maintenance/importPhase2.php

   1 <?php
   2 # MediaWiki 'phase 2' to current format import script
   3 # (import format current as of 1.2.0, March 2004)
   4 #
   5 # Copyright (C) 2004 Brion Vibber <brion@pobox.com>
   6 # Portions by Lee Daniel Crocker, 2002
   7 # http://www.mediawiki.org/
   8 #
   9 # This program is free software; you can redistribute it and/or modify
  10 # it under the terms of the GNU General Public License as published by
  11 # the Free Software Foundation; either version 2 of the License, or
  12 # (at your option) any later version.
  13 #
  14 # This program is distributed in the hope that it will be useful,
  15 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  16 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17 # GNU General Public License for more details.
  18 #
  19 # You should have received a copy of the GNU General Public License along
  20 # with this program; if not, write to the Free Software Foundation, Inc.,
  21 # 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  22 # http://www.gnu.org/copyleft/gpl.html
  23
  24 if ( ! is_readable( "../LocalSettings.php" ) ) {
  25         print "A copy of your installation's LocalSettings.php\n" .
  26           "must exist in the source directory.\n";
  27         exit();
  28 }
  29
  30 $wgCommandLineMode = true;
  31 ini_set("implicit_flush", 1);
  32
  33 $DP = "../includes";
  34 require_once( "../LocalSettings.php" );
  35 require_once( "../AdminSettings.php" );
  36
  37 $wgDBuser = $wgDBadminuser;
  38 $wgDBpassword = $wgDBadminpassword;
  39
  40 $sep = ( DIRECTORY_SEPARATOR == "\\" ) ? ";" : ":";
  41 ini_set( "include_path", "$IP$sep$include_path" );
  42
  43 require_once( "Setup.php" );
  44
  45 require_once( "../install-utils.inc" );
  46 require_once( "InitialiseMessages.inc" );
  47 require_once( "rebuildlinks.inc" );
  48 require_once( "rebuildrecentchanges.inc" );
  49 require_once( "rebuildtextindex.inc" );
  50
  51 class Phase2Importer {
  52         var $olddb, $titleCache;
  53
  54         function Phase2Importer( $database ) {
  55                 $this->olddb = $database;
  56                 $this->titleCache = new TitleCache;
  57         }
  58
  59         function importAll() {
  60                 $this->importCurData();
  61                 $this->fixCurTitles();
  62
  63                 $this->importOldData();
  64                 $this->fixOldTitles();
  65
  66                 $this->importUserData();
  67                 $this->fixUserOptions();
  68
  69                 $this->importWatchlists();
  70
  71                 $this->importLinkData();
  72
  73                 /*
  74                 # For some reason this is broken. RecentChanges will just start anew...
  75                 rebuildRecentChangesTablePass1();
  76                 rebuildRecentChangesTablePass2();
  77                 */
  78
  79                 print "Rebuilding search index:\n";
  80                 dropTextIndex();
  81                 rebuildTextIndex();
  82                 createTextIndex();
  83
  84                 initialiseMessages();
  85         }
  86
  87         # Simple import functions; for the most part these are pretty straightforward.
  88         # MySQL copies everything over to the new database and tweaks a few things.
  89         function importCurData() {
  90                 print "Clearing pages from default install, if any...\n";
  91                 wfQuery( "DELETE FROM cur", DB_WRITE );
  92
  93                 print "Importing current revision data...\n";
  94                 wfQuery( "INSERT INTO cur (cur_id,cur_namespace,cur_title,cur_text,cur_comment,
  95                         cur_user,cur_user_text,cur_timestamp,cur_restrictions,cur_counter,
  96                         cur_is_redirect,cur_minor_edit,cur_is_new,cur_random,cur_touched,inverse_timestamp)
  97                         SELECT cur_id,0,cur_title,cur_text,cur_comment,
  98                                 cur_user,cur_user_text,cur_timestamp,REPLACE(cur_restrictions,'is_',''),cur_counter,
  99                                 cur_text like '#redirect%',cur_minor_edit,0,RAND(),NOW()+0,99999999999999-cur_timestamp
 100                         FROM {$this->olddb}.cur", DB_WRITE );
 101                 $n = mysql_affected_rows();
 102                 print "$n rows imported.\n";
 103         }
 104
 105         function importOldData() {
 106                 print "Clearing old revision data from default install, if any...\n";
 107                 wfQuery( "DELETE FROM old", DB_WRITE );
 108
 109                 print "Importing old revision data...\n";
 110                 wfQuery( "INSERT INTO old (old_id,old_namespace,old_title,old_text,old_comment,
 111                         old_user,old_user_text,old_timestamp,old_minor_edit,old_flags,inverse_timestamp)
 112                         SELECT old_id,0,old_title,old_text,old_comment,
 113                                 old_user,old_user_text,old_timestamp,old_minor_edit,'',99999999999999-old_timestamp
 114                         FROM {$this->olddb}.old", DB_WRITE );
 115                 $n = mysql_affected_rows();
 116                 print "$n rows imported.\n";
 117         }
 118
 119         function importUserData() {
 120                 print "Clearing users from default install, if any...\n";
 121                 wfQuery( "DELETE FROM user", DB_WRITE );
 122
 123                 print "Importing user data...\n";
 124                 wfQuery( "INSERT INTO $newdb.user (user_id,user_name,user_rights,
 125                         user_password,user_newpassword,user_email,user_options,user_touched)
 126                         SELECT user_id,user_name,REPLACE(user_rights,'is_',''),
 127                                 MD5(CONCAT(user_id,'-',MD5(user_password))),'',user_email,user_options,NOW()+0
 128                         FROM {$this->olddb}.user", DB_WRITE );
 129                 $n = mysql_affected_rows();
 130                 print "$n rows imported.\n";
 131         }
 132
 133         # A little less clean...
 134         function importWatchlists() {
 135                 print "Clearing watchlists from default install, if any...\n";
 136                 wfQuery( "DELETE FROM watchlist", DB_WRITE );
 137
 138                 print "Importing watchlists...";
 139                 $res = wfQuery( "SELECT user_id,user_watch FROM {$this->olddb}.user WHERE user_watch != ''", DB_WRITE );
 140                 $total = wfNumRows( $res );
 141                 $n = 0;
 142                 print " ($total total)\n";
 143
 144                 while( $row = wfFetchObject( $res ) ) {
 145                         $id = IntVal( $row->user_id );
 146                         $list = explode( "\n", $row->user_watch );
 147                         foreach( $list as $page ) {
 148                                 $title = $this->titleCache->fetch( $page );
 149                                 if( is_null( $title ) ) {
 150                                         print "Caught bad title '{$row->title}'\n";
 151                                 } else {
 152                                         $ns = $title->getNamespace();
 153                                         $t = wfStrencode( $title->getDBkey() );
 154                                         wfQuery( "INSERT INTO watchlist(wl_user,wl_namespace,wl_title) VALUES ($id,$ns,'$t')", DB_WRITE );
 155                                 }
 156                         }
 157                         if( ++$n % 50 == 0 ) {
 158                                 print "$n\n";
 159                         }
 160                 }
 161                 wfFreeResult( $res );
 162         }
 163
 164         function importLinkData() {
 165                 # MUST BE CALLED BEFORE! fixCurTitles()
 166                 print "Clearing links from default install, if any...\n";
 167                 wfQuery( "DELETE FROM links", DB_WRITE );
 168                 wfQuery( "DELETE FROM brokenlinks", DB_WRITE );
 169
 170                 print "Importing live links...";
 171                 wfQuery( "INSERT INTO links (l_from, l_to)
 172                                         SELECT DISTINCT linked_from,cur_id
 173                                         FROM {$this->olddb}.linked,{$this->olddb}.cur
 174                                         WHERE linked_to=cur_title", DB_WRITE );
 175                 $n = mysql_affected_rows();
 176                 print "$n rows imported.\n";
 177
 178                 print "Importing broken links...";
 179                 wfQuery( "INSERT INTO brokenlinks (bl_from, bl_to)
 180                                         SELECT DISTINCT cur_id,unlinked_to
 181                                         FROM {$this->olddb}.unlinked,{$this->olddb}.cur
 182                                         WHERE unlinked_from=cur_title", DB_WRITE );
 183                 $n = mysql_affected_rows();
 184                 print "$n rows imported.\n";
 185         }
 186
 187         # Fixup functions: munge data that's already been brought into tables
 188         function fixCurTitles() {
 189                 $this->fixTitles( "cur" );
 190         }
 191
 192         function fixOldTitles() {
 193                 $this->fixTitles( "old" );
 194         }
 195
 196         function fixTitles( $table ) {
 197                 print "Fixing titles in $table...";
 198                 $res = wfQuery( "SELECT DISTINCT {$table}_title AS title FROM $table", DB_WRITE );
 199                 $total = wfNumRows( $res );
 200                 $n = 0;
 201                 print " ($total total)\n";
 202
 203                 while( $row = wfFetchObject( $res ) ) {
 204                         $xt = wfStrencode( $row->title );
 205                         $title = $this->titleCache->fetch( $row->title );
 206                         if( is_null( $title ) ) {
 207                                 print "Caught bad title '{$row->title}'\n";
 208                         } else {
 209                                 $ns = $title->getNamespace();
 210                                 $t = wfStrencode( $title->getDBkey() );
 211                                 wfQuery( "UPDATE $table SET {$table}_namespace=$ns,{$table}_title='$t'
 212                                                                 WHERE {$table}_namespace=0 AND {$table}_title='$xt'", DB_WRITE );
 213                         }
 214                         if( ++$n % 50 == 0 ) {
 215                                 print "$n\n";
 216                         }
 217                 }
 218                 wfFreeResult( $res );
 219         }
 220
 221         function rewriteUserOptions( $in )
 222         {
 223                 $s = urldecode( $in );
 224                 $a = explode( "\n", $s );
 225
 226                 foreach ( $a as $l ) {
 227                         if ( preg_match( "/^([A-Za-z0-9_]+)=(.*)/", $l, $m ) ) {
 228                                 $ops[$m[1]] = $m[2];
 229                         }
 230                 }
 231                 $nops = array();
 232
 233                 $q = strtolower( $ops["quickBar"] );
 234                 if ( $q == "none" ) { $q = 0; }
 235                 else { $q = 1; } # Default to left
 236                 $nops["quickbar"] = $q;
 237
 238                 if ( $ops["markupNewTopics"] == "inverse" ) {
 239                         $nops["highlightbroken"] = 1;
 240                 }
 241                 $sk = substr( strtolower( $ops["skin"] ), 0, 4 );
 242                 if ( "star" == $sk ) { $sk = 0; }
 243                 else if ( "nost" == $sk ) { $sk = 1; }
 244                 else if ( "colo" == $sk ) { $sk = 2; }
 245                 else { $sk = 0; }
 246                 $nops["skin"] = $sk;
 247
 248                 $u = strtolower( $ops["underlineLinks"] );
 249                 if ( "yes" == $u || "on" == $u ) { $nops["underline"] = 1; }
 250                 else { $nops["underline"] = 0; }
 251
 252                 $t = ( (int) ($ops["hourDiff"]) );
 253                 if ( $t < -23 || $t > 23 ) { $t = 0; }
 254                 if ( 0 != $t ) { $nops["timecorrection"] = $t; }
 255
 256                 $j = strtolower( $ops["justify"] );
 257                 if ( "yes" == $j || "on" == $j ) { $nops["justify"] = 1; }
 258                 $n = strtolower( $ops["numberHeadings"] );
 259                 if ( "yes" == $n || "on" == $n ) { $nops["numberheadings"] = 1; }
 260                 $h = strtolower( $ops["hideMinor"] );
 261                 if ( "yes" == $h || "on" == $h ) { $nops["hideminor"] = 1; }
 262                 $r = strtolower( $ops["rememberPassword"] );
 263                 if ( "yes" == $r || "on" == $r ) { $nops["rememberpassword"] = 1; }
 264                 $s = strtolower( $ops["showHover"] );
 265                 if ( "yes" == $s || "on" == $s ) { $nops["hover"] = 1; }
 266
 267                 $c = $ops["cols"];
 268                 if ( $c < 20 || c > 200 ) { $nops["cols"] = 80; }
 269                 else { $nops["cols"] = $c; }
 270                 $r = $ops["rows"];
 271                 if ( $r < 5 || $r > 100 ) { $nops["rows"] = 20; }
 272                 else { $nops["rows"] = $r; }
 273                 $r = $ops["resultsPerPage"];
 274                 if ( $r < 3 || $r > 500 ) { $nops["searchlimit"] = 20; }
 275                 else { $nops["searchlimit"] = $r; }
 276                 $r = $ops["viewRecentChanges"];
 277                 if ( $r < 10 || $r > 1000 ) { $nops["rclimit"] = 50; }
 278                 else { $nops["rclimit"] = $r; }
 279                 $nops["rcdays"] = 3;
 280
 281                 $a = array();
 282                 foreach ( $nops as $oname => $oval ) {
 283                         array_push( $a, "$oname=$oval" );
 284                 }
 285                 $s = implode( "\n", $a );
 286                 return $s;
 287         }
 288
 289         function fixUserOptions() {
 290                 print "Fixing user options...";
 291                 $res = wfQuery( "SELECT user_id,user_options FROM user", DB_WRITE );
 292                 $total = wfNumRows( $res );
 293                 $n = 0;
 294                 print " ($total total)\n";
 295
 296                 while( $row = wfFetchObject( $res ) ) {
 297                         $id = IntVal( $row->user_id );
 298                         $option = wfStrencode( $this->rewriteUserOptions( $row->user_options ) );
 299                         wfQuery( "UPDATE user SET user_options='$option' WHERE user_id=$id LIMIT 1", DB_WRITE );
 300                         if( ++$n % 50 == 0 ) {
 301                                 print "$n\n";
 302                         }
 303                 }
 304                 wfFreeResult( $res );
 305         }
 306
 307 }
 308
 309 class TitleCache {
 310         var $hash = array();
 311
 312         function &fetch( $dbkey ) {
 313                 if( !isset( $hash[$dbkey] ) ) {
 314                         $hash[$dbkey] = Title::newFromDBkey( $dbkey );
 315                 }
 316                 return $hash[$dbkey];
 317         }
 318
 319 }
 320
 321 #
 322 print "You should have already run the installer to create a fresh, blank database.\n";
 323 print "Data will be inserted into '$wgDBname'. THIS SHOULD BE EMPTY AND ANY DATA IN IN WILL BE ERASED!\n";
 324 print "\nIf that's not what you want, ABORT NOW!\n\n";
 325
 326 print "Please enter the name of the old 'phase 2'-format database that will be used as a source:\n";
 327 print "Old database name [enciclopedia]: ";
 328 $olddb = readconsole();
 329 if( empty( $olddb ) ) $olddb = "enciclopedia";
 330
 331 if( $olddb == $wgDBname ) {
 332         die( "Can't upgrade in-place! You must create a new database and copy data into it.\n" );
 333 }
 334
 335 print "\nSource database: '$olddb'\n";
 336 print "  Dest database: '$wgDBname'\n";
 337 print "Is this correct? Anything in '$wgDBname' WILL BE DESTROYED. [y/N] ";
 338 $response = readconsole();
 339 if( strtolower( $response{0} ) != 'y' ) {
 340         die( "\nAborted by user.\n" );
 341 }
 342
 343 print "Starting import....\n";
 344
 345 $wgTitle = Title::newFromText( "Conversion script" );
 346 $importer = new Phase2Importer( $olddb );
 347 $importer->importAll();
 348
 349 ?>