| Index: trunk/extensions/ActiveAbstract/AbstractFilter.php |
| — | — | @@ -0,0 +1,171 @@ |
| | 2 | +<?php |
| | 3 | + |
| | 4 | +/** |
| | 5 | + * Generate XML feed for Yahoo's Active Abstracts project |
| | 6 | + * Plugin for dumpBackup.php; call as eg: |
| | 7 | + * |
| | 8 | + * php dumpBackup.php \ |
| | 9 | + * --plugin=AbstractFilter:extensions/ActiveAbstract/AbstractFilter.php \ |
| | 10 | + * --current \ |
| | 11 | + * --output=gzip:/dumps/abstract.xml.gz \ |
| | 12 | + * --filter=namespace:NS_MAIN \ |
| | 13 | + * --filter=noredirect \ |
| | 14 | + * --filter=abstract |
| | 15 | + */ |
| | 16 | + |
| | 17 | +require_once 'includes/EditPage.php'; // hack; for section anchor code |
| | 18 | + |
| | 19 | +/** |
| | 20 | + * Tosses away the MediaWiki XML and generates new output |
| | 21 | + */ |
| | 22 | +class AbstractFilter { |
| | 23 | + /** |
| | 24 | + * Register the filter function with the dump manager |
| | 25 | + * @param BackupDumper $dumper |
| | 26 | + * @static |
| | 27 | + */ |
| | 28 | + function register( &$dumper ) { |
| | 29 | + $dumper->registerFilter( 'abstract', 'AbstractFilter' ); |
| | 30 | + $dumper->registerFilter( 'noredirect', 'NoredirectFilter' ); |
| | 31 | + } |
| | 32 | + |
| | 33 | + function AbstractFilter( &$sink ) { |
| | 34 | + $this->sink =& $sink; |
| | 35 | + } |
| | 36 | + |
| | 37 | + function writeOpenStream( $string ) { |
| | 38 | + $this->sink->writeOpenStream( "<feed>\n" ); |
| | 39 | + } |
| | 40 | + |
| | 41 | + function writeCloseStream( $string ) { |
| | 42 | + $this->sink->writeCloseStream( "</feed>\n" ); |
| | 43 | + } |
| | 44 | + |
| | 45 | + function writeOpenPage( $page, $string ) { |
| | 46 | + global $wgSitename; |
| | 47 | + $this->title = Title::makeTitle( $page->page_namespace, $page->page_title ); |
| | 48 | + |
| | 49 | + $xml = "<doc>\n"; |
| | 50 | + $xml .= wfElement( 'url', null, $this->title->getFullUrl() ) . "\n"; |
| | 51 | + $xml .= wfElement( 'title', null, $wgSitename . ': ' . $this->title->getPrefixedText() ) . "\n"; |
| | 52 | + |
| | 53 | + // add abstract and links when we have revision data... |
| | 54 | + $this->revision = null; |
| | 55 | + |
| | 56 | + $this->sink->writeOpenPage( $page, $xml ); |
| | 57 | + } |
| | 58 | + |
| | 59 | + function writeClosePage( $string ) { |
| | 60 | + $xml = ''; |
| | 61 | + if( $this->revision ) { |
| | 62 | + $xml .= wfElement( 'abstract', null, $this->_abstract( $this->revision ) ) . "\n"; |
| | 63 | + $xml .= "<links>\n"; |
| | 64 | + foreach( $this->_links( $this->revision ) as $url ) { |
| | 65 | + $xml .= wfElement( 'link', null, $url ) . "\n"; |
| | 66 | + } |
| | 67 | + $xml .= "</links>\n"; |
| | 68 | + } |
| | 69 | + $xml .= "</doc>\n"; |
| | 70 | + $this->sink->writeClosePage( $xml ); |
| | 71 | + $this->title = null; |
| | 72 | + $this->revision = null; |
| | 73 | + } |
| | 74 | + |
| | 75 | + function writeRevision( $rev, $string ) { |
| | 76 | + // Only use one revision's worth of data to output |
| | 77 | + $this->revision = $rev; |
| | 78 | + } |
| | 79 | + |
| | 80 | + /** |
| | 81 | + * Extract an abstract from the page |
| | 82 | + * @params object $rev Database rows with revision data |
| | 83 | + * @return string |
| | 84 | + * @access private |
| | 85 | + */ |
| | 86 | + function _abstract( $rev ) { |
| | 87 | + $text = Revision::getRevisionText( $rev ); // FIXME cache this |
| | 88 | + |
| | 89 | + $stripped = $this->_stripMarkup( $text ); |
| | 90 | + $extract = $this->_extractStart( $stripped ); |
| | 91 | + |
| | 92 | + return substr( $extract, 0, 1024 ); // not too long pls |
| | 93 | + } |
| | 94 | + |
| | 95 | + /** |
| | 96 | + * Strip markup to show plaintext |
| | 97 | + * @param string $text |
| | 98 | + * @return string |
| | 99 | + * @access private |
| | 100 | + */ |
| | 101 | + function _stripMarkup( $text ) { |
| | 102 | + $text = str_replace( "'''", "", $text ); |
| | 103 | + $text = str_replace( "''", "", $text ); |
| | 104 | + $text = preg_replace( '#<!--.*?-->#s', '', $text ); // HTML-style comments |
| | 105 | + $text = preg_replace( '#</?[a-z0-9]+.*?>#s', '', $text ); // HTML-style tags |
| | 106 | + $text = preg_replace( '#\\[[a-z]+:.*? (.*?)\\]#s', '$1', $text ); // URL links |
| | 107 | + $text = preg_replace( '#\\{\\{\\{.*?\\}\\}\\}#s', '', $text ); // template parameters |
| | 108 | + $text = preg_replace( '#\\{\\{.*?\\}\\}#s', '', $text ); // template calls |
| | 109 | + $text = preg_replace( '#\\[\\[([^|\\]]*\\|)?(.*?)\\]\\]#s', '$2', $text ); // links |
| | 110 | + $text = Sanitizer::decodeCharReferences( $text ); |
| | 111 | + return trim( $text ); |
| | 112 | + } |
| | 113 | + |
| | 114 | + /** |
| | 115 | + * Extract the first two sentences, if detectable, from the text. |
| | 116 | + * @param string $text |
| | 117 | + * @return string |
| | 118 | + * @access private |
| | 119 | + */ |
| | 120 | + function _extractStart( $text ) { |
| | 121 | + $endchars = array( |
| | 122 | + '.', '!', '?', // regular ASCII |
| | 123 | + '。', // full-width ideographic full-stop |
| | 124 | + '．', '！', '？', // double-width roman forms |
| | 125 | + '｡', // half-width ideographic full stop |
| | 126 | + ); |
| | 127 | + |
| | 128 | + $endgroup = implode( '', array_map( 'preg_quote', $endchars ) ); |
| | 129 | + $end = "[$endgroup]"; |
| | 130 | + $sentence = ".*?$end+"; |
| | 131 | + $firsttwo = "/^($sentence$sentence)/"; |
| | 132 | + |
| | 133 | + if( preg_match( $firsttwo, $text, $matches ) ) { |
| | 134 | + return $matches[1]; |
| | 135 | + } else { |
| | 136 | + return $text; |
| | 137 | + } |
| | 138 | + } |
| | 139 | + |
| | 140 | + /** |
| | 141 | + * Extract a list of TOC links |
| | 142 | + * @params object $rev Database rows with revision data |
| | 143 | + * @return array of URL strings |
| | 144 | + * @access private |
| | 145 | + * @fixme extract TOC items |
| | 146 | + */ |
| | 147 | + function _links( $rev ) { |
| | 148 | + $text = Revision::getRevisionText( $rev ); |
| | 149 | + $secs = |
| | 150 | + preg_split( |
| | 151 | + '/(^=+.+?=+|^<h[1-6].*?' . '>.*?<\/h[1-6].*?' . '>)(?!\S)/mi', |
| | 152 | + $text, -1, |
| | 153 | + PREG_SPLIT_DELIM_CAPTURE ); |
| | 154 | + |
| | 155 | + $headers = array(); |
| | 156 | + for( $i = 1; $i < count( $secs ); $i += 2 ) { |
| | 157 | + $header = preg_replace( '/^=+\s*(.*?)\s*=+/', '$1', $secs[$i] ); |
| | 158 | + $anchor = EditPage::sectionAnchor( $header ); |
| | 159 | + $url = $this->title->getFullUrl() . $anchor; |
| | 160 | + $headers[] = $url; |
| | 161 | + } |
| | 162 | + return $headers; |
| | 163 | + } |
| | 164 | +} |
| | 165 | + |
| | 166 | +class NoredirectFilter extends DumpFilter { |
| | 167 | + function pass( $page, $string ) { |
| | 168 | + return !$page->page_is_redirect; |
| | 169 | + } |
| | 170 | +} |
| | 171 | + |
| | 172 | +?> |
| \ No newline at end of file |
| Property changes on: trunk/extensions/ActiveAbstract/AbstractFilter.php |
| ___________________________________________________________________ |
| Added: svn:keywords |
| 1 | 173 | + Author Date Id Revision |
| Added: svn:eol-style |
| 2 | 174 | + native |