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 = MW_EXPORT_FULL;
 103         var $fetchCount = 0;
 104         var $prefetchCount = 0;
 105
 106         function dump() {
 107                 # This shouldn't happen if on console... ;)
 108                 header( 'Content-type: text/html; charset=UTF-8' );
 109
 110                 # Notice messages will foul up your XML output even if they're
 111                 # relatively harmless.
 112 //              ini_set( 'display_errors', false );
 113
 114                 $this->initProgress( $this->history );
 115
 116                 $this->db =& $this->backupDb();
 117
 118                 $this->egress = new ExportProgressFilter( $this->sink, $this );
 119
 120                 $input = fopen( $this->input, "rt" );
 121                 $result = $this->readDump( $input );
 122
 123                 if( WikiError::isError( $result ) ) {
 124                         wfDie( $result->getMessage() );
 125                 }
 126
 127                 $this->report( true );
 128         }
 129
 130         function processOption( $opt, $val, $param ) {
 131                 $url = $this->processFileOpt( $val, $param );
 132
 133                 switch( $opt ) {
 134                 case 'prefetch':
 135                         require_once 'maintenance/backupPrefetch.inc';
 136                         $this->prefetch = new BaseDump( $url );
 137                         break;
 138                 case 'stub':
 139                         $this->input = $url;
 140                         break;
 141                 case 'current':
 142                         $this->history = MW_EXPORT_CURRENT;
 143                         break;
 144                 case 'full':
 145                         $this->history = MW_EXPORT_FULL;
 146                         break;
 147                 }
 148         }
 149
 150         function processFileOpt( $val, $param ) {
 151                 switch( $val ) {
 152                 case "file":
 153                         return $param;
 154                 case "gzip":
 155                         return "compress.zlib://$param";
 156                 case "bzip2":
 157                         return "compress.bzip2://$param";
 158                 case "7zip":
 159                         return "mediawiki.compress.7z://$param";
 160                 default:
 161                         return $val;
 162                 }
 163         }
 164
 165         /**
 166          * Overridden to include prefetch ratio if enabled.
 167          */
 168         function showReport() {
 169                 if( !$this->prefetch ) {
 170                         return parent::showReport();
 171                 }
 172
 173                 if( $this->reporting ) {
 174                         $delta = wfTime() - $this->startTime;
 175                         $now = wfTimestamp( TS_DB );
 176                         if( $delta ) {
 177                                 $rate = $this->pageCount / $delta;
 178                                 $revrate = $this->revCount / $delta;
 179                                 $portion = $this->revCount / $this->maxCount;
 180                                 $eta = $this->startTime + $delta / $portion;
 181                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 182                                 $fetchrate = 100.0 * $this->prefetchCount / $this->fetchCount;
 183                         } else {
 184                                 $rate = '-';
 185                                 $revrate = '-';
 186                                 $etats = '-';
 187                                 $fetchrate = '-';
 188                         }
 189                         global $wgDBname;
 190                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), %0.1f%% prefetched, ETA %s [max %d]",
 191                                 $now, $wgDBname, $this->pageCount, $rate, $this->revCount, $revrate, $fetchrate, $etats, $this->maxCount ) );
 192                 }
 193         }
 194
 195         function readDump( $input ) {
 196                 $this->buffer = "";
 197                 $this->openElement = false;
 198                 $this->atStart = true;
 199                 $this->state = "";
 200                 $this->lastName = "";
 201                 $this->thisPage = 0;
 202                 $this->thisRev = 0;
 203
 204                 $parser = xml_parser_create( "UTF-8" );
 205                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 206
 207                 xml_set_element_handler( $parser, array( &$this, 'startElement' ), array( &$this, 'endElement' ) );
 208                 xml_set_character_data_handler( $parser, array( &$this, 'characterData' ) );
 209
 210                 $offset = 0; // for context extraction on error reporting
 211                 $bufferSize = 512 * 1024;
 212                 do {
 213                         $chunk = fread( $input, $bufferSize );
 214                         if( !xml_parse( $parser, $chunk, feof( $input ) ) ) {
 215                                 wfDebug( "TextDumpPass::readDump encountered XML parsing error\n" );
 216                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 217                         }
 218                         $offset += strlen( $chunk );
 219                 } while( $chunk !== false && !feof( $input ) );
 220                 xml_parser_free( $parser );
 221
 222                 return true;
 223         }
 224
 225         function getText( $id ) {
 226                 $this->fetchCount++;
 227                 if( isset( $this->prefetch ) ) {
 228                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 229                         if( $text === null ) {
 230                                 // Entry missing from prefetch dump
 231                         } elseif( $text === "" ) {
 232                                 // Blank entries may indicate that the prior dump was broken.
 233                                 // To be safe, reload it.
 234                         } else {
 235                                 $this->prefetchCount++;
 236                                 return $text;
 237                         }
 238                 }
 239                 $id = intval( $id );
 240                 $row = $this->db->selectRow( 'text',
 241                         array( 'old_text', 'old_flags' ),
 242                         array( 'old_id' => $id ),
 243                         'TextPassDumper::getText' );
 244                 $text = Revision::getRevisionText( $row );
 245                 $stripped = str_replace( "\r", "", $text );
 246                 $normalized = UtfNormal::cleanUp( $stripped );
 247                 return $normalized;
 248         }
 249
 250         function startElement( $parser, $name, $attribs ) {
 251                 $this->clearOpenElement( null );
 252                 $this->lastName = $name;
 253
 254                 if( $name == 'revision' ) {
 255                         $this->state = $name;
 256                         $this->egress->writeOpenPage( null, $this->buffer );
 257                         $this->buffer = "";
 258                 } elseif( $name == 'page' ) {
 259                         $this->state = $name;
 260                         if( $this->atStart ) {
 261                                 $this->egress->writeOpenStream( $this->buffer );
 262                                 $this->buffer = "";
 263                                 $this->atStart = false;
 264                         }
 265                 }
 266
 267                 if( $name == "text" && isset( $attribs['id'] ) ) {
 268                         $text = $this->getText( $attribs['id'] );
 269                         $this->openElement = array( $name, array( 'xml:space' => 'preserve' ) );
 270                         if( strlen( $text ) > 0 ) {
 271                                 $this->characterData( $parser, $text );
 272                         }
 273                 } else {
 274                         $this->openElement = array( $name, $attribs );
 275                 }
 276         }
 277
 278         function endElement( $parser, $name ) {
 279                 if( $this->openElement ) {
 280                         $this->clearOpenElement( "" );
 281                 } else {
 282                         $this->buffer .= "</$name>";
 283                 }
 284
 285                 if( $name == 'revision' ) {
 286                         $this->egress->writeRevision( null, $this->buffer );
 287                         $this->buffer = "";
 288                         $this->thisRev = "";
 289                 } elseif( $name == 'page' ) {
 290                         $this->egress->writeClosePage( $this->buffer );
 291                         $this->buffer = "";
 292                         $this->thisPage = "";
 293                 } elseif( $name == 'mediawiki' ) {
 294                         $this->egress->writeCloseStream( $this->buffer );
 295                         $this->buffer = "";
 296                 }
 297         }
 298
 299         function characterData( $parser, $data ) {
 300                 $this->clearOpenElement( null );
 301                 if( $this->lastName == "id" ) {
 302                         if( $this->state == "revision" ) {
 303                                 $this->thisRev .= $data;
 304                         } elseif( $this->state == "page" ) {
 305                                 $this->thisPage .= $data;
 306                         }
 307                 }
 308                 $this->buffer .= htmlspecialchars( $data );
 309         }
 310
 311         function clearOpenElement( $style ) {
 312                 if( $this->openElement ) {
 313                         $this->buffer .= wfElement( $this->openElement[0], $this->openElement[1], $style );
 314                         $this->openElement = false;
 315                 }
 316         }
 317 }
 318
 319
 320 $dumper = new TextPassDumper( $argv );
 321
 322 if( true ) {
 323         $dumper->dump();
 324 } else {
 325         $dumper->progress( <<<END
 326 This script postprocesses XML dumps from dumpBackup.php to add
 327 page text which was stubbed out (using --stub).
 328
 329 XML input is accepted on stdin.
 330 XML output is sent to stdout; progress reports are sent to stderr.
 331
 332 Usage: php dumpTextPass.php [<options>]
 333 Options:
 334   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 335   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 336               pressure on the database.
 337               (Requires PHP 5.0+ and the XMLReader PECL extension)
 338   --quiet     Don't dump status reports to stderr.
 339   --report=n  Report position and speed after every n pages processed.
 340               (Default: 100)
 341   --server=h  Force reading from MySQL server h
 342   --current   Base ETA on number of pages in database instead of all revisions
 343 END
 344 );
 345 }
 346
 347 ?>