namespaceDupes.php

Go to the documentation of this file.

<?php
require_once __DIR__ . '/Maintenance.php';

class NamespaceConflictChecker extends Maintenance {

    protected $db;

    public function __construct() {
        parent::__construct();
        $this->mDescription = "";
        $this->addOption( 'fix', 'Attempt to automatically fix errors' );
        $this->addOption( 'suffix', "Dupes will be renamed with correct namespace with " .
                                    "<text> appended after the article name", false, true );
        $this->addOption( 'prefix', "Do an explicit check for the given title prefix " .
                                    "appended after the article name", false, true );
    }

    public function execute() {
        $this->db = wfGetDB( DB_MASTER );

        $fix = $this->hasOption( 'fix' );
        $suffix = $this->getOption( 'suffix', '' );
        $prefix = $this->getOption( 'prefix', '' );
        $key = intval( $this->getOption( 'key', 0 ) );

        if ( $prefix ) {
            $retval = $this->checkPrefix( $key, $prefix, $fix, $suffix );
        } else {
            $retval = $this->checkAll( $fix, $suffix );
        }

        if ( $retval ) {
            $this->output( "\nLooks good!\n" );
        } else {
            $this->output( "\nOh noeees\n" );
        }
    }

    private function checkAll( $fix, $suffix = '' ) {
        global $wgContLang, $wgNamespaceAliases, $wgCapitalLinks;

        $spaces = array();

        // List interwikis first, so they'll be overridden
        // by any conflicting local namespaces.
        foreach ( $this->getInterwikiList() as $prefix ) {
            $name = $wgContLang->ucfirst( $prefix );
            $spaces[$name] = 0;
        }

        // Now pull in all canonical and alias namespaces...
        foreach ( MWNamespace::getCanonicalNamespaces() as $ns => $name ) {
            // This includes $wgExtraNamespaces
            if ( $name !== '' ) {
                $spaces[$name] = $ns;
            }
        }
        foreach ( $wgContLang->getNamespaces() as $ns => $name ) {
            if ( $name !== '' ) {
                $spaces[$name] = $ns;
            }
        }
        foreach ( $wgNamespaceAliases as $name => $ns ) {
            $spaces[$name] = $ns;
        }
        foreach ( $wgContLang->getNamespaceAliases() as $name => $ns ) {
            $spaces[$name] = $ns;
        }

        // We'll need to check for lowercase keys as well,
        // since we're doing case-sensitive searches in the db.
        foreach ( $spaces as $name => $ns ) {
            $moreNames = array();
            $moreNames[] = $wgContLang->uc( $name );
            $moreNames[] = $wgContLang->ucfirst( $wgContLang->lc( $name ) );
            $moreNames[] = $wgContLang->ucwords( $name );
            $moreNames[] = $wgContLang->ucwords( $wgContLang->lc( $name ) );
            $moreNames[] = $wgContLang->ucwordbreaks( $name );
            $moreNames[] = $wgContLang->ucwordbreaks( $wgContLang->lc( $name ) );
            if ( !$wgCapitalLinks ) {
                foreach ( $moreNames as $altName ) {
                    $moreNames[] = $wgContLang->lcfirst( $altName );
                }
                $moreNames[] = $wgContLang->lcfirst( $name );
            }
            foreach ( array_unique( $moreNames ) as $altName ) {
                if ( $altName !== $name ) {
                    $spaces[$altName] = $ns;
                }
            }
        }

        ksort( $spaces );
        asort( $spaces );

        $ok = true;
        foreach ( $spaces as $name => $ns ) {
            $ok = $this->checkNamespace( $ns, $name, $fix, $suffix ) && $ok;
        }
        return $ok;
    }

    private function getInterwikiList() {
        $result = Interwiki::getAllPrefixes();
        $prefixes = array();
        foreach ( $result as $row ) {
            $prefixes[] = $row['iw_prefix'];
        }
        return $prefixes;
    }

    private function checkNamespace( $ns, $name, $fix, $suffix = '' ) {
        $conflicts = $this->getConflicts( $ns, $name );
        $count = count( $conflicts );
        if ( $count == 0 ) {
            return true;
        }

        $ok = true;
        foreach ( $conflicts as $row ) {
            $resolvable = $this->reportConflict( $row, $suffix );
            $ok = $ok && $resolvable;
            if ( $fix && ( $resolvable || $suffix != '' ) ) {
                $ok = $this->resolveConflict( $row, $resolvable, $suffix ) && $ok;
            }
        }
        return $ok;
    }

    private function checkPrefix( $key, $prefix, $fix, $suffix = '' ) {
        $this->output( "Checking prefix \"$prefix\" vs namespace $key\n" );
        return $this->checkNamespace( $key, $prefix, $fix, $suffix );
    }

    private function getConflicts( $ns, $name ) {
        $page = 'page';
        $table = $this->db->tableName( $page );

        $prefix = $this->db->strencode( $name );
        $encNamespace = $this->db->addQuotes( $ns );

        $titleSql = "TRIM(LEADING '$prefix:' FROM {$page}_title)";
        if ( $ns == 0 ) {
            // An interwiki; try an alternate encoding with '-' for ':'
            $titleSql = $this->db->buildConcat( array( "'$prefix-'", $titleSql ) );
        }

        $sql = "SELECT {$page}_id    AS id,
                       {$page}_title AS oldtitle,
                       $encNamespace + {$page}_namespace AS namespace,
                   $titleSql     AS title,
                   {$page}_namespace AS oldnamespace
                  FROM {$table}
                 WHERE ( {$page}_namespace=0 OR {$page}_namespace=1 )
                   AND {$page}_title " . $this->db->buildLike( $name . ':', $this->db->anyString() );

        $result = $this->db->query( $sql, __METHOD__ );

        $set = array();
        foreach ( $result as $row ) {
            $set[] = $row;
        }
        return $set;
    }

    private function reportConflict( $row, $suffix ) {
        $newTitle = Title::makeTitleSafe( $row->namespace, $row->title );
        if ( is_null( $newTitle ) || !$newTitle->canExist() ) {
            // Title is also an illegal title...
            // For the moment we'll let these slide to cleanupTitles or whoever.
            $this->output( sprintf( "... %d (%d,\"%s\")\n",
                $row->id,
                $row->oldnamespace,
                $row->oldtitle ) );
            $this->output( "...  *** cannot resolve automatically; illegal title ***\n" );
            return false;
        }

        $this->output( sprintf( "... %d (%d,\"%s\") -> (%d,\"%s\") [[%s]]\n",
            $row->id,
            $row->oldnamespace,
            $row->oldtitle,
            $newTitle->getNamespace(),
            $newTitle->getDBkey(),
            $newTitle->getPrefixedText() ) );

        $id = $newTitle->getArticleID();
        if ( $id ) {
            $this->output( "...  *** cannot resolve automatically; page exists with ID $id ***\n" );
            return false;
        } else {
            return true;
        }
    }

    private function resolveConflict( $row, $resolvable, $suffix ) {
        if ( !$resolvable ) {
            $this->output( "...  *** old title {$row->title}\n" );
            while ( true ) {
                $row->title .= $suffix;
                $this->output( "...  *** new title {$row->title}\n" );
                $title = Title::makeTitleSafe( $row->namespace, $row->title );
                if ( !$title ) {
                    $this->output( "... !!! invalid title\n" );
                    return false;
                }
                $id = $title->getArticleID();
                if ( $id ) {
                    $this->output( "...  *** page exists with ID $id ***\n" );
                } else {
                    break;
                }
            }
            $this->output( "...  *** using suffixed form [[" . $title->getPrefixedText() . "]] ***\n" );
        }
        $this->resolveConflictOn( $row, 'page', 'page' );
        return true;
    }

    private function resolveConflictOn( $row, $table, $prefix ) {
        $this->output( "... resolving on $table... " );
        $newTitle = Title::makeTitleSafe( $row->namespace, $row->title );
        $this->db->update( $table,
            array(
                "{$prefix}_namespace" => $newTitle->getNamespace(),
                "{$prefix}_title" => $newTitle->getDBkey(),
            ),
            array(
                // "{$prefix}_namespace" => 0,
                // "{$prefix}_title" => $row->oldtitle,
                "{$prefix}_id" => $row->id,
            ),
            __METHOD__ );
        $this->output( "ok.\n" );
        return true;
    }
}

$maintClass = "NamespaceConflictChecker";
require_once RUN_MAINTENANCE_IF_MAIN;