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 require_once( 'WikiError.php' );
  28
  29 /**
  30  * Constructor
  31  */
  32 function wfSpecialImport( $page = '' ) {
  33         global $wgUser, $wgOut, $wgRequest, $wgTitle, $wgImportSources;
  34
  35         ###
  36 #       $wgOut->addWikiText( "Special:Import is not ready for this beta release, sorry." );
  37 #       return;
  38         ###
  39
  40         if( $wgRequest->wasPosted() && $wgRequest->getVal( 'action' ) == 'submit') {
  41                 switch( $wgRequest->getVal( "source" ) ) {
  42                 case "upload":
  43                         if( $wgUser->isAllowed( 'importupload' ) ) {
  44                                 $source = ImportStreamSource::newFromUpload( "xmlimport" );
  45                         } else {
  46                                 return $wgOut->permissionRequired( 'importupload' );
  47                         }
  48                         break;
  49                 case "interwiki":
  50                         $source = ImportStreamSource::newFromInterwiki(
  51                                 $wgRequest->getVal( "interwiki" ),
  52                                 $wgRequest->getText( "frompage" ) );
  53                         break;
  54                 default:
  55                         $source = new WikiError( "Unknown import source type" );
  56                 }
  57
  58                 if( WikiError::isError( $source ) ) {
  59                         $wgOut->addWikiText( wfEscapeWikiText( $source->getMessage() ) );
  60                 } else {
  61                         $importer = new WikiImporter( $source );
  62                         $result = $importer->doImport();
  63                         if( WikiError::isError( $result ) ) {
  64                                 $wgOut->addWikiText( wfMsg( "importfailed",
  65                                         wfEscapeWikiText( $result->getMessage() ) ) );
  66                         } else {
  67                                 # Success!
  68                                 $wgOut->addWikiText( wfMsg( "importsuccess" ) );
  69                         }
  70                 }
  71         }
  72
  73         $action = $wgTitle->escapeLocalUrl( 'action=submit' );
  74
  75         if( $wgUser->isAllowed( 'importupload' ) ) {
  76                 $wgOut->addWikiText( wfMsg( "importtext" ) );
  77                 $wgOut->addHTML( "
  78 <fieldset>
  79         <legend>" . wfMsgHtml('upload') . "</legend>
  80         <form enctype='multipart/form-data' method='post' action=\"$action\">
  81                 <input type='hidden' name='action' value='submit' />
  82                 <input type='hidden' name='source' value='upload' />
  83                 <input type='hidden' name='MAX_FILE_SIZE' value='2000000' />
  84                 <input type='file' name='xmlimport' value='' size='30' />
  85                 <input type='submit' value=\"" . wfMsgHtml( "uploadbtn" ) . "\" />
  86         </form>
  87 </fieldset>
  88 " );
  89         } else {
  90                 if( empty( $wgImportSources ) ) {
  91                         $wgOut->addWikiText( wfMsg( 'importnosources' ) );
  92                 }
  93         }
  94
  95         if( !empty( $wgImportSources ) ) {
  96                 $wgOut->addHTML( "
  97 <fieldset>
  98         <legend>" . wfMsgHtml('importinterwiki') . "</legend>
  99         <form method='post' action=\"$action\">
 100                 <input type='hidden' name='action' value='submit' />
 101                 <input type='hidden' name='source' value='interwiki' />
 102                 <select name='interwiki'>
 103 " );
 104                 foreach( $wgImportSources as $interwiki ) {
 105                         $iw = htmlspecialchars( $interwiki );
 106                         $wgOut->addHTML( "<option value=\"$iw\">$iw</option>\n" );
 107                 }
 108                 $wgOut->addHTML( "
 109                 </select>
 110                 <input name='frompage' />
 111                 <input type='submit' />
 112         </form>
 113 </fieldset>
 114 " );
 115         }
 116 }
 117
 118 /**
 119  *
 120  * @package MediaWiki
 121  * @subpackage SpecialPage
 122  */
 123 class WikiRevision {
 124         var $title = NULL;
 125         var $id = 0;
 126         var $timestamp = "20010115000000";
 127         var $user = 0;
 128         var $user_text = "";
 129         var $text = "";
 130         var $comment = "";
 131         var $minor = false;
 132
 133         function setTitle( $text ) {
 134                 $this->title = Title::newFromText( $text );
 135         }
 136
 137         function setID( $id ) {
 138                 $this->id = $id;
 139         }
 140
 141         function setTimestamp( $ts ) {
 142                 # 2003-08-05T18:30:02Z
 143                 $this->timestamp = wfTimestamp( TS_MW, $ts );
 144         }
 145
 146         function setUsername( $user ) {
 147                 $this->user_text = $user;
 148         }
 149
 150         function setUserIP( $ip ) {
 151                 $this->user_text = $ip;
 152         }
 153
 154         function setText( $text ) {
 155                 $this->text = $text;
 156         }
 157
 158         function setComment( $text ) {
 159                 $this->comment = $text;
 160         }
 161
 162         function setMinor( $minor ) {
 163                 $this->minor = (bool)$minor;
 164         }
 165
 166         function getTitle() {
 167                 return $this->title;
 168         }
 169
 170         function getID() {
 171                 return $this->id;
 172         }
 173
 174         function getTimestamp() {
 175                 return $this->timestamp;
 176         }
 177
 178         function getUser() {
 179                 return $this->user_text;
 180         }
 181
 182         function getText() {
 183                 return $this->text;
 184         }
 185
 186         function getComment() {
 187                 return $this->comment;
 188         }
 189
 190         function getMinor() {
 191                 return $this->minor;
 192         }
 193
 194         function importOldRevision() {
 195                 $fname = "WikiImporter::importOldRevision";
 196                 $dbw =& wfGetDB( DB_MASTER );
 197
 198                 # Sneak a single revision into place
 199                 $user = User::newFromName( $this->getUser() );
 200                 if( $user ) {
 201                         $userId = intval( $user->getId() );
 202                         $userText = $user->getName();
 203                 } else {
 204                         $userId = 0;
 205                         $userText = $this->getUser();
 206                 }
 207
 208                 // avoid memory leak...?
 209                 $linkCache =& LinkCache::singleton();
 210                 $linkCache->clear();
 211
 212                 $article = new Article( $this->title );
 213                 $pageId = $article->getId();
 214                 if( $pageId == 0 ) {
 215                         # must create the page...
 216                         $pageId = $article->insertOn( $dbw );
 217                 }
 218
 219                 # FIXME: Check for exact conflicts
 220                 # FIXME: Use original rev_id optionally
 221                 # FIXME: blah blah blah
 222
 223                 #if( $numrows > 0 ) {
 224                 #       return wfMsg( "importhistoryconflict" );
 225                 #}
 226
 227                 # Insert the row
 228                 $revision = new Revision( array(
 229                         'page'       => $pageId,
 230                         'text'       => $this->getText(),
 231                         'comment'    => $this->getComment(),
 232                         'user'       => $userId,
 233                         'user_text'  => $userText,
 234                         'timestamp'  => $this->timestamp,
 235                         'minor_edit' => $this->minor,
 236                         ) );
 237                 $revId = $revision->insertOn( $dbw );
 238                 $article->updateIfNewerOn( $dbw, $revision );
 239
 240                 return true;
 241         }
 242
 243 }
 244
 245 /**
 246  *
 247  * @package MediaWiki
 248  * @subpackage SpecialPage
 249  */
 250 class WikiImporter {
 251         var $mSource = null;
 252         var $mPageCallback = null;
 253         var $mRevisionCallback = null;
 254         var $lastfield;
 255
 256         function WikiImporter( $source ) {
 257                 $this->setRevisionCallback( array( &$this, "importRevision" ) );
 258                 $this->mSource = $source;
 259         }
 260
 261         function throwXmlError( $err ) {
 262                 $this->debug( "FAILURE: $err" );
 263                 wfDebug( "WikiImporter XML error: $err\n" );
 264         }
 265
 266         # --------------
 267
 268         function doImport() {
 269                 if( empty( $this->mSource ) ) {
 270                         return new WikiErrorMsg( "importnotext" );
 271                 }
 272
 273                 $parser = xml_parser_create( "UTF-8" );
 274
 275                 # case folding violates XML standard, turn it off
 276                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 277
 278                 xml_set_object( $parser, $this );
 279                 xml_set_element_handler( $parser, "in_start", "" );
 280
 281                 $offset = 0; // for context extraction on error reporting
 282                 do {
 283                         $chunk = $this->mSource->readChunk();
 284                         if( !xml_parse( $parser, $chunk, $this->mSource->atEnd() ) ) {
 285                                 wfDebug( "WikiImporter::doImport encountered XML parsing error\n" );
 286                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 287                         }
 288                         $offset += strlen( $chunk );
 289                 } while( $chunk !== false && !$this->mSource->atEnd() );
 290                 xml_parser_free( $parser );
 291
 292                 return true;
 293         }
 294
 295         function debug( $data ) {
 296                 #wfDebug( "IMPORT: $data\n" );
 297         }
 298
 299         function notice( $data ) {
 300                 global $wgCommandLineMode;
 301                 if( $wgCommandLineMode ) {
 302                         print "$data\n";
 303                 } else {
 304                         global $wgOut;
 305                         $wgOut->addHTML( "<li>$data</li>\n" );
 306                 }
 307         }
 308
 309         /**
 310          * Sets the action to perform as each new page in the stream is reached.
 311          * @param callable $callback
 312          * @return callable
 313          */
 314         function setPageCallback( $callback ) {
 315                 $previous = $this->mPageCallback;
 316                 $this->mPageCallback = $callback;
 317                 return $previous;
 318         }
 319
 320         /**
 321          * Sets the action to perform as each page revision is reached.
 322          * @param callable $callback
 323          * @return callable
 324          */
 325         function setRevisionCallback( $callback ) {
 326                 $previous = $this->mRevisionCallback;
 327                 $this->mRevisionCallback = $callback;
 328                 return $previous;
 329         }
 330
 331         /**
 332          * Default per-revision callback, performs the import.
 333          * @param WikiRevision $revision
 334          * @access private
 335          */
 336         function importRevision( &$revision ) {
 337                 $dbw =& wfGetDB( DB_MASTER );
 338                 $dbw->deadlockLoop( array( &$revision, 'importOldRevision' ) );
 339         }
 340
 341         /**
 342          * Alternate per-revision callback, for debugging.
 343          * @param WikiRevision $revision
 344          * @access private
 345          */
 346         function debugRevisionHandler( &$revision ) {
 347                 $this->debug( "Got revision:" );
 348                 if( is_object( $revision->title ) ) {
 349                         $this->debug( "-- Title: " . $revision->title->getPrefixedText() );
 350                 } else {
 351                         $this->debug( "-- Title: <invalid>" );
 352                 }
 353                 $this->debug( "-- User: " . $revision->user_text );
 354                 $this->debug( "-- Timestamp: " . $revision->timestamp );
 355                 $this->debug( "-- Comment: " . $revision->comment );
 356                 $this->debug( "-- Text: " . $revision->text );
 357         }
 358
 359         /**
 360          * Notify the callback function when a new <page> is reached.
 361          * @param Title $title
 362          * @access private
 363          */
 364         function pageCallback( $title ) {
 365                 if( is_callable( $this->mPageCallback ) ) {
 366                         call_user_func( $this->mPageCallback, $title );
 367                 }
 368         }
 369
 370
 371         # XML parser callbacks from here out -- beware!
 372         function donothing( $parser, $x, $y="" ) {
 373                 #$this->debug( "donothing" );
 374         }
 375
 376         function in_start( $parser, $name, $attribs ) {
 377                 $this->debug( "in_start $name" );
 378                 if( $name != "mediawiki" ) {
 379                         return $this->throwXMLerror( "Expected <mediawiki>, got <$name>" );
 380                 }
 381                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 382         }
 383
 384         function in_mediawiki( $parser, $name, $attribs ) {
 385                 $this->debug( "in_mediawiki $name" );
 386                 if( $name == 'siteinfo' ) {
 387                         xml_set_element_handler( $parser, "in_siteinfo", "out_siteinfo" );
 388                 } elseif( $name == 'page' ) {
 389                         xml_set_element_handler( $parser, "in_page", "out_page" );
 390                 } else {
 391                         return $this->throwXMLerror( "Expected <page>, got <$name>" );
 392                 }
 393         }
 394         function out_mediawiki( $parser, $name ) {
 395                 $this->debug( "out_mediawiki $name" );
 396                 if( $name != "mediawiki" ) {
 397                         return $this->throwXMLerror( "Expected </mediawiki>, got </$name>" );
 398                 }
 399                 xml_set_element_handler( $parser, "donothing", "donothing" );
 400         }
 401
 402
 403         function in_siteinfo( $parser, $name, $attribs ) {
 404                 // no-ops for now
 405                 $this->debug( "in_siteinfo $name" );
 406                 switch( $name ) {
 407                 case "sitename":
 408                 case "base":
 409                 case "generator":
 410                 case "case":
 411                 case "namespaces":
 412                 case "namespace":
 413                         break;
 414                 default:
 415                         return $this->throwXMLerror( "Element <$name> not allowed in <siteinfo>." );
 416                 }
 417         }
 418
 419         function out_siteinfo( $parser, $name ) {
 420                 if( $name == "siteinfo" ) {
 421                         xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 422                 }
 423         }
 424
 425
 426         function in_page( $parser, $name, $attribs ) {
 427                 $this->debug( "in_page $name" );
 428                 switch( $name ) {
 429                 case "id":
 430                 case "title":
 431                 case "restrictions":
 432                         $this->appendfield = $name;
 433                         $this->appenddata = "";
 434                         $this->parenttag = "page";
 435                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 436                         xml_set_character_data_handler( $parser, "char_append" );
 437                         break;
 438                 case "revision":
 439                         $this->workRevision = new WikiRevision;
 440                         $this->workRevision->setTitle( $this->workTitle );
 441                         xml_set_element_handler( $parser, "in_revision", "out_revision" );
 442                         break;
 443                 default:
 444                         return $this->throwXMLerror( "Element <$name> not allowed in a <page>." );
 445                 }
 446         }
 447
 448         function out_page( $parser, $name ) {
 449                 $this->debug( "out_page $name" );
 450                 if( $name != "page" ) {
 451                         return $this->throwXMLerror( "Expected </page>, got </$name>" );
 452                 }
 453                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 454
 455                 $this->workTitle = NULL;
 456                 $this->workRevision = NULL;
 457         }
 458
 459         function in_nothing( $parser, $name, $attribs ) {
 460                 $this->debug( "in_nothing $name" );
 461                 return $this->throwXMLerror( "No child elements allowed here; got <$name>" );
 462         }
 463         function char_append( $parser, $data ) {
 464                 $this->debug( "char_append '$data'" );
 465                 $this->appenddata .= $data;
 466         }
 467         function out_append( $parser, $name ) {
 468                 $this->debug( "out_append $name" );
 469                 if( $name != $this->appendfield ) {
 470                         return $this->throwXMLerror( "Expected </{$this->appendfield}>, got </$name>" );
 471                 }
 472                 xml_set_element_handler( $parser, "in_$this->parenttag", "out_$this->parenttag" );
 473                 xml_set_character_data_handler( $parser, "donothing" );
 474
 475                 switch( $this->appendfield ) {
 476                 case "title":
 477                         $this->workTitle = $this->appenddata;
 478                         $this->pageCallback( $this->workTitle );
 479                         break;
 480                 case "id":
 481                         if ( $this->parenttag == 'revision' ) {
 482                                 $this->workRevision->setID( $this->appenddata );
 483                         }
 484                         break;
 485                 case "text":
 486                         $this->workRevision->setText( $this->appenddata );
 487                         break;
 488                 case "username":
 489                         $this->workRevision->setUsername( $this->appenddata );
 490                         break;
 491                 case "ip":
 492                         $this->workRevision->setUserIP( $this->appenddata );
 493                         break;
 494                 case "timestamp":
 495                         $this->workRevision->setTimestamp( $this->appenddata );
 496                         break;
 497                 case "comment":
 498                         $this->workRevision->setComment( $this->appenddata );
 499                         break;
 500                 case "minor":
 501                         $this->workRevision->setMinor( true );
 502                         break;
 503                 default:
 504                         $this->debug( "Bad append: {$this->appendfield}" );
 505                 }
 506                 $this->appendfield = "";
 507                 $this->appenddata = "";
 508         }
 509
 510         function in_revision( $parser, $name, $attribs ) {
 511                 $this->debug( "in_revision $name" );
 512                 switch( $name ) {
 513                 case "id":
 514                 case "timestamp":
 515                 case "comment":
 516                 case "minor":
 517                 case "text":
 518                         $this->parenttag = "revision";
 519                         $this->appendfield = $name;
 520                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 521                         xml_set_character_data_handler( $parser, "char_append" );
 522                         break;
 523                 case "contributor":
 524                         xml_set_element_handler( $parser, "in_contributor", "out_contributor" );
 525                         break;
 526                 default:
 527                         return $this->throwXMLerror( "Element <$name> not allowed in a <revision>." );
 528                 }
 529         }
 530
 531         function out_revision( $parser, $name ) {
 532                 $this->debug( "out_revision $name" );
 533                 if( $name != "revision" ) {
 534                         return $this->throwXMLerror( "Expected </revision>, got </$name>" );
 535                 }
 536                 xml_set_element_handler( $parser, "in_page", "out_page" );
 537
 538                 $out = call_user_func_array( $this->mRevisionCallback,
 539                         array( &$this->workRevision, &$this ) );
 540                 if( !empty( $out ) ) {
 541                         global $wgOut;
 542                         $wgOut->addHTML( "<li>" . $out . "</li>\n" );
 543                 }
 544         }
 545
 546         function in_contributor( $parser, $name, $attribs ) {
 547                 $this->debug( "in_contributor $name" );
 548                 switch( $name ) {
 549                 case "username":
 550                 case "ip":
 551                 case "id":
 552                         $this->parenttag = "contributor";
 553                         $this->appendfield = $name;
 554                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 555                         xml_set_character_data_handler( $parser, "char_append" );
 556                         break;
 557                 default:
 558                         $this->throwXMLerror( "Invalid tag <$name> in <contributor>" );
 559                 }
 560         }
 561
 562         function out_contributor( $parser, $name ) {
 563                 $this->debug( "out_contributor $name" );
 564                 if( $name != "contributor" ) {
 565                         return $this->throwXMLerror( "Expected </contributor>, got </$name>" );
 566                 }
 567                 xml_set_element_handler( $parser, "in_revision", "out_revision" );
 568         }
 569
 570 }
 571
 572 /** @package MediaWiki */
 573 class ImportStringSource {
 574         function ImportStringSource( $string ) {
 575                 $this->mString = $string;
 576                 $this->mRead = false;
 577         }
 578
 579         function atEnd() {
 580                 return $this->mRead;
 581         }
 582
 583         function readChunk() {
 584                 if( $this->atEnd() ) {
 585                         return false;
 586                 } else {
 587                         $this->mRead = true;
 588                         return $this->mString;
 589                 }
 590         }
 591 }
 592
 593 /** @package MediaWiki */
 594 class ImportStreamSource {
 595         function ImportStreamSource( $handle ) {
 596                 $this->mHandle = $handle;
 597         }
 598
 599         function atEnd() {
 600                 return feof( $this->mHandle );
 601         }
 602
 603         function readChunk() {
 604                 return fread( $this->mHandle, 32768 );
 605         }
 606
 607         function newFromFile( $filename ) {
 608                 $file = @fopen( $filename, 'rt' );
 609                 if( !$file ) {
 610                         return new WikiError( "Couldn't open import file" );
 611                 }
 612                 return new ImportStreamSource( $file );
 613         }
 614
 615         function newFromUpload( $fieldname = "xmlimport" ) {
 616                 $upload =& $_FILES[$fieldname];
 617
 618                 if( !isset( $upload ) || !$upload['name'] ) {
 619                         return new WikiErrorMsg( 'importnofile' );
 620                 }
 621                 if( !empty( $upload['error'] ) ) {
 622                         return new WikiErrorMsg( 'importuploaderror', $upload['error'] );
 623                 }
 624                 $fname = $upload['tmp_name'];
 625                 if( is_uploaded_file( $fname ) ) {
 626                         return ImportStreamSource::newFromFile( $fname );
 627                 } else {
 628                         return new WikiErrorMsg( 'importnofile' );
 629                 }
 630         }
 631
 632         function newFromURL( $url ) {
 633                 # fopen-wrappers are normally turned off for security.
 634                 ini_set( "allow_url_fopen", true );
 635                 $ret = ImportStreamSource::newFromFile( $url );
 636                 ini_set( "allow_url_fopen", false );
 637                 return $ret;
 638         }
 639
 640         function newFromInterwiki( $interwiki, $page ) {
 641                 $base = Title::getInterwikiLink( $interwiki );
 642                 if( empty( $base ) ) {
 643                         return new WikiError( 'Bad interwiki link' );
 644                 } else {
 645                         $import = wfUrlencode( "Special:Export/$page" );
 646                         $url = str_replace( "$1", $import, $base );
 647                         return ImportStreamSource::newFromURL( $url );
 648                 }
 649         }
 650 }
 651
 652
 653 ?>