maintenance/importUseModWiki.php

   1 <?php
   2
   3 /**
   4  * Import data from a UseModWiki into a MediaWiki wiki
   5  * 2003-02-09 Brion VIBBER <brion@pobox.com>
   6  * Based loosely on Magnus's code from 2001-2002
   7  *
   8  * Updated limited version to get something working temporarily
   9  * 2003-10-09
  10  * Be sure to run the link & index rebuilding scripts!
  11  *
  12  * Some more munging for charsets etc
  13  * 2003-11-28
  14  *
  15  * Partial fix for pages starting with lowercase letters (??)
  16  * and CamelCase and /Subpage link conversion
  17  * 2004-11-17
  18  *
  19  * Rewrite output to create Special:Export format for import
  20  * instead of raw SQL. Should be 'future-proof' against future
  21  * schema changes.
  22  * 2005-03-14
  23  *
  24  * @todo document
  25  * @package MediaWiki
  26  * @subpackage Maintenance
  27  */
  28
  29 if( php_sapi_name() != 'cli' ) {
  30         die( "Please customize the settings and run me from the command line." );
  31 }
  32
  33 /** Set these correctly! */
  34 $wgImportEncoding = "CP1252"; /* We convert all to UTF-8 */
  35 $wgRootDirectory = "/kalman/Projects/wiki2002/wiki/lib-http/db/wiki";
  36
  37 /* On a large wiki, you might run out of memory */
  38 @ini_set( 'memory_limit', '40M' );
  39
  40 /* globals */
  41 $wgFieldSeparator = "\xb3"; # Some wikis may use different char
  42         $FS = $wgFieldSeparator ;
  43         $FS1 = $FS."1" ;
  44         $FS2 = $FS."2" ;
  45         $FS3 = $FS."3" ;
  46
  47 # Unicode sanitization tools
  48 require_once( '../includes/normal/UtfNormal.php' );
  49
  50 $usercache = array();
  51
  52 importPages();
  53
  54 # ------------------------------------------------------------------------------
  55
  56 function importPages()
  57 {
  58         global $wgRootDirectory;
  59
  60         $gt = '>';
  61         echo <<<END
  62 <?xml version="1.0" encoding="UTF-8" ?$gt
  63 <mediawiki xmlns="http://www.mediawiki.org/xml/export-0.1/"
  64            xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  65            xsi:schemaLocation="http://www.mediawiki.org/xml/export-0.1/
  66                                http://www.mediawiki.org/xml/export-0.1.xsd"
  67            version="0.1">
  68 <!-- generated by importUseModWiki.php -->
  69
  70 END;
  71         $letters = array(
  72                 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I',
  73                 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R',
  74                 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 'other' );
  75         foreach( $letters as $letter ) {
  76                 $dir = "$wgRootDirectory/page/$letter";
  77                 if( is_dir( $dir ) )
  78                         importPageDirectory( $dir );
  79         }
  80         echo <<<END
  81 </mediawiki>
  82
  83 END;
  84 }
  85
  86 function importPageDirectory( $dir, $prefix = "" )
  87 {
  88         echo "\n<!-- Checking page directory $dir -->\n";
  89         $mydir = opendir( $dir );
  90         while( $entry = readdir( $mydir ) ) {
  91                 if( preg_match( '/^(.+)\.db$/', $entry, $m ) ) {
  92                         echo importPage( $prefix . $m[1] );
  93                 } else {
  94                         if( is_dir( "$dir/$entry" ) ) {
  95                                 if( $entry != '.' && $entry != '..' ) {
  96                                         importPageDirectory( "$dir/$entry", "$entry/" );
  97                                 }
  98                         } else {
  99                                 echo "<!-- File '" . xmlCommentSafe( $entry ) . "' doesn't seem to contain an article. Skipping. -->\n";
 100                         }
 101                 }
 102         }
 103 }
 104
 105
 106 # ------------------------------------------------------------------------------
 107
 108 /* fetch_ functions
 109         Grab a given item from the database
 110         */
 111
 112 function useModFilename( $title ) {
 113         $c = substr( $title, 0, 1 );
 114         if(preg_match( '/[A-Z]/i', $c ) ) {
 115                 return strtoupper( $c ) . "/$title";
 116         }
 117         return "other/$title";
 118 }
 119
 120 function fetchPage( $title )
 121 {
 122         global $FS,$FS1,$FS2,$FS3, $wgRootDirectory;
 123
 124         $fname = $wgRootDirectory . "/page/" . useModFilename( $title ) . ".db";
 125         if( !file_exists( $fname ) ) {
 126                 die( "Couldn't open file '$fname' for page '$title'.\n" );
 127         }
 128
 129         $page = splitHash( $FS1, file_get_contents( $fname ) );
 130         $section = splitHash( $FS2, $page["text_default"] );
 131         $text = splitHash( $FS3, $section["data"] );
 132
 133         return array2object( array( "text" => $text["text"] , "summary" => $text["summary"] ,
 134                 "minor" => $text["minor"] , "ts" => $section["ts"] ,
 135                 "username" => $section["username"] , "host" => $section["host"] ) );
 136 }
 137
 138 function fetchKeptPages( $title )
 139 {
 140         global $FS,$FS1,$FS2,$FS3, $wgRootDirectory, $wgTimezoneCorrection;
 141
 142         $fname = $wgRootDirectory . "/keep/" . useModFilename( $title ) . ".kp";
 143         if( !file_exists( $fname ) ) return array();
 144
 145         $keptlist = explode( $FS1, file_get_contents( $fname ) );
 146         array_shift( $keptlist ); # Drop the junk at beginning of file
 147
 148         $revisions = array();
 149         foreach( $keptlist as $rev ) {
 150                 $section = splitHash( $FS2, $rev );
 151                 $text = splitHash( $FS3, $section["data"] );
 152                 if ( $text["text"] && $text["minor"] != "" && ( $section["ts"]*1 > 0 ) ) {
 153                         array_push( $revisions, array2object( array ( "text" => $text["text"] , "summary" => $text["summary"] ,
 154                                 "minor" => $text["minor"] , "ts" => $section["ts"] ,
 155                                 "username" => $section["username"] , "host" => $section["host"] ) ) );
 156                 } else {
 157                         echo "-- skipped a bad old revision\n";
 158                 }
 159         }
 160         return $revisions;
 161 }
 162
 163 function splitHash ( $sep , $str ) {
 164         $temp = explode ( $sep , $str ) ;
 165         $ret = array () ;
 166         for ( $i = 0; $i+1 < count ( $temp ) ; $i++ ) {
 167                 $ret[$temp[$i]] = $temp[++$i] ;
 168                 }
 169         return $ret ;
 170         }
 171
 172
 173 /* import_ functions
 174         Take a fetched item and produce SQL
 175         */
 176
 177 function checkUserCache( $name, $host )
 178 {
 179         global $usercache;
 180
 181         if( $name ) {
 182                 if( in_array( $name, $usercache ) ) {
 183                         $userid = $usercache[$name];
 184                 } else {
 185                         # If we haven't imported user accounts
 186                         $userid = 0;
 187                 }
 188                 $username = str_replace( '_', ' ', $name );
 189         } else {
 190                 $userid = 0;
 191                 $username = $host;
 192         }
 193         return array( $userid, $username );
 194 }
 195
 196 function importPage( $title )
 197 {
 198         global $usercache;
 199
 200         echo "\n<!-- Importing page " . xmlCommentSafe( $title ) . " -->\n";
 201         $page = fetchPage( $title );
 202
 203         $newtitle = xmlsafe( str_replace( '_', ' ', recodeText( $title ) ) );
 204
 205         $munged = mungeFormat( $page->text );
 206         if( $munged != $page->text ) {
 207                 /**
 208                  * Save a *new* revision with the conversion, and put the
 209                  * previous last version into the history.
 210                  */
 211                 $next = array2object( array(
 212                         'text'     => $munged,
 213                         'minor'    => 1,
 214                         'username' => 'Conversion script',
 215                         'host'     => '127.0.0.1',
 216                         'ts'       => time(),
 217                         'summary'  => 'link fix',
 218                         ) );
 219                 $revisions = array( $page, $next );
 220         } else {
 221                 /**
 222                  * Current revision:
 223                  */
 224                 $revisions = array( $page );
 225         }
 226         $xml = <<<END
 227         <page>
 228                 <title>$newtitle</title>
 229
 230 END;
 231
 232         # History
 233         $revisions = array_merge( $revisions, fetchKeptPages( $title ) );
 234         if(count( $revisions ) == 0 ) {
 235                 return $sql;
 236         }
 237
 238         foreach( $revisions as $rev ) {
 239                 $text      = xmlsafe( recodeText( $rev->text ) );
 240                 $minor     = ($rev->minor ? '<minor/>' : '');
 241                 list( $userid, $username ) = checkUserCache( $rev->username, $rev->host );
 242                 $username  = xmlsafe( recodeText( $username ) );
 243                 $timestamp = xmlsafe( timestamp2ISO8601( $rev->ts ) );
 244                 $comment   = xmlsafe( recodeText( $rev->summary ) );
 245
 246                 $xml .= <<<END
 247                 <revision>
 248                         <timestamp>$timestamp</timestamp>
 249                         <contributor><username>$username</username></contributor>
 250                         <comment>$comment</comment>
 251                         $minor
 252                         <text>$text</text>
 253                 </revision>
 254
 255 END;
 256         }
 257         $xml .= "</page>\n\n";
 258         return $xml;
 259 }
 260
 261 # Whee!
 262 function recodeText( $string ) {
 263         global $wgImportEncoding;
 264         # For currently latin-1 wikis
 265         $string = str_replace( "\r\n", "\n", $string );
 266         $string = @iconv( $wgImportEncoding, "UTF-8", $string );
 267         $string = wfMungeToUtf8( $string ); # Any old &#1234; stuff
 268         return $string;
 269 }
 270
 271 function wfUtf8Sequence($codepoint) {
 272         if($codepoint <     0x80) return chr($codepoint);
 273         if($codepoint <    0x800) return chr($codepoint >>  6 & 0x3f | 0xc0) .
 274                                      chr($codepoint       & 0x3f | 0x80);
 275     if($codepoint <  0x10000) return chr($codepoint >> 12 & 0x0f | 0xe0) .
 276                                      chr($codepoint >>  6 & 0x3f | 0x80) .
 277                                      chr($codepoint       & 0x3f | 0x80);
 278         if($codepoint < 0x100000) return chr($codepoint >> 18 & 0x07 | 0xf0) . # Double-check this
 279                                          chr($codepoint >> 12 & 0x3f | 0x80) .
 280                                      chr($codepoint >>  6 & 0x3f | 0x80) .
 281                                      chr($codepoint       & 0x3f | 0x80);
 282         # Doesn't yet handle outside the BMP
 283         return "&#$codepoint;";
 284 }
 285
 286 function wfMungeToUtf8($string) {
 287         $string = preg_replace ( '/&#([0-9]+);/e', 'wfUtf8Sequence($1)', $string );
 288         $string = preg_replace ( '/&#x([0-9a-f]+);/ie', 'wfUtf8Sequence(0x$1)', $string );
 289         # Should also do named entities here
 290         return $string;
 291 }
 292
 293 function timestamp2ISO8601( $ts ) {
 294         #2003-08-05T18:30:02Z
 295         return gmdate( 'Y-m-d', $ts ) . 'T' . gmdate( 'H:i:s', $ts ) . 'Z';
 296 }
 297
 298 function xmlsafe( $string ) {
 299         /**
 300          * The page may contain old data which has not been properly normalized.
 301          * Invalid UTF-8 sequences or forbidden control characters will make our
 302          * XML output invalid, so be sure to strip them out.
 303          */
 304         $string = UtfNormal::cleanUp( $string );
 305
 306         $string = htmlspecialchars( $string );
 307         return $string;
 308 }
 309
 310 function xmlCommentSafe( $text ) {
 311         return str_replace( '--', '\\-\\-', xmlsafe( $text ) );
 312 }
 313
 314
 315 function array2object( $arr ) {
 316         $o = (object)0;
 317         foreach( $arr as $x => $y ) {
 318                 $o->$x = $y;
 319         }
 320         return $o;
 321 }
 322
 323
 324 /**
 325  * Make CamelCase and /Talk links work
 326  */
 327 function mungeFormat( $text ) {
 328         global $nowiki;
 329         $nowiki = array();
 330         $staged = preg_replace_callback(
 331                 '/(<nowiki>.*?<\\/nowiki>|(?:http|https|ftp):\\S+|\[\[[^]\\n]+]])/s',
 332                 'nowikiPlaceholder', $text );
 333
 334         # This is probably not  100% correct, I'm just
 335         # glancing at the UseModWiki code.
 336         $upper   = "[A-Z]";
 337         $lower   = "[a-z_0-9]";
 338         $any     = "[A-Za-z_0-9]";
 339         $camel   = "(?:$upper+$lower+$upper+$any*)";
 340         $subpage = "(?:\\/$any+)";
 341         $substart = "(?:\\/$upper$any*)";
 342
 343         $munged = preg_replace( "/(?!\\[\\[)($camel$subpage*|$substart$subpage*)\\b(?!\\]\\]|>)/",
 344                 '[[$1]]', $staged );
 345
 346         $final = preg_replace( '/' . preg_quote( placeholder() ) . '/es',
 347                 'array_shift( $nowiki )', $munged );
 348         return $final;
 349 }
 350
 351
 352 function placeholder( $x = null ) {
 353         return '\xffplaceholder\xff';
 354 }
 355
 356 function nowikiPlaceholder( $matches ) {
 357         global $nowiki;
 358         $nowiki[] = $matches[1];
 359         return placeholder();
 360 }
 361
 362 ?>