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  * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @package MediaWiki
  23  * @subpackage SpecialPage
  24  */
  25
  26 /** */
  27 require_once( 'WikiError.php' );
  28
  29 /**
  30  * Constructor
  31  */
  32 function wfSpecialImport( $page = '' ) {
  33         global $wgUser, $wgOut, $wgLang, $wgRequest, $wgTitle;
  34         global $wgImportSources;
  35
  36         ###
  37 #       $wgOut->addWikiText( "Special:Import is not ready for this beta release, sorry." );
  38 #       return;
  39         ###
  40
  41         if( $wgRequest->wasPosted() && $wgRequest->getVal( 'action' ) == 'submit') {
  42                 switch( $wgRequest->getVal( "source" ) ) {
  43                 case "upload":
  44                         if( $wgUser->isAllowed( 'importupload' ) ) {
  45                                 $source = ImportStreamSource::newFromUpload( "xmlimport" );
  46                         } else {
  47                                 return $wgOut->permissionRequired( 'importupload' );
  48                         }
  49                         break;
  50                 case "interwiki":
  51                         $source = ImportStreamSource::newFromInterwiki(
  52                                 $wgRequest->getVal( "interwiki" ),
  53                                 $wgRequest->getText( "frompage" ) );
  54                         break;
  55                 default:
  56                         $source = new WikiError( "Unknown import source type" );
  57                 }
  58
  59                 if( WikiError::isError( $source ) ) {
  60                         $wgOut->addWikiText( wfEscapeWikiText( $source->getMessage() ) );
  61                 } else {
  62                         $importer = new WikiImporter( $source );
  63                         $result = $importer->doImport();
  64                         if( WikiError::isError( $result ) ) {
  65                                 $wgOut->addWikiText( wfMsg( "importfailed",
  66                                         wfEscapeWikiText( $result->getMessage() ) ) );
  67                         } else {
  68                                 # Success!
  69                                 $wgOut->addWikiText( wfMsg( "importsuccess" ) );
  70                         }
  71                 }
  72         }
  73
  74         $action = $wgTitle->escapeLocalUrl( 'action=submit' );
  75
  76         if( $wgUser->isAllowed( 'importupload' ) ) {
  77                 $wgOut->addWikiText( wfMsg( "importtext" ) );
  78                 $wgOut->addHTML( "
  79 <fieldset>
  80         <legend>" . wfMsgHtml('upload') . "</legend>
  81         <form enctype='multipart/form-data' method='post' action=\"$action\">
  82                 <input type='hidden' name='action' value='submit' />
  83                 <input type='hidden' name='source' value='upload' />
  84                 <input type='hidden' name='MAX_FILE_SIZE' value='2000000' />
  85                 <input type='file' name='xmlimport' value='' size='30' />
  86                 <input type='submit' value='" . wfMsgHtml( "uploadbtn" ) . "'/>
  87         </form>
  88 </fieldset>
  89 " );
  90         } else {
  91                 if( empty( $wgImportSources ) ) {
  92                         $wgOut->addWikiText( wfMsg( 'importnosources' ) );
  93                 }
  94         }
  95
  96         if( !empty( $wgImportSources ) ) {
  97                 $wgOut->addHTML( "
  98 <fieldset>
  99         <legend>" . wfMsgHtml('importinterwiki') . "</legend>
 100         <form method='post' action=\"$action\">
 101                 <input type='hidden' name='action' value='submit' />
 102                 <input type='hidden' name='source' value='interwiki' />
 103                 <select name='interwiki'>
 104 " );
 105                 foreach( $wgImportSources as $interwiki ) {
 106                         $iw = htmlspecialchars( $interwiki );
 107                         $wgOut->addHTML( "<option value=\"$iw\">$iw</option>\n" );
 108                 }
 109                 $wgOut->addHTML( "
 110                 </select>
 111                 <input name='frompage' />
 112                 <input type='submit' />
 113         </form>
 114 </fieldset>
 115 " );
 116         }
 117 }
 118
 119 /**
 120  *
 121  * @package MediaWiki
 122  * @subpackage SpecialPage
 123  */
 124 class WikiRevision {
 125         var $title = NULL;
 126         var $timestamp = "20010115000000";
 127         var $user = 0;
 128         var $user_text = "";
 129         var $text = "";
 130         var $comment = "";
 131
 132         function setTitle( $text ) {
 133                 $this->title = Title::newFromText( $text );
 134         }
 135
 136         function setTimestamp( $ts ) {
 137                 # 2003-08-05T18:30:02Z
 138                 $this->timestamp = preg_replace( '/^(....)-(..)-(..)T(..):(..):(..)Z$/', '$1$2$3$4$5$6', $ts );
 139         }
 140
 141         function setUsername( $user ) {
 142                 $this->user_text = $user;
 143         }
 144
 145         function setUserIP( $ip ) {
 146                 $this->user_text = $ip;
 147         }
 148
 149         function setText( $text ) {
 150                 $this->text = $text;
 151         }
 152
 153         function setComment( $text ) {
 154                 $this->comment = $text;
 155         }
 156
 157         function getTitle() {
 158                 return $this->title;
 159         }
 160
 161         function getTimestamp() {
 162                 return $this->timestamp;
 163         }
 164
 165         function getUser() {
 166                 return $this->user_text;
 167         }
 168
 169         function getText() {
 170                 return $this->text;
 171         }
 172
 173         function getComment() {
 174                 return $this->comment;
 175         }
 176
 177         function importOldRevision() {
 178                 $fname = "WikiImporter::importOldRevision";
 179                 $dbw =& wfGetDB( DB_MASTER );
 180
 181                 # Sneak a single revision into place
 182                 $user = User::newFromName( $this->getUser() );
 183                 if( $user ) {
 184                         $userId = IntVal( $user->getId() );
 185                         $userText = $user->getName();
 186                 } else {
 187                         $userId = 0;
 188                         $userText = $this->getUser();
 189                 }
 190
 191                 $article = new Article( $this->title );
 192                 $pageId = $article->getId();
 193                 if( $pageId == 0 ) {
 194                         # must create the page...
 195                         $pageId = $article->insertOn( $dbw );
 196                 }
 197
 198                 # FIXME: Check for exact conflicts
 199                 # FIXME: Use original rev_id optionally
 200                 # FIXME: blah blah blah
 201
 202                 #if( $numrows > 0 ) {
 203                 #       return wfMsg( "importhistoryconflict" );
 204                 #}
 205
 206                 # Insert the row
 207                 $revision = new Revision( array(
 208                         'page'       => $pageId,
 209                         'text'       => $this->getText(),
 210                         'comment'    => $this->getComment(),
 211                         'user'       => $userId,
 212                         'user_text'  => $userText,
 213                         'timestamp'  => $this->timestamp,
 214                         'minor_edit' => 0
 215                         ) );
 216                 $revId = $revision->insertOn( $dbw );
 217                 $article->updateIfNewerOn( $dbw, $revision );
 218
 219                 return true;
 220         }
 221
 222 }
 223
 224 /**
 225  *
 226  * @package MediaWiki
 227  * @subpackage SpecialPage
 228  */
 229 class WikiImporter {
 230         var $mSource = null;
 231         var $mPageCallback = null;
 232         var $mRevisionCallback = null;
 233         var $lastfield;
 234
 235         function WikiImporter( $source ) {
 236                 $this->setRevisionCallback( array( &$this, "importRevision" ) );
 237                 $this->mSource = $source;
 238         }
 239
 240         function throwXmlError( $err ) {
 241                 $this->debug( "FAILURE: $err" );
 242         }
 243
 244         # --------------
 245
 246         function doImport() {
 247                 if( empty( $this->mSource ) ) {
 248                         return new WikiErrorMsg( "importnotext" );
 249                 }
 250
 251                 $parser = xml_parser_create( "UTF-8" );
 252
 253                 # case folding violates XML standard, turn it off
 254                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 255
 256                 xml_set_object( $parser, &$this );
 257                 xml_set_element_handler( $parser, "in_start", "" );
 258
 259                 do {
 260                         $chunk = $this->mSource->readChunk();
 261                         if( !xml_parse( $parser, $chunk, $this->mSource->atEnd() ) ) {
 262                                 return new WikiXmlError( $parser );
 263                         }
 264                 } while( $chunk !== false && !$this->mSource->atEnd() );
 265                 xml_parser_free( $parser );
 266
 267                 return true;
 268         }
 269
 270         function debug( $data ) {
 271                 #wfDebug( "IMPORT: $data\n" );
 272         }
 273
 274         function notice( $data ) {
 275                 global $wgCommandLineMode;
 276                 if( $wgCommandLineMode ) {
 277                         print "$data\n";
 278                 } else {
 279                         global $wgOut;
 280                         $wgOut->addHTML( "<li>$data</li>\n" );
 281                 }
 282         }
 283
 284         /**
 285          * Sets the action to perform as each new page in the stream is reached.
 286          * @param callable $callback
 287          * @return callable
 288          */
 289         function setPageCallback( $callback ) {
 290                 $previous = $this->mPageCallback;
 291                 $this->mPageCallback = $callback;
 292                 return $previous;
 293         }
 294
 295         /**
 296          * Sets the action to perform as each page revision is reached.
 297          * @param callable $callback
 298          * @return callable
 299          */
 300         function setRevisionCallback( $callback ) {
 301                 $previous = $this->mRevisionCallback;
 302                 $this->mRevisionCallback = $callback;
 303                 return $previous;
 304         }
 305
 306         /**
 307          * Default per-revision callback, performs the import.
 308          * @param WikiRevision $revision
 309          * @access private
 310          */
 311         function importRevision( &$revision ) {
 312                 $dbw =& wfGetDB( DB_MASTER );
 313                 $dbw->deadlockLoop( array( &$revision, 'importOldRevision' ) );
 314         }
 315
 316         /**
 317          * Alternate per-revision callback, for debugging.
 318          * @param WikiRevision $revision
 319          * @access private
 320          */
 321         function debugRevisionHandler( &$revision ) {
 322                 $this->debug( "Got revision:" );
 323                 if( is_object( $revision->title ) ) {
 324                         $this->debug( "-- Title: " . $revision->title->getPrefixedText() );
 325                 } else {
 326                         $this->debug( "-- Title: <invalid>" );
 327                 }
 328                 $this->debug( "-- User: " . $revision->user_text );
 329                 $this->debug( "-- Timestamp: " . $revision->timestamp );
 330                 $this->debug( "-- Comment: " . $revision->comment );
 331                 $this->debug( "-- Text: " . $revision->text );
 332         }
 333
 334         /**
 335          * Notify the callback function when a new <page> is reached.
 336          * @param Title $title
 337          * @access private
 338          */
 339         function pageCallback( $title ) {
 340                 if( is_callable( $this->mPageCallback ) ) {
 341                         call_user_func( $this->mPageCallback, $title );
 342                 }
 343         }
 344
 345
 346         # XML parser callbacks from here out -- beware!
 347         function donothing( $parser, $x, $y="" ) {
 348                 #$this->debug( "donothing" );
 349         }
 350
 351         function in_start( $parser, $name, $attribs ) {
 352                 $this->debug( "in_start $name" );
 353                 if( $name != "mediawiki" ) {
 354                         return $this->throwXMLerror( "Expected <mediawiki>, got <$name>" );
 355                 }
 356                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 357         }
 358
 359         function in_mediawiki( $parser, $name, $attribs ) {
 360                 $this->debug( "in_mediawiki $name" );
 361                 if( $name == 'siteinfo' ) {
 362                         xml_set_element_handler( $parser, "in_siteinfo", "out_siteinfo" );
 363                 } elseif( $name == 'page' ) {
 364                         xml_set_element_handler( $parser, "in_page", "out_page" );
 365                 } else {
 366                         return $this->throwXMLerror( "Expected <page>, got <$name>" );
 367                 }
 368         }
 369         function out_mediawiki( $parser, $name ) {
 370                 $this->debug( "out_mediawiki $name" );
 371                 if( $name != "mediawiki" ) {
 372                         return $this->throwXMLerror( "Expected </mediawiki>, got </$name>" );
 373                 }
 374                 xml_set_element_handler( $parser, "donothing", "donothing" );
 375         }
 376
 377
 378         function in_siteinfo( $parser, $name, $attribs ) {
 379                 // no-ops for now
 380                 $this->debug( "in_siteinfo $name" );
 381                 switch( $name ) {
 382                 case "sitename":
 383                 case "generator":
 384                 case "case":
 385                 case "namespaces":
 386                 case "namespace":
 387                         break;
 388                 default:
 389                         return $this->throwXMLerror( "Element <$name> not allowed in <siteinfo>." );
 390                 }
 391         }
 392
 393         function out_siteinfo( $parser, $name ) {
 394                 if( $name == "siteinfo" ) {
 395                         xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 396                 }
 397         }
 398
 399
 400         function in_page( $parser, $name, $attribs ) {
 401                 $this->debug( "in_page $name" );
 402                 switch( $name ) {
 403                 case "id":
 404                 case "title":
 405                 case "restrictions":
 406                         $this->appendfield = $name;
 407                         $this->appenddata = "";
 408                         $this->parenttag = "page";
 409                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 410                         xml_set_character_data_handler( $parser, "char_append" );
 411                         break;
 412                 case "revision":
 413                         $this->workRevision = new WikiRevision;
 414                         $this->workRevision->setTitle( $this->workTitle );
 415                         xml_set_element_handler( $parser, "in_revision", "out_revision" );
 416                         break;
 417                 default:
 418                         return $this->throwXMLerror( "Element <$name> not allowed in a <page>." );
 419                 }
 420         }
 421
 422         function out_page( $parser, $name ) {
 423                 $this->debug( "out_page $name" );
 424                 if( $name != "page" ) {
 425                         return $this->throwXMLerror( "Expected </page>, got </$name>" );
 426                 }
 427                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 428
 429                 $this->workTitle = NULL;
 430                 $this->workRevision = NULL;
 431         }
 432
 433         function in_nothing( $parser, $name, $attribs ) {
 434                 $this->debug( "in_nothing $name" );
 435                 return $this->throwXMLerror( "No child elements allowed here; got <$name>" );
 436         }
 437         function char_append( $parser, $data ) {
 438                 $this->debug( "char_append '$data'" );
 439                 $this->appenddata .= $data;
 440         }
 441         function out_append( $parser, $name ) {
 442                 $this->debug( "out_append $name" );
 443                 if( $name != $this->appendfield ) {
 444                         return $this->throwXMLerror( "Expected </{$this->appendfield}>, got </$name>" );
 445                 }
 446                 xml_set_element_handler( $parser, "in_$this->parenttag", "out_$this->parenttag" );
 447                 xml_set_character_data_handler( $parser, "donothing" );
 448
 449                 switch( $this->appendfield ) {
 450                 case "title":
 451                         $this->workTitle = $this->appenddata;
 452                         $this->pageCallback( $this->workTitle );
 453                         break;
 454                 case "text":
 455                         $this->workRevision->setText( $this->appenddata );
 456                         break;
 457                 case "username":
 458                         $this->workRevision->setUsername( $this->appenddata );
 459                         break;
 460                 case "ip":
 461                         $this->workRevision->setUserIP( $this->appenddata );
 462                         break;
 463                 case "timestamp":
 464                         $this->workRevision->setTimestamp( $this->appenddata );
 465                         break;
 466                 case "comment":
 467                         $this->workRevision->setComment( $this->appenddata );
 468                         break;
 469                 default:
 470                         $this->debug( "Bad append: {$this->appendfield}" );
 471                 }
 472                 $this->appendfield = "";
 473                 $this->appenddata = "";
 474         }
 475
 476         function in_revision( $parser, $name, $attribs ) {
 477                 $this->debug( "in_revision $name" );
 478                 switch( $name ) {
 479                 case "id":
 480                 case "timestamp":
 481                 case "comment":
 482                 case "text":
 483                         $this->parenttag = "revision";
 484                         $this->appendfield = $name;
 485                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 486                         xml_set_character_data_handler( $parser, "char_append" );
 487                         break;
 488                 case "contributor":
 489                         xml_set_element_handler( $parser, "in_contributor", "out_contributor" );
 490                         break;
 491                 default:
 492                         return $this->throwXMLerror( "Element <$name> not allowed in a <revision>." );
 493                 }
 494         }
 495
 496         function out_revision( $parser, $name ) {
 497                 $this->debug( "out_revision $name" );
 498                 if( $name != "revision" ) {
 499                         return $this->throwXMLerror( "Expected </revision>, got </$name>" );
 500                 }
 501                 xml_set_element_handler( $parser, "in_page", "out_page" );
 502
 503                 $out = call_user_func( $this->mRevisionCallback,
 504                         &$this->workRevision,
 505                         &$this );
 506                 if( !empty( $out ) ) {
 507                         global $wgOut;
 508                         $wgOut->addHTML( "<li>" . $out . "</li>\n" );
 509                 }
 510         }
 511
 512         function in_contributor( $parser, $name, $attribs ) {
 513                 $this->debug( "in_contributor $name" );
 514                 switch( $name ) {
 515                 case "username":
 516                 case "ip":
 517                         $this->parenttag = "contributor";
 518                         $this->appendfield = $name;
 519                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 520                         xml_set_character_data_handler( $parser, "char_append" );
 521                         break;
 522                 default:
 523                         $this->throwXMLerror( "Invalid tag <$name> in <contributor>" );
 524                 }
 525         }
 526
 527         function out_contributor( $parser, $name ) {
 528                 $this->debug( "out_contributor $name" );
 529                 if( $name != "contributor" ) {
 530                         return $this->throwXMLerror( "Expected </contributor>, got </$name>" );
 531                 }
 532                 xml_set_element_handler( $parser, "in_revision", "out_revision" );
 533         }
 534
 535 }
 536
 537 /** @package MediaWiki */
 538 class ImportStringSource {
 539         function ImportStringSource( $string ) {
 540                 $this->mString = $string;
 541                 $this->mRead = false;
 542         }
 543
 544         function atEnd() {
 545                 return $this->mRead;
 546         }
 547
 548         function readChunk() {
 549                 if( $this->atEnd() ) {
 550                         return false;
 551                 } else {
 552                         $this->mRead = true;
 553                         return $this->mString;
 554                 }
 555         }
 556 }
 557
 558 /** @package MediaWiki */
 559 class ImportStreamSource {
 560         function ImportStreamSource( $handle ) {
 561                 $this->mHandle = $handle;
 562         }
 563
 564         function atEnd() {
 565                 return feof( $this->mHandle );
 566         }
 567
 568         function readChunk() {
 569                 return fread( $this->mHandle, 32768 );
 570         }
 571
 572         function newFromFile( $filename ) {
 573                 $file = @fopen( $filename, 'rt' );
 574                 if( !$file ) {
 575                         return new WikiError( "Couldn't open import file" );
 576                 }
 577                 return new ImportStreamSource( $file );
 578         }
 579
 580         function newFromUpload( $fieldname = "xmlimport" ) {
 581                 global $wgOut;
 582
 583                 $upload =& $_FILES[$fieldname];
 584
 585                 if( !isset( $upload ) ) {
 586                         return new WikiErrorMsg( 'importnofile' );
 587                 }
 588                 if( !empty( $upload['error'] ) ) {
 589                         return new WikiErrorMsg( 'importuploaderror', $upload['error'] );
 590                 }
 591                 $fname = $upload['tmp_name'];
 592                 if( is_uploaded_file( $fname ) ) {
 593                         return ImportStreamSource::newFromFile( $fname );
 594                 } else {
 595                         return new WikiErrorMsg( 'importnofile' );
 596                 }
 597         }
 598
 599         function newFromURL( $url ) {
 600                 # fopen-wrappers are normally turned off for security.
 601                 ini_set( "allow_url_fopen", true );
 602                 $ret = ImportStreamSource::newFromFile( $url );
 603                 ini_set( "allow_url_fopen", false );
 604                 return $ret;
 605         }
 606
 607         function newFromInterwiki( $interwiki, $page ) {
 608                 $base = Title::getInterwikiLink( $interwiki );
 609                 if( empty( $base ) ) {
 610                         return new WikiError( 'Bad interwiki link' );
 611                 } else {
 612                         $import = wfUrlencode( "Special:Export/$page" );
 613                         $url = str_replace( "$1", $import, $base );
 614                         return ImportStreamSource::newFromURL( $url );
 615                 }
 616         }
 617 }
 618
 619
 620 ?>