diff --git a/includes/boz-mw b/includes/boz-mw index 5279cb6..619f053 160000 --- a/includes/boz-mw +++ b/includes/boz-mw @@ -1 +1 @@ -Subproject commit 5279cb690357c8fc0b0f25285b0276d983fb4372 +Subproject commit 619f053eb0dbeb4121b6830f3aa4a2516d2e2474 diff --git a/includes/class-PDCs.php b/includes/class-PDCs.php index 70a5ed6..a001a3b 100644 --- a/includes/class-PDCs.php +++ b/includes/class-PDCs.php @@ -1,209 +1,182 @@ . namespace itwikidelbot; use DateTime; use wm\WikipediaIt; +use mw\API\PageMatcher; /** * Handle some PDCs */ class PDCs { /** * Sort some PDCs by creation date * * @param $pdcs array */ public static function sortByCreationDate( & $pdcs ) { // sort PDCs by start date usort( $pdcs, function ( $a, $b ) { return $a->getCreationDate() > $b->getCreationDate(); } ); } /** * Index some PDCs by their type * * @param $pdcs array * @return array */ public static function indexByType( $pdcs ) { $pdcs_by_type = []; foreach( CategoryYearMonthDayTypes::all() as $Type ) { $pdcs_by_type[ $Type::PDC_TYPE ] = []; } foreach( $pdcs as $pdc ) { $type = $pdc->getType(); $pdcs_by_type[ $type ][] = $pdc; } return $pdcs_by_type; } /** * Discard all the PDCs that do not belong to a certain date * * @param $pdcs array * @param $date DateTime * @return array */ public static function filterByDate( $pdcs, DateTime $date ) { $y_m_d = $date->format( 'Y-m-d' ); return array_filter( $pdcs, function ( $pdc ) use ( $y_m_d ) { return $pdc->getStartDate()->format( 'Y-m-d' ) === $y_m_d; } ); } /** * Discard all the PDCs that are not multiple * * @param $pdcs array * @return $pdcs */ public static function filterNotMultiple( $pdcs ) { return array_filter( $pdcs, function ( $pdc ) { return ! $pdc->isMultiple(); } ); } /** * Populate the PDC 'subject themes' field of some PDCs * * This method is intended to do less API requests as possible. * * @param $pdcs array * @return array * @see PDC::getSubjectThemes() */ public static function populateSubjectThemes( $pdcs ) { // Pattern to match spaces, newline, tabulations $_ = '[ \t\n]*'; /* * Pattern to match a PDC turnover (actually this information is unuseful) * * {{cancellazione|9}} * or explicitly: * {{cancellazione|1 = 9}} */ $PARAM_NUM = '(?:' . '\|' . $_ . '(?:' . '1' . $_ . '=' . $_ . ')?' . // 1 = '[0-9]+' . $_ . // turnover number ')?'; /* * Pattern to match a PDC subject theme * * {{cancellazione|arg = something}} * or * {{cancellazione|arg2 = something}} */ $PARAM_THEME = '(?:' . '\|' . $_ . 'arg2?' . $_ . '=' . $_ . '([0-9a-zA-ZàèìòùÀÈÌÒÙ\-\/\'_. ]+?)' . $_ . ')?'; /* * Complete pattern to match all the PDC arguments * * {{Cancellazione|9|arg=something|arg2=something}} * * This pattern is a bit repetitive because PCRE does not support to * match a group multiple times. * Yes, every group can be repeated, but it will be matched only once. * * @TODO: use a wikitext parser */ $PATTERN = '/' . '{{' . $_ . '[Cc]ancellazione' . $_ . $PARAM_NUM . $PARAM_THEME . $PARAM_THEME . '}}/'; // only non multiple PDCs $pdcs = self::filterNotMultiple( $pdcs ); // subject titles $titles = array_map( function ( $pdc ) { return $pdc->getTitleSubject(); }, $pdcs ); // API to retrieve the wikitext in the first section of all the pages $site = WikipediaIt::getInstance(); $query = $site->createQuery( [ 'action' => 'query', 'titles' => $titles, 'prop' => 'revisions', 'rvprop' => 'content', 'rvsection' => '0', ] ); foreach( $query->getGenerator() as $response ) { - - // all normalized PDC titles - $normalizeds = []; - if( isset( $response->query->normalized ) ) { - foreach( $response->query->normalized as $fromto ) { - $normalizeds[ $fromto->from ] = $fromto->to; - } - } - - foreach( $response->query->pages as $id => $page ) { - - // avoid unexisting pages - if( ! isset( $page->revisions ) ) { - continue; - } - - // page content (only the section 0) - $page_content = $page->revisions[ 0 ]->{ '*' }; - - // find the matching PDC - foreach( $pdcs as $i => $pdc ) { - - // normalized PDC title - $title = $pdc->getTitleSubject(); - if( isset( $normalizeds[ $title ] ) ) { - $title = $normalizeds[ $title ]; - } - - // does the normalized PDC title match the page title? - if( $title === $page->title ) { - - // find the {{cancellazione|arg=|arg2=}} + ( new PageMatcher( $response->query, $pdcs ) )->matchByTitle( + // callback fired for every match between response pages and PDCs + function ( $page, $pdc ) use ( $PATTERN ) { + // find the {{cancellazione|arg=|arg2=}} + if( isset( $page->revisions ) ) { + $page_content = $page->revisions[ 0 ]->{ '*' }; preg_match( $PATTERN, $page_content, $matches ); for( $j = 1; $j < count( $matches ); $j++ ) { $pdc->addSubjectTheme( trim( $matches[ $j ] ) ); } - - // do not fill again the themes of this PDC - unset( $pdcs[ $i ] ); - - // next page - break; } + }, + // callback that must returns the PDC page title + function ( $pdc ) { + return $pdc->getTitleSubject(); } - } + ); } } }