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