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         function dump() {
 109                 # This shouldn't happen if on console... ;)
 110                 header( 'Content-type: text/html; charset=UTF-8' );
 111
 112                 # Notice messages will foul up your XML output even if they're
 113                 # relatively harmless.
 114 //              ini_set( 'display_errors', false );
 115
 116                 $this->initProgress( $this->history );
 117
 118                 $this->db = $this->backupDb();
 119
 120                 $this->egress = new ExportProgressFilter( $this->sink, $this );
 121
 122                 $input = fopen( $this->input, "rt" );
 123                 $result = $this->readDump( $input );
 124
 125                 if( WikiError::isError( $result ) ) {
 126                         wfDie( $result->getMessage() );
 127                 }
 128
 129                 $this->report( true );
 130         }
 131
 132         function processOption( $opt, $val, $param ) {
 133                 $url = $this->processFileOpt( $val, $param );
 134
 135                 switch( $opt ) {
 136                 case 'prefetch':
 137                         require_once 'maintenance/backupPrefetch.inc';
 138                         $this->prefetch = new BaseDump( $url );
 139                         break;
 140                 case 'stub':
 141                         $this->input = $url;
 142                         break;
 143                 case 'current':
 144                         $this->history = WikiExporter::CURRENT;
 145                         break;
 146                 case 'full':
 147                         $this->history = WikiExporter::FULL;
 148                         break;
 149                 }
 150         }
 151
 152         function processFileOpt( $val, $param ) {
 153                 switch( $val ) {
 154                 case "file":
 155                         return $param;
 156                 case "gzip":
 157                         return "compress.zlib://$param";
 158                 case "bzip2":
 159                         return "compress.bzip2://$param";
 160                 case "7zip":
 161                         return "mediawiki.compress.7z://$param";
 162                 default:
 163                         return $val;
 164                 }
 165         }
 166
 167         /**
 168          * Overridden to include prefetch ratio if enabled.
 169          */
 170         function showReport() {
 171                 if( !$this->prefetch ) {
 172                         return parent::showReport();
 173                 }
 174
 175                 if( $this->reporting ) {
 176                         $delta = wfTime() - $this->startTime;
 177                         $now = wfTimestamp( TS_DB );
 178                         if( $delta ) {
 179                                 $rate = $this->pageCount / $delta;
 180                                 $revrate = $this->revCount / $delta;
 181                                 $portion = $this->revCount / $this->maxCount;
 182                                 $eta = $this->startTime + $delta / $portion;
 183                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 184                                 $fetchrate = 100.0 * $this->prefetchCount / $this->fetchCount;
 185                         } else {
 186                                 $rate = '-';
 187                                 $revrate = '-';
 188                                 $etats = '-';
 189                                 $fetchrate = '-';
 190                         }
 191                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), %0.1f%% prefetched, ETA %s [max %d]",
 192                                 $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $fetchrate, $etats, $this->maxCount ) );
 193                 }
 194         }
 195
 196         function readDump( $input ) {
 197                 $this->buffer = "";
 198                 $this->openElement = false;
 199                 $this->atStart = true;
 200                 $this->state = "";
 201                 $this->lastName = "";
 202                 $this->thisPage = 0;
 203                 $this->thisRev = 0;
 204
 205                 $parser = xml_parser_create( "UTF-8" );
 206                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 207
 208                 xml_set_element_handler( $parser, array( &$this, 'startElement' ), array( &$this, 'endElement' ) );
 209                 xml_set_character_data_handler( $parser, array( &$this, 'characterData' ) );
 210
 211                 $offset = 0; // for context extraction on error reporting
 212                 $bufferSize = 512 * 1024;
 213                 do {
 214                         $chunk = fread( $input, $bufferSize );
 215                         if( !xml_parse( $parser, $chunk, feof( $input ) ) ) {
 216                                 wfDebug( "TextDumpPass::readDump encountered XML parsing error\n" );
 217                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 218                         }
 219                         $offset += strlen( $chunk );
 220                 } while( $chunk !== false && !feof( $input ) );
 221                 xml_parser_free( $parser );
 222
 223                 return true;
 224         }
 225
 226         function getText( $id ) {
 227                 $this->fetchCount++;
 228                 if( isset( $this->prefetch ) ) {
 229                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 230                         if( $text === null ) {
 231                                 // Entry missing from prefetch dump
 232                         } elseif( $text === "" ) {
 233                                 // Blank entries may indicate that the prior dump was broken.
 234                                 // To be safe, reload it.
 235                         } else {
 236                                 $this->prefetchCount++;
 237                                 return $text;
 238                         }
 239                 }
 240                 while( true ) {
 241                         try {
 242                                 $text = $this->doGetText( $id );
 243                                 $ex = new MWException("Graceful storage failure");
 244                         } catch (DBQueryError $ex) {
 245                                 $text = false;
 246                         }
 247                         if( $text === false ) {
 248                                 $this->failures++;
 249                                 if( $this->failures > $this->maxFailures ) {
 250                                         throw $ex;
 251                                 } else {
 252                                         $this->progress( "Database failure $this->failures " .
 253                                                 "of allowed $this->maxFailures for revision $id! " .
 254                                                 "Pausing $this->failureTimeout seconds..." );
 255                                         sleep( $this->failureTimeout );
 256                                 }
 257                         } else {
 258                                 return $text;
 259                         }
 260                 }
 261         }
 262
 263         /**
 264          * May throw a database error if, say, the server dies during query.
 265          */
 266         private function doGetText( $id ) {
 267                 $id = intval( $id );
 268                 $row = $this->db->selectRow( 'text',
 269                         array( 'old_text', 'old_flags' ),
 270                         array( 'old_id' => $id ),
 271                         'TextPassDumper::getText' );
 272                 $text = Revision::getRevisionText( $row );
 273                 if( $text === false ) {
 274                         return false;
 275                 }
 276                 $stripped = str_replace( "\r", "", $text );
 277                 $normalized = UtfNormal::cleanUp( $stripped );
 278                 return $normalized;
 279         }
 280
 281         function startElement( $parser, $name, $attribs ) {
 282                 $this->clearOpenElement( null );
 283                 $this->lastName = $name;
 284
 285                 if( $name == 'revision' ) {
 286                         $this->state = $name;
 287                         $this->egress->writeOpenPage( null, $this->buffer );
 288                         $this->buffer = "";
 289                 } elseif( $name == 'page' ) {
 290                         $this->state = $name;
 291                         if( $this->atStart ) {
 292                                 $this->egress->writeOpenStream( $this->buffer );
 293                                 $this->buffer = "";
 294                                 $this->atStart = false;
 295                         }
 296                 }
 297
 298                 if( $name == "text" && isset( $attribs['id'] ) ) {
 299                         $text = $this->getText( $attribs['id'] );
 300                         $this->openElement = array( $name, array( 'xml:space' => 'preserve' ) );
 301                         if( strlen( $text ) > 0 ) {
 302                                 $this->characterData( $parser, $text );
 303                         }
 304                 } else {
 305                         $this->openElement = array( $name, $attribs );
 306                 }
 307         }
 308
 309         function endElement( $parser, $name ) {
 310                 if( $this->openElement ) {
 311                         $this->clearOpenElement( "" );
 312                 } else {
 313                         $this->buffer .= "</$name>";
 314                 }
 315
 316                 if( $name == 'revision' ) {
 317                         $this->egress->writeRevision( null, $this->buffer );
 318                         $this->buffer = "";
 319                         $this->thisRev = "";
 320                 } elseif( $name == 'page' ) {
 321                         $this->egress->writeClosePage( $this->buffer );
 322                         $this->buffer = "";
 323                         $this->thisPage = "";
 324                 } elseif( $name == 'mediawiki' ) {
 325                         $this->egress->writeCloseStream( $this->buffer );
 326                         $this->buffer = "";
 327                 }
 328         }
 329
 330         function characterData( $parser, $data ) {
 331                 $this->clearOpenElement( null );
 332                 if( $this->lastName == "id" ) {
 333                         if( $this->state == "revision" ) {
 334                                 $this->thisRev .= $data;
 335                         } elseif( $this->state == "page" ) {
 336                                 $this->thisPage .= $data;
 337                         }
 338                 }
 339                 $this->buffer .= htmlspecialchars( $data );
 340         }
 341
 342         function clearOpenElement( $style ) {
 343                 if( $this->openElement ) {
 344                         $this->buffer .= wfElement( $this->openElement[0], $this->openElement[1], $style );
 345                         $this->openElement = false;
 346                 }
 347         }
 348 }
 349
 350
 351 $dumper = new TextPassDumper( $argv );
 352
 353 if( true ) {
 354         $dumper->dump();
 355 } else {
 356         $dumper->progress( <<<END
 357 This script postprocesses XML dumps from dumpBackup.php to add
 358 page text which was stubbed out (using --stub).
 359
 360 XML input is accepted on stdin.
 361 XML output is sent to stdout; progress reports are sent to stderr.
 362
 363 Usage: php dumpTextPass.php [<options>]
 364 Options:
 365   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 366   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 367               pressure on the database.
 368               (Requires PHP 5.0+ and the XMLReader PECL extension)
 369   --quiet     Don't dump status reports to stderr.
 370   --report=n  Report position and speed after every n pages processed.
 371               (Default: 100)
 372   --server=h  Force reading from MySQL server h
 373   --current   Base ETA on number of pages in database instead of all revisions
 374 END
 375 );
 376 }
 377
 378