includes/specials/SpecialMediaStatistics.php

   1 <?php
   2 /**
   3  * Implements Special:MediaStatistics
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup SpecialPage
  22  * @author Brian Wolff
  23  */
  24
  25 /**
  26  * @ingroup SpecialPage
  27  */
  28 class MediaStatisticsPage extends QueryPage {
  29         protected $totalCount = 0, $totalBytes = 0;
  30
  31         function __construct( $name = 'MediaStatistics' ) {
  32                 parent::__construct( $name );
  33                 // Generally speaking there is only a small number of file types,
  34                 // so just show all of them.
  35                 $this->limit = 5000;
  36                 $this->shownavigation = false;
  37         }
  38
  39         function isExpensive() {
  40                 return true;
  41         }
  42
  43         /**
  44          * Query to do.
  45          *
  46          * This abuses the query cache table by storing mime types as "titles".
  47          *
  48          * This will store entries like [[Media:BITMAP;image/jpeg;200;20000]]
  49          * where the form is Media type;mime type;count;bytes.
  50          *
  51          * This relies on the behaviour that when value is tied, the order things
  52          * come out of querycache table is the order they went in. Which is hacky.
  53          * However, other special pages like Special:Deadendpages and
  54          * Special:BrokenRedirects also rely on this.
  55          */
  56         public function getQueryInfo() {
  57                 $dbr = wfGetDB( DB_SLAVE );
  58                 $fakeTitle = $dbr->buildConcat( array(
  59                         'img_media_type',
  60                         $dbr->addQuotes( ';' ),
  61                         'img_major_mime',
  62                         $dbr->addQuotes( '/' ),
  63                         'img_minor_mime',
  64                         $dbr->addQuotes( ';' ),
  65                         'COUNT(*)',
  66                         $dbr->addQuotes( ';' ),
  67                         'SUM( img_size )'
  68                 ) );
  69                 return array(
  70                         'tables' => array( 'image' ),
  71                         'fields' => array(
  72                                 'title' => $fakeTitle,
  73                                 'namespace' => NS_MEDIA, /* needs to be something */
  74                                 'value' => '1'
  75                         ),
  76                         'conds' => array(
  77                                 // WMF has a random null row in the db
  78                                 'img_media_type IS NOT NULL'
  79                         ),
  80                         'options' => array(
  81                                 'GROUP BY' => array(
  82                                         'img_media_type',
  83                                         'img_major_mime',
  84                                         'img_minor_mime',
  85                                 )
  86                         )
  87                 );
  88         }
  89
  90         /**
  91          * How to sort the results
  92          *
  93          * It's important that img_media_type come first, otherwise the
  94          * tables will be fragmented.
  95          * @return Array Fields to sort by
  96          */
  97         function getOrderFields() {
  98                 return array( 'img_media_type', 'count(*)', 'img_major_mime', 'img_minor_mime' );
  99         }
 100
 101         /**
 102          * Output the results of the query.
 103          *
 104          * @param $out OutputPage
 105          * @param $skin Skin (deprecated presumably)
 106          * @param $dbr IDatabase
 107          * @param $res ResultWrapper Results from query
 108          * @param $num integer Number of results
 109          * @param $offset integer Paging offset (Should always be 0 in our case)
 110          */
 111         protected function outputResults( $out, $skin, $dbr, $res, $num, $offset ) {
 112                 $prevMediaType = null;
 113                 foreach ( $res as $row ) {
 114                         $mediaStats = $this->splitFakeTitle( $row->title );
 115                         if ( count( $mediaStats ) < 4 ) {
 116                                 continue;
 117                         }
 118                         list( $mediaType, $mime, $totalCount, $totalBytes ) = $mediaStats;
 119                         if ( $prevMediaType !== $mediaType ) {
 120                                 if ( $prevMediaType !== null ) {
 121                                         // We're not at beginning, so we have to
 122                                         // close the previous table.
 123                                         $this->outputTableEnd();
 124                                 }
 125                                 $this->outputMediaType( $mediaType );
 126                                 $this->outputTableStart( $mediaType );
 127                                 $prevMediaType = $mediaType;
 128                         }
 129                         $this->outputTableRow( $mime, intval( $totalCount ), intval( $totalBytes ) );
 130                 }
 131                 if ( $prevMediaType !== null ) {
 132                         $this->outputTableEnd();
 133                 }
 134         }
 135
 136         /**
 137          * Output closing </table>
 138          */
 139         protected function outputTableEnd() {
 140                 $this->getOutput()->addHtml( Html::closeElement( 'table' ) );
 141         }
 142
 143         /**
 144          * Output a row of the stats table
 145          *
 146          * @param $mime String mime type (e.g. image/jpeg)
 147          * @param $count integer Number of images of this type
 148          * @param $totalBytes integer Total space for images of this type
 149          */
 150         protected function outputTableRow( $mime, $count, $bytes ) {
 151                 $mimeSearch = SpecialPage::getTitleFor( 'MIMEsearch', $mime );
 152                 $row = Html::rawElement(
 153                         'td',
 154                         array(),
 155                         Linker::link( $mimeSearch, htmlspecialchars( $mime ) )
 156                 );
 157                 $row .= Html::element(
 158                         'td',
 159                         array(),
 160                         $this->getExtensionList( $mime )
 161                 );
 162                 $row .= Html::rawElement(
 163                         'td',
 164                         // Make sure js sorts it in numeric order
 165                         array( 'data-sort-value' => $count ),
 166                         $this->msg( 'mediastatistics-nfiles' )
 167                                 ->numParams( $count )
 168                                 /** @todo Check to be sure this really should have number formatting */
 169                                 ->numParams( $this->makePercentPretty( $count / $this->totalCount ) )
 170                                 ->parse()
 171                 );
 172                 $row .= Html::rawElement(
 173                         'td',
 174                         // Make sure js sorts it in numeric order
 175                         array( 'data-sort-value' =>  $bytes ),
 176                         $this->msg( 'mediastatistics-nbytes' )
 177                                 ->numParams( $bytes )
 178                                 ->sizeParams( $bytes )
 179                                 /** @todo Check to be sure this really should have number formatting */
 180                                 ->numParams( $this->makePercentPretty( $bytes / $this->totalBytes ) )
 181                                 ->parse()
 182                 );
 183
 184                 $this->getOutput()->addHTML( Html::rawElement( 'tr', array(), $row ) );
 185         }
 186
 187         /**
 188          * @param float $decimal A decimal percentage (ie for 12.3%, this would be 0.123)
 189          * @return String The percentage formatted so that 3 significant digits are shown.
 190          */
 191         protected function makePercentPretty( $decimal ) {
 192                 $decimal *= 100;
 193                 // Always show three useful digits
 194                 if ( $decimal == 0 ) {
 195                         return '0';
 196                 }
 197                 if ( $decimal >= 100 ) {
 198                         return '100';
 199                 }
 200                 $percent = sprintf( "%." . max( 0, 2 - floor( log10( $decimal ) ) ) . "f", $decimal );
 201                 // Then remove any trailing 0's
 202                 return preg_replace( '/\.?0*$/', '', $percent );
 203         }
 204
 205         /**
 206          * Given a mime type, return a comma separated list of allowed extensions.
 207          *
 208          * @param $mime String mime type
 209          * @return String Comma separated list of allowed extensions (e.g. ".ogg, .oga")
 210          */
 211         private function getExtensionList( $mime ) {
 212                 $exts = MimeMagic::singleton()->getExtensionsForType( $mime );
 213                 if ( $exts === null ) {
 214                         return '';
 215                 }
 216                 $extArray = explode( ' ', $exts );
 217                 $extArray = array_unique( $extArray );
 218                 foreach ( $extArray as &$ext ) {
 219                         $ext = '.' . $ext;
 220                 }
 221
 222                 return $this->getLanguage()->commaList( $extArray );
 223         }
 224
 225         /**
 226          * Output the start of the table
 227          *
 228          * Including opening <table>, and first <tr> with column headers.
 229          */
 230         protected function outputTableStart( $mediaType ) {
 231                 $this->getOutput()->addHTML(
 232                         Html::openElement(
 233                                 'table',
 234                                 array( 'class' => array(
 235                                         'mw-mediastats-table',
 236                                         'mw-mediastats-table-' . strtolower( $mediaType ),
 237                                         'sortable',
 238                                         'wikitable'
 239                                 ) )
 240                         )
 241                 );
 242                 $this->getOutput()->addHTML( $this->getTableHeaderRow() );
 243         }
 244
 245         /**
 246          * Get (not output) the header row for the table
 247          *
 248          * @return String the header row of the able
 249          */
 250         protected function getTableHeaderRow() {
 251                 $headers = array( 'mimetype', 'extensions', 'count', 'totalbytes' );
 252                 $ths = '';
 253                 foreach ( $headers as $header ) {
 254                         $ths .= Html::rawElement(
 255                                 'th',
 256                                 array(),
 257                                 // for grep:
 258                                 // mediastatistics-table-mimetype, mediastatistics-table-extensions
 259                                 // tatistics-table-count, mediastatistics-table-totalbytes
 260                                 $this->msg( 'mediastatistics-table-' . $header )->parse()
 261                         );
 262                 }
 263                 return Html::rawElement( 'tr', array(), $ths );
 264         }
 265
 266         /**
 267          * Output a header for a new media type section
 268          *
 269          * @param $mediaType string A media type (e.g. from the MEDIATYPE_xxx constants)
 270          */
 271         protected function outputMediaType( $mediaType ) {
 272                 $this->getOutput()->addHTML(
 273                         Html::element(
 274                                 'h2',
 275                                 array( 'class' => array(
 276                                         'mw-mediastats-mediatype',
 277                                         'mw-mediastats-mediatype-' . strtolower( $mediaType )
 278                                 ) ),
 279                                 // for grep
 280                                 // mediastatistics-header-unknown, mediastatistics-header-bitmap,
 281                                 // mediastatistics-header-drawing, mediastatistics-header-audio,
 282                                 // mediastatistics-header-video, mediastatistics-header-multimedia,
 283                                 // mediastatistics-header-office, mediastatistics-header-text,
 284                                 // mediastatistics-header-executable, mediastatistics-header-archive,
 285                                 $this->msg( 'mediastatistics-header-' . strtolower( $mediaType ) )->text()
 286                         )
 287                 );
 288                 /** @todo Possibly could add a message here explaining what the different types are.
 289                  *    not sure if it is needed though.
 290                  */
 291         }
 292
 293         /**
 294          * parse the fake title format that this special page abuses querycache with.
 295          *
 296          * @param $fakeTitle String A string formatted as <media type>;<mime type>;<count>;<bytes>
 297          * @return Array The constituant parts of $fakeTitle
 298          */
 299         private function splitFakeTitle( $fakeTitle ) {
 300                 return explode( ';', $fakeTitle, 4 );
 301         }
 302
 303         /**
 304          * What group to put the page in
 305          * @return string
 306          */
 307         protected function getGroupName() {
 308                 return 'media';
 309         }
 310
 311         /**
 312          * This method isn't used, since we override outputResults, but
 313          * we need to implement since abstract in parent class.
 314          *
 315          * @param $skin Skin
 316          * @param $result stdObject Result row
 317          * @return bool|string|void
 318          * @throws MWException
 319          */
 320         public function formatResult( $skin, $result ) {
 321                 throw new MWException( "unimplemented" );
 322         }
 323
 324         /**
 325          * Initialize total values so we can figure out percentages later.
 326          *
 327          * @param $dbr IDatabase
 328          * @param $res ResultWrapper
 329          */
 330         public function preprocessResults( $dbr, $res ) {
 331                 $this->totalCount = $this->totalBytes = 0;
 332                 foreach ( $res as $row ) {
 333                         $mediaStats = $this->splitFakeTitle( $row->title );
 334                         $this->totalCount += isset( $mediaStats[2] ) ? $mediaStats[2] : 0;
 335                         $this->totalBytes += isset( $mediaStats[3] ) ? $mediaStats[3] : 0;
 336                 }
 337                 $res->seek( 0 );
 338         }
 339 }