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  * @file
  26  * @ingroup Maintenance
  27  */
  28
  29 if ( php_sapi_name() != 'cli' ) {
  30         echo "Please customize the settings and run me from the command line.";
  31         die( -1 );
  32 }
  33
  34 /** Set these correctly! */
  35 $wgImportEncoding = "CP1252"; /* We convert all to UTF-8 */
  36 $wgRootDirectory = "/kalman/Projects/wiki2002/wiki/lib-http/db/wiki";
  37
  38 /* On a large wiki, you might run out of memory */
  39 @ini_set( 'memory_limit', '40M' );
  40
  41 /* globals */
  42 $wgFieldSeparator = "\xb3"; # Some wikis may use different char
  43         $FS = $wgFieldSeparator ;
  44         $FS1 = $FS . "1" ;
  45         $FS2 = $FS . "2" ;
  46         $FS3 = $FS . "3" ;
  47
  48 # Unicode sanitization tools
  49 require_once( dirname( dirname( __FILE__ ) ) . '/includes/normal/UtfNormal.php' );
  50
  51 $usercache = array();
  52
  53 importPages();
  54
  55 # ------------------------------------------------------------------------------
  56
  57 function importPages()
  58 {
  59         global $wgRootDirectory;
  60
  61         $gt = '>';
  62         echo <<<XML
  63 <?xml version="1.0" encoding="UTF-8" ?$gt
  64 <mediawiki xmlns="http://www.mediawiki.org/xml/export-0.1/"
  65            xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  66            xsi:schemaLocation="http://www.mediawiki.org/xml/export-0.1/
  67                                http://www.mediawiki.org/xml/export-0.1.xsd"
  68            version="0.1"
  69            xml:lang="en">
  70 <!-- generated by importUseModWiki.php -->
  71
  72 XML;
  73         $letters = array(
  74                 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I',
  75                 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R',
  76                 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 'other' );
  77         foreach ( $letters as $letter ) {
  78                 $dir = "$wgRootDirectory/page/$letter";
  79                 if ( is_dir( $dir ) )
  80                         importPageDirectory( $dir );
  81         }
  82         echo <<<XML
  83 </mediawiki>
  84
  85 XML;
  86 }
  87
  88 function importPageDirectory( $dir, $prefix = "" )
  89 {
  90         echo "\n<!-- Checking page directory " . xmlCommentSafe( $dir ) . " -->\n";
  91         $mydir = opendir( $dir );
  92         while ( $entry = readdir( $mydir ) ) {
  93                 $m = array();
  94                 if ( preg_match( '/^(.+)\.db$/', $entry, $m ) ) {
  95                         echo importPage( $prefix . $m[1] );
  96                 } else {
  97                         if ( is_dir( "$dir/$entry" ) ) {
  98                                 if ( $entry != '.' && $entry != '..' ) {
  99                                         importPageDirectory( "$dir/$entry", "$entry/" );
 100                                 }
 101                         } else {
 102                                 echo "<!-- File '" . xmlCommentSafe( $entry ) . "' doesn't seem to contain an article. Skipping. -->\n";
 103                         }
 104                 }
 105         }
 106 }
 107
 108
 109 # ------------------------------------------------------------------------------
 110
 111 /* fetch_ functions
 112         Grab a given item from the database
 113         */
 114
 115 function useModFilename( $title ) {
 116         $c = substr( $title, 0, 1 );
 117         if ( preg_match( '/[A-Z]/i', $c ) ) {
 118                 return strtoupper( $c ) . "/$title";
 119         }
 120         return "other/$title";
 121 }
 122
 123 function fetchPage( $title )
 124 {
 125         global $FS1, $FS2, $FS3, $wgRootDirectory;
 126
 127         $fname = $wgRootDirectory . "/page/" . useModFilename( $title ) . ".db";
 128         if ( !file_exists( $fname ) ) {
 129                 echo "Couldn't open file '$fname' for page '$title'.\n";
 130                 die( -1 );
 131         }
 132
 133         $page = splitHash( $FS1, file_get_contents( $fname ) );
 134         $section = splitHash( $FS2, $page["text_default"] );
 135         $text = splitHash( $FS3, $section["data"] );
 136
 137         return array2object( array( "text" => $text["text"] , "summary" => $text["summary"] ,
 138                 "minor" => $text["minor"] , "ts" => $section["ts"] ,
 139                 "username" => $section["username"] , "host" => $section["host"] ) );
 140 }
 141
 142 function fetchKeptPages( $title )
 143 {
 144         global $FS1, $FS2, $FS3, $wgRootDirectory;
 145
 146         $fname = $wgRootDirectory . "/keep/" . useModFilename( $title ) . ".kp";
 147         if ( !file_exists( $fname ) ) return array();
 148
 149         $keptlist = explode( $FS1, file_get_contents( $fname ) );
 150         array_shift( $keptlist ); # Drop the junk at beginning of file
 151
 152         $revisions = array();
 153         foreach ( $keptlist as $rev ) {
 154                 $section = splitHash( $FS2, $rev );
 155                 $text = splitHash( $FS3, $section["data"] );
 156                 if ( $text["text"] && $text["minor"] != "" && ( $section["ts"] * 1 > 0 ) ) {
 157                         array_push( $revisions, array2object( array ( "text" => $text["text"] , "summary" => $text["summary"] ,
 158                                 "minor" => $text["minor"] , "ts" => $section["ts"] ,
 159                                 "username" => $section["username"] , "host" => $section["host"] ) ) );
 160                 } else {
 161                         echo "<!-- skipped a bad old revision -->\n";
 162                 }
 163         }
 164         return $revisions;
 165 }
 166
 167 function splitHash ( $sep , $str ) {
 168         $temp = explode ( $sep , $str ) ;
 169         $ret = array () ;
 170         for ( $i = 0; $i + 1 < count ( $temp ) ; $i++ ) {
 171                 $ret[$temp[$i]] = $temp[++$i] ;
 172                 }
 173         return $ret ;
 174         }
 175
 176
 177 /* import_ functions
 178         Take a fetched item and produce SQL
 179         */
 180
 181 function checkUserCache( $name, $host )
 182 {
 183         global $usercache;
 184
 185         if ( $name ) {
 186                 if ( in_array( $name, $usercache ) ) {
 187                         $userid = $usercache[$name];
 188                 } else {
 189                         # If we haven't imported user accounts
 190                         $userid = 0;
 191                 }
 192                 $username = str_replace( '_', ' ', $name );
 193         } else {
 194                 $userid = 0;
 195                 $username = $host;
 196         }
 197         return array( $userid, $username );
 198 }
 199
 200 function importPage( $title )
 201 {
 202         echo "\n<!-- Importing page " . xmlCommentSafe( $title ) . " -->\n";
 203         $page = fetchPage( $title );
 204
 205         $newtitle = xmlsafe( str_replace( '_', ' ', recodeText( $title ) ) );
 206
 207         $munged = mungeFormat( $page->text );
 208         if ( $munged != $page->text ) {
 209                 /**
 210                  * Save a *new* revision with the conversion, and put the
 211                  * previous last version into the history.
 212                  */
 213                 $next = array2object( array(
 214                         'text'     => $munged,
 215                         'minor'    => 1,
 216                         'username' => 'Conversion script',
 217                         'host'     => '127.0.0.1',
 218                         'ts'       => time(),
 219                         'summary'  => 'link fix',
 220                         ) );
 221                 $revisions = array( $page, $next );
 222         } else {
 223                 /**
 224                  * Current revision:
 225                  */
 226                 $revisions = array( $page );
 227         }
 228         $xml = <<<XML
 229         <page>
 230                 <title>$newtitle</title>
 231
 232 XML;
 233
 234         # History
 235         $revisions = array_merge( $revisions, fetchKeptPages( $title ) );
 236         if ( count( $revisions ) == 0 ) {
 237                 return NULL; // Was "$sql", which does not appear to be defined.
 238         }
 239
 240         foreach ( $revisions as $rev ) {
 241                 $text      = xmlsafe( recodeText( $rev->text ) );
 242                 $minor     = ( $rev->minor ? '<minor/>' : '' );
 243                 list( /* $userid */ , $username ) = checkUserCache( $rev->username, $rev->host );
 244                 $username  = xmlsafe( recodeText( $username ) );
 245                 $timestamp = xmlsafe( timestamp2ISO8601( $rev->ts ) );
 246                 $comment   = xmlsafe( recodeText( $rev->summary ) );
 247
 248                 $xml .= <<<XML
 249                 <revision>
 250                         <timestamp>$timestamp</timestamp>
 251                         <contributor><username>$username</username></contributor>
 252                         $minor
 253                         <comment>$comment</comment>
 254                         <text>$text</text>
 255                 </revision>
 256
 257 XML;
 258         }
 259         $xml .= "</page>\n\n";
 260         return $xml;
 261 }
 262
 263 # Whee!
 264 function recodeText( $string ) {
 265         global $wgImportEncoding;
 266         # For currently latin-1 wikis
 267         $string = str_replace( "\r\n", "\n", $string );
 268         $string = @iconv( $wgImportEncoding, "UTF-8", $string );
 269         $string = wfMungeToUtf8( $string ); # Any old &#1234; stuff
 270         return $string;
 271 }
 272
 273 function wfUtf8Sequence( $codepoint ) {
 274         if ( $codepoint <     0x80 ) return chr( $codepoint );
 275         if ( $codepoint <    0x800 ) return chr( $codepoint >>  6 & 0x3f | 0xc0 ) .
 276                                      chr( $codepoint       & 0x3f | 0x80 );
 277     if ( $codepoint <  0x10000 ) return chr( $codepoint >> 12 & 0x0f | 0xe0 ) .
 278                                      chr( $codepoint >>  6 & 0x3f | 0x80 ) .
 279                                      chr( $codepoint       & 0x3f | 0x80 );
 280         if ( $codepoint < 0x100000 ) return chr( $codepoint >> 18 & 0x07 | 0xf0 ) . # Double-check this
 281                                          chr( $codepoint >> 12 & 0x3f | 0x80 ) .
 282                                      chr( $codepoint >>  6 & 0x3f | 0x80 ) .
 283                                      chr( $codepoint       & 0x3f | 0x80 );
 284         # Doesn't yet handle outside the BMP
 285         return "&#$codepoint;";
 286 }
 287
 288 function wfMungeToUtf8( $string ) {
 289         $string = preg_replace ( '/&#([0-9]+);/e', 'wfUtf8Sequence($1)', $string );
 290         $string = preg_replace ( '/&#x([0-9a-f]+);/ie', 'wfUtf8Sequence(0x$1)', $string );
 291         # Should also do named entities here
 292         return $string;
 293 }
 294
 295 function timestamp2ISO8601( $ts ) {
 296         # 2003-08-05T18:30:02Z
 297         return gmdate( 'Y-m-d', $ts ) . 'T' . gmdate( 'H:i:s', $ts ) . 'Z';
 298 }
 299
 300 function xmlsafe( $string ) {
 301         /**
 302          * The page may contain old data which has not been properly normalized.
 303          * Invalid UTF-8 sequences or forbidden control characters will make our
 304          * XML output invalid, so be sure to strip them out.
 305          */
 306         $string = UtfNormal::cleanUp( $string );
 307
 308         $string = htmlspecialchars( $string );
 309         return $string;
 310 }
 311
 312 function xmlCommentSafe( $text ) {
 313         return str_replace( '--', '\\-\\-', xmlsafe( recodeText( $text ) ) );
 314 }
 315
 316
 317 function array2object( $arr ) {
 318         $o = (object)0;
 319         foreach ( $arr as $x => $y ) {
 320                 $o->$x = $y;
 321         }
 322         return $o;
 323 }
 324
 325
 326 /**
 327  * Make CamelCase and /Talk links work
 328  */
 329 function mungeFormat( $text ) {
 330         global $nowiki;
 331         $nowiki = array();
 332         $staged = preg_replace_callback(
 333                 '/(<nowiki>.*?<\\/nowiki>|(?:http|https|ftp):\\S+|\[\[[^]\\n]+]])/s',
 334                 'nowikiPlaceholder', $text );
 335
 336         # This is probably not  100% correct, I'm just
 337         # glancing at the UseModWiki code.
 338         $upper   = "[A-Z]";
 339         $lower   = "[a-z_0-9]";
 340         $any     = "[A-Za-z_0-9]";
 341         $camel   = "(?:$upper+$lower+$upper+$any*)";
 342         $subpage = "(?:\\/$any+)";
 343         $substart = "(?:\\/$upper$any*)";
 344
 345         $munged = preg_replace( "/(?!\\[\\[)($camel$subpage*|$substart$subpage*)\\b(?!\\]\\]|>)/",
 346                 '[[$1]]', $staged );
 347
 348         $final = preg_replace( '/' . preg_quote( placeholder() ) . '/es',
 349                 'array_shift( $nowiki )', $munged );
 350         return $final;
 351 }
 352
 353
 354 function placeholder( $x = null ) {
 355         return '\xffplaceholder\xff';
 356 }
 357
 358 function nowikiPlaceholder( $matches ) {
 359         global $nowiki;
 360         $nowiki[] = $matches[1];
 361         return placeholder();
 362 }
 363
 364