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