diff --git a/orphanize.php b/orphanize.php index 6080d5a..32ddacb 100755 --- a/orphanize.php +++ b/orphanize.php @@ -1,212 +1,221 @@ #!/usr/bin/php . namespace itwikidelbot; // report every error -error_reporting( E_STRICT ); +//error_reporting( E_STRICT ); // autoload classes and configuration require __DIR__ . DIRECTORY_SEPARATOR . 'includes' . DIRECTORY_SEPARATOR . 'autoload.php'; require __DIR__ . DIRECTORY_SEPARATOR . 'config.php'; // how much titles at time requesting - this is a MediaWiki limit define( 'MAX_TRANCHE_TITLES', 50 ); //$TITLE_SOURCE = 'Utente:.avgas/Wikilink da orfanizzare'; $TITLE_SOURCE = 'Utente:Valerio Bozzolan'; // classes used use \cli\Log; use \web\MediaWikis; +use \mw\Wikilink; // wiki identifier $wiki_uid = 'itwiki'; // wiki instance $wiki = Mediawikis::findFromUid( $wiki_uid ); // query last revision $revision = $wiki->fetch( [ 'action' => 'query', 'prop' => 'revisions', 'titles' => $TITLE_SOURCE, 'rvslots' => 'main', 'rvprop' => 'content', 'rvlimit' => 1, ] ); // array of page titles to be orphanized $involved_pagetitles = []; // associative array of page IDs as key and a boolean as value containg pages to be orphanized $involved_pageids = []; // for each page (well, just one) foreach( $revision->query->pages as $sourcepage ) { // for each revision (well, just one) foreach( $sourcepage->revisions as $revision ) { // pure wikitext $wikitextRaw = $revision->slots->main->{ '*' }; // wikitext object $wikitext = $wiki->createWikitext( $revision->slots->main->{ '*' } ); // identify wikilinks $n = $wikitext->pregMatchAll( '~\[\[(.*?)\]\]~', $matches ); if( $n === false ) { Log::error( 'wtf' ); exit( 1 ); } // collect these titles $titles_to_be_orphanized = []; for( $i = 0; $i < $n; $i++ ) { // can be both 'title' and 'title|alias' $wlink = $matches[ 1 ][ $i ]; // just the page title $title = explode( '|', $wlink )[0]; // normalize titles $title = str_replace( '_', ' ', ucfirst( $title ) ); // append $titles_to_be_orphanized[] = $title; } // drop duplicates $titles_to_be_orphanized = array_unique( $titles_to_be_orphanized ); // order sort( $titles_to_be_orphanized, SORT_STRING ); // keep a copy $involved_pagetitles = $titles_to_be_orphanized; // log titles Log::info( "Pages to be orphanized:" ); foreach( $titles_to_be_orphanized as $title ) { Log::info( $title ); } // note that the API accepts a maximum trance of titles while( $less_titles_to_be_orphanized = array_splice( $titles_to_be_orphanized, 0, MAX_TRANCHE_TITLES ) ) { // for each of these titles, query linksto $linksto = $wiki->createQuery( [ 'action' => 'query', 'titles' => $less_titles_to_be_orphanized, 'prop' => 'linkshere', 'lhprop' => 'pageid', 'lhnamespace' => 0, 'lhlimit' => 500, ] ); // cumulate the linkshere page ids Log::info( "requesting linkshere..." ); foreach( $linksto as $response ) { foreach( $response->query->pages as $page ) { if( isset( $page->linkshere ) ) { foreach( $page->linkshere as $linkingpage ) { $pageid = (int) $linkingpage->pageid; $involved_pageids[ $pageid ] = false; } } } } } } } // create a clean array or page ids $involved_pageids = array_keys( $involved_pageids ); // count of involved pages Log::info( sprintf( "found %d pages containing the %d involved wlinks", count( $involved_pageids ), count( $involved_pagetitles ) ) ); // note that the API accepts a maximum tranche of IDs while( $less_involved_pageids = array_splice( $involved_pageids, 0, MAX_TRANCHE_TITLES ) ) { // query last revision - $queries = + $responses = $wiki->createQuery( [ 'action' => 'query', 'pageids' => $less_involved_pageids, 'prop' => 'revisions', 'rvslots' => 'main', 'rvprop' => 'content', ] ); - // for each revision query - foreach( $queries as $query ) { + // for each response + foreach( $responses as $response ) { - // page wikitext raw - $wikitext_raw = $page->revisions[ 0 ]->revision->slots->main->{ '*' }; + // for each page + foreach( $response->query->pages as $page ) { - // wikitext object - $wikitext = $wiki->createWikitext( $wikitext_raw ); + // does it have a revision? + if( isset( $page->revisions[ 0 ] ) ) { - // orphanize these titles - foreach( $involved_pagetitles as $title_complete ) { + // get wikitext from the main slot + $wikitext_raw = $page->revisions[ 0 ]->slots->main->{ '*' }; - // spaces can be underscores - $title_complete = str_replace( ' ', '[ _]', $title_complete ); + // create a Wikitext object + $wikitext = $wiki->createWikitext( $wikitext_raw ); - // the title can contain a namespace - $ns = ''; - $parts = explode( ':', $title_complete, 2 ); - if( count( $parts ) === 2 ) { - $ns = $parts[ 0 ]; - $title = $parts[ 1 ]; + // for each of the titles to be orphanized + foreach( $involved_pagetitles as $involved_pagetitle ) { - if( $ns ) { - $ns .= "$ns:"; - } - } + // parse the orphanizing title + $title = $wiki->createTitleParsing( $involved_pagetitle ); - // verify the namespace - if( $ns ) { - if( $ns_object = $wiki->findNamespace( $ns ) ) { - $ns = $ns_object->getRegex(); - } else { - // Oh, that was not a namespace but title part - $title = "$ns:$title"; - $ns = ''; - } - } + // a wikilink without alias + $wikilink_simple = $wiki->createWikilink( $title, Wikilink::NO_ALIAS ); + + // a wikilink with whatever alias + $wikilink_alias = $wiki->createWikilink( $title, Wikilink::WHATEVER_ALIAS ); + + // sobstitute simple links e.g. [[Hello]] + $wikilink_regex_simple = $wikilink_simple->getRegex( [ + 'title-group-name' => 'title' + ] ); - // sanitize title and namespace - $ns = preg_quote( $ns ); - $title = preg_quote( $title ); + // sobstitute links with alias e.g. [[Hello|whatever]] + $wikilink_regex_alias = $wikilink_alias->getRegex( [ + 'alias-group-name' => 'alias' + ] ); - $wikitext->pregReplace( - '/\[\[ *' . $title . ' *\]\]/' - ); + // convert '[[Hello]]' to 'Hello' + $wikitext->pregReplace( + "/$wikilink_regex_simple/", + \regex\Generic::groupName( 'title' ) + ); + + // convert '[[Hello|world]]' to 'world' + $wikitext->pregReplace( + "/$wikilink_regex_alias/", + \regex\Generic::groupName( 'alias' ) + ); + + Log::info( implode( "; ", $wikitext->getHumanUniqueSobstitutions() ) ); + + // @TODO: save! + } + } } } - }