includes/SpecialImport.php

   1 <?php
   2 /**
   3  * MediaWiki page data importer
   4  * Copyright (C) 2003,2005 Brion Vibber <brion@pobox.com>
   5  * http://www.mediawiki.org/
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License along
  18  * with this program; if not, write to the Free Software Foundation, Inc.,
  19  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @package MediaWiki
  23  * @subpackage SpecialPage
  24  */
  25
  26 /**
  27  * Constructor
  28  */
  29 function wfSpecialImport( $page = '' ) {
  30         global $wgUser, $wgOut, $wgRequest, $wgTitle, $wgImportSources;
  31
  32         ###
  33 #       $wgOut->addWikiText( "Special:Import is not ready for this beta release, sorry." );
  34 #       return;
  35         ###
  36
  37         if( $wgRequest->wasPosted() && $wgRequest->getVal( 'action' ) == 'submit') {
  38                 switch( $wgRequest->getVal( "source" ) ) {
  39                 case "upload":
  40                         if( $wgUser->isAllowed( 'importupload' ) ) {
  41                                 $source = ImportStreamSource::newFromUpload( "xmlimport" );
  42                         } else {
  43                                 return $wgOut->permissionRequired( 'importupload' );
  44                         }
  45                         break;
  46                 case "interwiki":
  47                         $source = ImportStreamSource::newFromInterwiki(
  48                                 $wgRequest->getVal( "interwiki" ),
  49                                 $wgRequest->getText( "frompage" ) );
  50                         break;
  51                 default:
  52                         $source = new WikiError( "Unknown import source type" );
  53                 }
  54
  55                 if( WikiError::isError( $source ) ) {
  56                         $wgOut->addWikiText( wfEscapeWikiText( $source->getMessage() ) );
  57                 } else {
  58                         $importer = new WikiImporter( $source );
  59                         $result = $importer->doImport();
  60                         if( WikiError::isError( $result ) ) {
  61                                 $wgOut->addWikiText( wfMsg( "importfailed",
  62                                         wfEscapeWikiText( $result->getMessage() ) ) );
  63                         } else {
  64                                 # Success!
  65                                 $wgOut->addWikiText( wfMsg( "importsuccess" ) );
  66                         }
  67                 }
  68         }
  69
  70         $action = $wgTitle->escapeLocalUrl( 'action=submit' );
  71
  72         if( $wgUser->isAllowed( 'importupload' ) ) {
  73                 $wgOut->addWikiText( wfMsg( "importtext" ) );
  74                 $wgOut->addHTML( "
  75 <fieldset>
  76         <legend>" . wfMsgHtml('upload') . "</legend>
  77         <form enctype='multipart/form-data' method='post' action=\"$action\">
  78                 <input type='hidden' name='action' value='submit' />
  79                 <input type='hidden' name='source' value='upload' />
  80                 <input type='hidden' name='MAX_FILE_SIZE' value='2000000' />
  81                 <input type='file' name='xmlimport' value='' size='30' />
  82                 <input type='submit' value=\"" . wfMsgHtml( "uploadbtn" ) . "\" />
  83         </form>
  84 </fieldset>
  85 " );
  86         } else {
  87                 if( empty( $wgImportSources ) ) {
  88                         $wgOut->addWikiText( wfMsg( 'importnosources' ) );
  89                 }
  90         }
  91
  92         if( !empty( $wgImportSources ) ) {
  93                 $wgOut->addHTML( "
  94 <fieldset>
  95         <legend>" . wfMsgHtml('importinterwiki') . "</legend>
  96         <form method='post' action=\"$action\">
  97                 <input type='hidden' name='action' value='submit' />
  98                 <input type='hidden' name='source' value='interwiki' />
  99                 <select name='interwiki'>
 100 " );
 101                 foreach( $wgImportSources as $interwiki ) {
 102                         $iw = htmlspecialchars( $interwiki );
 103                         $wgOut->addHTML( "<option value=\"$iw\">$iw</option>\n" );
 104                 }
 105                 $wgOut->addHTML( "
 106                 </select>
 107                 <input name='frompage' />
 108                 <input type='submit' />
 109         </form>
 110 </fieldset>
 111 " );
 112         }
 113 }
 114
 115 /**
 116  *
 117  * @package MediaWiki
 118  * @subpackage SpecialPage
 119  */
 120 class WikiRevision {
 121         var $title = NULL;
 122         var $id = 0;
 123         var $timestamp = "20010115000000";
 124         var $user = 0;
 125         var $user_text = "";
 126         var $text = "";
 127         var $comment = "";
 128         var $minor = false;
 129
 130         function setTitle( $text ) {
 131                 $this->title = Title::newFromText( $text );
 132         }
 133
 134         function setID( $id ) {
 135                 $this->id = $id;
 136         }
 137
 138         function setTimestamp( $ts ) {
 139                 # 2003-08-05T18:30:02Z
 140                 $this->timestamp = wfTimestamp( TS_MW, $ts );
 141         }
 142
 143         function setUsername( $user ) {
 144                 $this->user_text = $user;
 145         }
 146
 147         function setUserIP( $ip ) {
 148                 $this->user_text = $ip;
 149         }
 150
 151         function setText( $text ) {
 152                 $this->text = $text;
 153         }
 154
 155         function setComment( $text ) {
 156                 $this->comment = $text;
 157         }
 158
 159         function setMinor( $minor ) {
 160                 $this->minor = (bool)$minor;
 161         }
 162
 163         function getTitle() {
 164                 return $this->title;
 165         }
 166
 167         function getID() {
 168                 return $this->id;
 169         }
 170
 171         function getTimestamp() {
 172                 return $this->timestamp;
 173         }
 174
 175         function getUser() {
 176                 return $this->user_text;
 177         }
 178
 179         function getText() {
 180                 return $this->text;
 181         }
 182
 183         function getComment() {
 184                 return $this->comment;
 185         }
 186
 187         function getMinor() {
 188                 return $this->minor;
 189         }
 190
 191         function importOldRevision() {
 192                 $fname = "WikiImporter::importOldRevision";
 193                 $dbw =& wfGetDB( DB_MASTER );
 194
 195                 # Sneak a single revision into place
 196                 $user = User::newFromName( $this->getUser() );
 197                 if( $user ) {
 198                         $userId = intval( $user->getId() );
 199                         $userText = $user->getName();
 200                 } else {
 201                         $userId = 0;
 202                         $userText = $this->getUser();
 203                 }
 204
 205                 // avoid memory leak...?
 206                 $linkCache =& LinkCache::singleton();
 207                 $linkCache->clear();
 208
 209                 $article = new Article( $this->title );
 210                 $pageId = $article->getId();
 211                 if( $pageId == 0 ) {
 212                         # must create the page...
 213                         $pageId = $article->insertOn( $dbw );
 214                 }
 215
 216                 # FIXME: Check for exact conflicts
 217                 # FIXME: Use original rev_id optionally
 218                 # FIXME: blah blah blah
 219
 220                 #if( $numrows > 0 ) {
 221                 #       return wfMsg( "importhistoryconflict" );
 222                 #}
 223
 224                 # Insert the row
 225                 $revision = new Revision( array(
 226                         'page'       => $pageId,
 227                         'text'       => $this->getText(),
 228                         'comment'    => $this->getComment(),
 229                         'user'       => $userId,
 230                         'user_text'  => $userText,
 231                         'timestamp'  => $this->timestamp,
 232                         'minor_edit' => $this->minor,
 233                         ) );
 234                 $revId = $revision->insertOn( $dbw );
 235                 $article->updateIfNewerOn( $dbw, $revision );
 236
 237                 return true;
 238         }
 239
 240 }
 241
 242 /**
 243  *
 244  * @package MediaWiki
 245  * @subpackage SpecialPage
 246  */
 247 class WikiImporter {
 248         var $mSource = null;
 249         var $mPageCallback = null;
 250         var $mRevisionCallback = null;
 251         var $lastfield;
 252
 253         function WikiImporter( $source ) {
 254                 $this->setRevisionCallback( array( &$this, "importRevision" ) );
 255                 $this->mSource = $source;
 256         }
 257
 258         function throwXmlError( $err ) {
 259                 $this->debug( "FAILURE: $err" );
 260                 wfDebug( "WikiImporter XML error: $err\n" );
 261         }
 262
 263         # --------------
 264
 265         function doImport() {
 266                 if( empty( $this->mSource ) ) {
 267                         return new WikiErrorMsg( "importnotext" );
 268                 }
 269
 270                 $parser = xml_parser_create( "UTF-8" );
 271
 272                 # case folding violates XML standard, turn it off
 273                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 274
 275                 xml_set_object( $parser, $this );
 276                 xml_set_element_handler( $parser, "in_start", "" );
 277
 278                 $offset = 0; // for context extraction on error reporting
 279                 do {
 280                         $chunk = $this->mSource->readChunk();
 281                         if( !xml_parse( $parser, $chunk, $this->mSource->atEnd() ) ) {
 282                                 wfDebug( "WikiImporter::doImport encountered XML parsing error\n" );
 283                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 284                         }
 285                         $offset += strlen( $chunk );
 286                 } while( $chunk !== false && !$this->mSource->atEnd() );
 287                 xml_parser_free( $parser );
 288
 289                 return true;
 290         }
 291
 292         function debug( $data ) {
 293                 #wfDebug( "IMPORT: $data\n" );
 294         }
 295
 296         function notice( $data ) {
 297                 global $wgCommandLineMode;
 298                 if( $wgCommandLineMode ) {
 299                         print "$data\n";
 300                 } else {
 301                         global $wgOut;
 302                         $wgOut->addHTML( "<li>$data</li>\n" );
 303                 }
 304         }
 305
 306         /**
 307          * Sets the action to perform as each new page in the stream is reached.
 308          * @param callable $callback
 309          * @return callable
 310          */
 311         function setPageCallback( $callback ) {
 312                 $previous = $this->mPageCallback;
 313                 $this->mPageCallback = $callback;
 314                 return $previous;
 315         }
 316
 317         /**
 318          * Sets the action to perform as each page revision is reached.
 319          * @param callable $callback
 320          * @return callable
 321          */
 322         function setRevisionCallback( $callback ) {
 323                 $previous = $this->mRevisionCallback;
 324                 $this->mRevisionCallback = $callback;
 325                 return $previous;
 326         }
 327
 328         /**
 329          * Default per-revision callback, performs the import.
 330          * @param WikiRevision $revision
 331          * @private
 332          */
 333         function importRevision( &$revision ) {
 334                 $dbw =& wfGetDB( DB_MASTER );
 335                 $dbw->deadlockLoop( array( &$revision, 'importOldRevision' ) );
 336         }
 337
 338         /**
 339          * Alternate per-revision callback, for debugging.
 340          * @param WikiRevision $revision
 341          * @private
 342          */
 343         function debugRevisionHandler( &$revision ) {
 344                 $this->debug( "Got revision:" );
 345                 if( is_object( $revision->title ) ) {
 346                         $this->debug( "-- Title: " . $revision->title->getPrefixedText() );
 347                 } else {
 348                         $this->debug( "-- Title: <invalid>" );
 349                 }
 350                 $this->debug( "-- User: " . $revision->user_text );
 351                 $this->debug( "-- Timestamp: " . $revision->timestamp );
 352                 $this->debug( "-- Comment: " . $revision->comment );
 353                 $this->debug( "-- Text: " . $revision->text );
 354         }
 355
 356         /**
 357          * Notify the callback function when a new <page> is reached.
 358          * @param Title $title
 359          * @private
 360          */
 361         function pageCallback( $title ) {
 362                 if( is_callable( $this->mPageCallback ) ) {
 363                         call_user_func( $this->mPageCallback, $title );
 364                 }
 365         }
 366
 367
 368         # XML parser callbacks from here out -- beware!
 369         function donothing( $parser, $x, $y="" ) {
 370                 #$this->debug( "donothing" );
 371         }
 372
 373         function in_start( $parser, $name, $attribs ) {
 374                 $this->debug( "in_start $name" );
 375                 if( $name != "mediawiki" ) {
 376                         return $this->throwXMLerror( "Expected <mediawiki>, got <$name>" );
 377                 }
 378                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 379         }
 380
 381         function in_mediawiki( $parser, $name, $attribs ) {
 382                 $this->debug( "in_mediawiki $name" );
 383                 if( $name == 'siteinfo' ) {
 384                         xml_set_element_handler( $parser, "in_siteinfo", "out_siteinfo" );
 385                 } elseif( $name == 'page' ) {
 386                         xml_set_element_handler( $parser, "in_page", "out_page" );
 387                 } else {
 388                         return $this->throwXMLerror( "Expected <page>, got <$name>" );
 389                 }
 390         }
 391         function out_mediawiki( $parser, $name ) {
 392                 $this->debug( "out_mediawiki $name" );
 393                 if( $name != "mediawiki" ) {
 394                         return $this->throwXMLerror( "Expected </mediawiki>, got </$name>" );
 395                 }
 396                 xml_set_element_handler( $parser, "donothing", "donothing" );
 397         }
 398
 399
 400         function in_siteinfo( $parser, $name, $attribs ) {
 401                 // no-ops for now
 402                 $this->debug( "in_siteinfo $name" );
 403                 switch( $name ) {
 404                 case "sitename":
 405                 case "base":
 406                 case "generator":
 407                 case "case":
 408                 case "namespaces":
 409                 case "namespace":
 410                         break;
 411                 default:
 412                         return $this->throwXMLerror( "Element <$name> not allowed in <siteinfo>." );
 413                 }
 414         }
 415
 416         function out_siteinfo( $parser, $name ) {
 417                 if( $name == "siteinfo" ) {
 418                         xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 419                 }
 420         }
 421
 422
 423         function in_page( $parser, $name, $attribs ) {
 424                 $this->debug( "in_page $name" );
 425                 switch( $name ) {
 426                 case "id":
 427                 case "title":
 428                 case "restrictions":
 429                         $this->appendfield = $name;
 430                         $this->appenddata = "";
 431                         $this->parenttag = "page";
 432                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 433                         xml_set_character_data_handler( $parser, "char_append" );
 434                         break;
 435                 case "revision":
 436                         $this->workRevision = new WikiRevision;
 437                         $this->workRevision->setTitle( $this->workTitle );
 438                         xml_set_element_handler( $parser, "in_revision", "out_revision" );
 439                         break;
 440                 default:
 441                         return $this->throwXMLerror( "Element <$name> not allowed in a <page>." );
 442                 }
 443         }
 444
 445         function out_page( $parser, $name ) {
 446                 $this->debug( "out_page $name" );
 447                 if( $name != "page" ) {
 448                         return $this->throwXMLerror( "Expected </page>, got </$name>" );
 449                 }
 450                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 451
 452                 $this->workTitle = NULL;
 453                 $this->workRevision = NULL;
 454         }
 455
 456         function in_nothing( $parser, $name, $attribs ) {
 457                 $this->debug( "in_nothing $name" );
 458                 return $this->throwXMLerror( "No child elements allowed here; got <$name>" );
 459         }
 460         function char_append( $parser, $data ) {
 461                 $this->debug( "char_append '$data'" );
 462                 $this->appenddata .= $data;
 463         }
 464         function out_append( $parser, $name ) {
 465                 $this->debug( "out_append $name" );
 466                 if( $name != $this->appendfield ) {
 467                         return $this->throwXMLerror( "Expected </{$this->appendfield}>, got </$name>" );
 468                 }
 469                 xml_set_element_handler( $parser, "in_$this->parenttag", "out_$this->parenttag" );
 470                 xml_set_character_data_handler( $parser, "donothing" );
 471
 472                 switch( $this->appendfield ) {
 473                 case "title":
 474                         $this->workTitle = $this->appenddata;
 475                         $this->pageCallback( $this->workTitle );
 476                         break;
 477                 case "id":
 478                         if ( $this->parenttag == 'revision' ) {
 479                                 $this->workRevision->setID( $this->appenddata );
 480                         }
 481                         break;
 482                 case "text":
 483                         $this->workRevision->setText( $this->appenddata );
 484                         break;
 485                 case "username":
 486                         $this->workRevision->setUsername( $this->appenddata );
 487                         break;
 488                 case "ip":
 489                         $this->workRevision->setUserIP( $this->appenddata );
 490                         break;
 491                 case "timestamp":
 492                         $this->workRevision->setTimestamp( $this->appenddata );
 493                         break;
 494                 case "comment":
 495                         $this->workRevision->setComment( $this->appenddata );
 496                         break;
 497                 case "minor":
 498                         $this->workRevision->setMinor( true );
 499                         break;
 500                 default:
 501                         $this->debug( "Bad append: {$this->appendfield}" );
 502                 }
 503                 $this->appendfield = "";
 504                 $this->appenddata = "";
 505         }
 506
 507         function in_revision( $parser, $name, $attribs ) {
 508                 $this->debug( "in_revision $name" );
 509                 switch( $name ) {
 510                 case "id":
 511                 case "timestamp":
 512                 case "comment":
 513                 case "minor":
 514                 case "text":
 515                         $this->parenttag = "revision";
 516                         $this->appendfield = $name;
 517                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 518                         xml_set_character_data_handler( $parser, "char_append" );
 519                         break;
 520                 case "contributor":
 521                         xml_set_element_handler( $parser, "in_contributor", "out_contributor" );
 522                         break;
 523                 default:
 524                         return $this->throwXMLerror( "Element <$name> not allowed in a <revision>." );
 525                 }
 526         }
 527
 528         function out_revision( $parser, $name ) {
 529                 $this->debug( "out_revision $name" );
 530                 if( $name != "revision" ) {
 531                         return $this->throwXMLerror( "Expected </revision>, got </$name>" );
 532                 }
 533                 xml_set_element_handler( $parser, "in_page", "out_page" );
 534
 535                 $out = call_user_func_array( $this->mRevisionCallback,
 536                         array( &$this->workRevision, &$this ) );
 537                 if( !empty( $out ) ) {
 538                         global $wgOut;
 539                         $wgOut->addHTML( "<li>" . $out . "</li>\n" );
 540                 }
 541         }
 542
 543         function in_contributor( $parser, $name, $attribs ) {
 544                 $this->debug( "in_contributor $name" );
 545                 switch( $name ) {
 546                 case "username":
 547                 case "ip":
 548                 case "id":
 549                         $this->parenttag = "contributor";
 550                         $this->appendfield = $name;
 551                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 552                         xml_set_character_data_handler( $parser, "char_append" );
 553                         break;
 554                 default:
 555                         $this->throwXMLerror( "Invalid tag <$name> in <contributor>" );
 556                 }
 557         }
 558
 559         function out_contributor( $parser, $name ) {
 560                 $this->debug( "out_contributor $name" );
 561                 if( $name != "contributor" ) {
 562                         return $this->throwXMLerror( "Expected </contributor>, got </$name>" );
 563                 }
 564                 xml_set_element_handler( $parser, "in_revision", "out_revision" );
 565         }
 566
 567 }
 568
 569 /** @package MediaWiki */
 570 class ImportStringSource {
 571         function ImportStringSource( $string ) {
 572                 $this->mString = $string;
 573                 $this->mRead = false;
 574         }
 575
 576         function atEnd() {
 577                 return $this->mRead;
 578         }
 579
 580         function readChunk() {
 581                 if( $this->atEnd() ) {
 582                         return false;
 583                 } else {
 584                         $this->mRead = true;
 585                         return $this->mString;
 586                 }
 587         }
 588 }
 589
 590 /** @package MediaWiki */
 591 class ImportStreamSource {
 592         function ImportStreamSource( $handle ) {
 593                 $this->mHandle = $handle;
 594         }
 595
 596         function atEnd() {
 597                 return feof( $this->mHandle );
 598         }
 599
 600         function readChunk() {
 601                 return fread( $this->mHandle, 32768 );
 602         }
 603
 604         function newFromFile( $filename ) {
 605                 $file = @fopen( $filename, 'rt' );
 606                 if( !$file ) {
 607                         return new WikiError( "Couldn't open import file" );
 608                 }
 609                 return new ImportStreamSource( $file );
 610         }
 611
 612         function newFromUpload( $fieldname = "xmlimport" ) {
 613                 $upload =& $_FILES[$fieldname];
 614
 615                 if( !isset( $upload ) || !$upload['name'] ) {
 616                         return new WikiErrorMsg( 'importnofile' );
 617                 }
 618                 if( !empty( $upload['error'] ) ) {
 619                         return new WikiErrorMsg( 'importuploaderror', $upload['error'] );
 620                 }
 621                 $fname = $upload['tmp_name'];
 622                 if( is_uploaded_file( $fname ) ) {
 623                         return ImportStreamSource::newFromFile( $fname );
 624                 } else {
 625                         return new WikiErrorMsg( 'importnofile' );
 626                 }
 627         }
 628
 629         function newFromURL( $url ) {
 630                 # fopen-wrappers are normally turned off for security.
 631                 ini_set( "allow_url_fopen", true );
 632                 $ret = ImportStreamSource::newFromFile( $url );
 633                 ini_set( "allow_url_fopen", false );
 634                 return $ret;
 635         }
 636
 637         function newFromInterwiki( $interwiki, $page ) {
 638                 $base = Title::getInterwikiLink( $interwiki );
 639                 if( empty( $base ) ) {
 640                         return new WikiError( 'Bad interwiki link' );
 641                 } else {
 642                         $import = wfUrlencode( "Special:Export/$page" );
 643                         $url = str_replace( "$1", $import, $base );
 644                         return ImportStreamSource::newFromURL( $url );
 645                 }
 646         }
 647 }
 648
 649
 650 ?>