 $eta = $this->startTime + $delta / $portion;
 185                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 186                                 $fetchrate = 100.0 * $this->prefetchCount / $this->fetchCount;
 187                         } else {
 188                                 $rate = '-';
 189                                 $revrate = '-';
 190                                 $etats = '-';
 191                                 $fetchrate = '-';
 192                         }
 193                         global $wgDBname;
 194                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), %0.1f%% prefetched, ETA %s [max %d]",
 195                                 $now, $wgDBname, $this->pageCount, $rate, $this->revCount, $revrate, $fetchrate, $etats, $this->maxCount ) );
 196                 }
 197         }
 198
 199         function readDump( $input ) {
 200                 $this->buffer = "";
 201                 $this->openElement = false;
 202                 $this->atStart = true;
 203                 $this->state = "";
 204                 $this->lastName = "";
 205                 $this->thisPage = 0;
 206                 $this->thisRev = 0;
 207
 208                 $parser = xml_parser_create( "UTF-8" );
 209                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 210
 211                 xml_set_element_handler( $parser, array( &$this, 'startElement' ), array( &$this, 'endElement' ) );
 212                 xml_set_character_data_handler( $parser, array( &$this, 'characterData' ) );
 213
 214                 $offset = 0; // for context extraction on error reporting
 215                 $bufferSize = 512 * 1024;
 216                 do {
 217                         $chunk = fread( $input, $bufferSize );
 218                         if( !xml_parse( $parser, $chunk, feof( $input ) ) ) {
 219                                 wfDebug( "TextDumpPass::readDump encountered XML parsing error\n" );
 220                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 221                         }
 222                         $offset += strlen( $chunk );
 223                 } while( $chunk !== false && !feof( $input ) );
 224                 xml_parser_free( $parser );
 225
 226                 return true;
 227         }
 228
 229         function getText( $id ) {
 230                 $this->fetchCount++;
 231                 if( isset( $this->prefetch ) ) {
 232                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 233                         if( $text === null ) {
 234                                 // Entry missing from prefetch dump
 235                         } elseif( $text === "" ) {
 236                                 // Blank entries may indicate that the prior dump was broken.
 237                                 // To be safe, reload it.
 238                         } else {
 239                                 $this->prefetchCount++;
 240                                 return $text;
 241                         }
 242                 }
 243                 while( true ) {
 244                         try {
 245                                 return $this->doGetText( $id );
 246                         } catch (DBQueryError $ex) {
 247                                 $this->failures++;
 248                                 if( $this->failures > $this->maxFailures ) {
 249                                         throw $ex;
 250                                 } else {
 251                                         $this->progress( "Database failure $this->failures " .
 252                                                 "of allowed $this->maxFailures! " .
 253                                                 "Pausing $this->failureTimeout seconds..." );
 254                                         sleep( $this->failureTimeout );
 255                                 }
 256                         }
 257                 }
 258         }
 259
 260         /**
 261          * May throw a database error if, say, the server dies during query.
 262          */
 263         private function doGetText( $id ) {
 264                 $id = intval( $id );
 265                 $row = $this->db->selectRow( 'text',
 266                         array( 'old_text', 'old_flags' ),
 267                         array( 'old_id' => $id ),
 268                         'TextPassDumper::getText' );
 269                 $text = Revision::getRevisionText( $row );
 270                 $stripped = str_replace( "\r", "", $text );
 271                 $normalized = UtfNormal::cleanUp( $stripped );
 272                 return $normalized;
 273         }
 274
 275         function startElement( $parser, $name, $attribs ) {
 276                 $this->clearOpenElement( null );
 277                 $this->lastName = $name;
 278
 279                 if( $name == 'revision' ) {
 280                         $this->state = $name;
 281                         $this->egress->writeOpenPage( null, $this->buffer );
 282                         $this->buffer = "";
 283                 } elseif( $name == 'page' ) {
 284                         $this->state = $name;
 285                         if( $this->atStart ) {
 286                                 $this->egress->writeOpenStream( $this->buffer );
 287                                 $this->buffer = "";
 288                                 $this->atStart = false;
 289                         }
 290                 }
 291
 292                 if( $name == "text" && isset( $attribs['id'] ) ) {
 293                         $text = $this->getText( $attribs['id'] );
 294                         $this->openElement = array( $name, array( 'xml:space' => 'preserve' ) );
 295                         if( strlen( $text ) > 0 ) {
 296                                 $this->characterData( $parser, $text );
 297                         }
 298                 } else {
 299                         $this->openElement = array( $name, $attribs );
 300                 }
 301         }
 302
 303         function endElement( $parser, $name ) {
 304                 if( $this->openElement ) {
 305                         $this->clearOpenElement( "" );
 306                 } else {
 307                         $this->buffer .= "</$name>";
 308                 }
 309
 310                 if( $name == 'revision' ) {
 311                         $this->egress->writeRevision( null, $this->buffer );
 312                         $this->buffer = "";
 313                         $this->thisRev = "";
 314                 } elseif( $name == 'page' ) {
 315                         $this->egress->writeClosePage( $this->buffer );
 316                         $this->buffer = "";
 317                         $this->thisPage = "";
 318                 } elseif( $name == 'mediawiki' ) {
 319                         $this->egress->writeCloseStream( $this->buffer );
 320                         $this->buffer = "";
 321                 }
 322         }
 323
 324         function characterData( $parser, $data ) {
 325                 $this->clearOpenElement( null );
 326                 if( $this->lastName == "id" ) {
 327                         if( $this->state == "revision" ) {
 328                                 $this->thisRev .= $data;
 329                         } elseif( $this->state == "page" ) {
 330                                 $this->thisPage .= $data;
 331                         }
 332                 }
 333                 $this->buffer .= htmlspecialchars( $data );
 334         }
 335
 336         function clearOpenElement( $style ) {
 337                 if( $this->openElement ) {
 338                         $this->buffer .= wfElement( $this->openElement[0], $this->openElement[1], $style );
 339                         $this->openElement = false;
 340                 }
 341         }
 342 }
 343
 344
 345 $dumper = new TextPassDumper( $argv );
 346
 347 if( true ) {
 348         $dumper->dump();
 349 } else {
 350         $dumper->progress( <<<END
 351 This script postprocesses XML dumps from dumpBackup.php to add
 352 page text which was stubbed out (using --stub).
 353
 354 XML input is accepted on stdin.
 355 XML output is sent to stdout; progress reports are sent to stderr.
 356
 357 Usage: php dumpTextPass.php [<options>]
 358 Options:
 359   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 360   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 361               pressure on the database.
 362               (Requires PHP 5.0+ and the XMLReader PECL extension)
 363   --quiet     Don't dump status reports to stderr.
 364   --report=n  Report position and speed after every n pages processed.
 365               (Default: 100)
 366   --server=h  Force reading from MySQL server h
 367   --current   Base ETA on number of pages in database instead of all revisions
 368 END
 369 );
 370 }
 371
 372 ?>