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         * @var integer $totalPerType Combined file size of all files in a section
  32         */
  33         protected $totalPerType = 0;
  34         /**
  35         * @var integer $totalSize Combined file size of all files
  36         */
  37         protected $totalSize = 0;
  38
  39         function __construct( $name = 'MediaStatistics' ) {
  40                 parent::__construct( $name );
  41                 // Generally speaking there is only a small number of file types,
  42                 // so just show all of them.
  43                 $this->limit = 5000;
  44                 $this->shownavigation = false;
  45         }
  46
  47         public function isExpensive() {
  48                 return true;
  49         }
  50
  51         /**
  52          * Query to do.
  53          *
  54          * This abuses the query cache table by storing mime types as "titles".
  55          *
  56          * This will store entries like [[Media:BITMAP;image/jpeg;200;20000]]
  57          * where the form is Media type;mime type;count;bytes.
  58          *
  59          * This relies on the behaviour that when value is tied, the order things
  60          * come out of querycache table is the order they went in. Which is hacky.
  61          * However, other special pages like Special:Deadendpages and
  62          * Special:BrokenRedirects also rely on this.
  63          */
  64         public function getQueryInfo() {
  65                 $dbr = wfGetDB( DB_REPLICA );
  66                 $fakeTitle = $dbr->buildConcat( [
  67                         'img_media_type',
  68                         $dbr->addQuotes( ';' ),
  69                         'img_major_mime',
  70                         $dbr->addQuotes( '/' ),
  71                         'img_minor_mime',
  72                         $dbr->addQuotes( ';' ),
  73                         'COUNT(*)',
  74                         $dbr->addQuotes( ';' ),
  75                         'SUM( img_size )'
  76                 ] );
  77                 return [
  78                         'tables' => [ 'image' ],
  79                         'fields' => [
  80                                 'title' => $fakeTitle,
  81                                 'namespace' => NS_MEDIA, /* needs to be something */
  82                                 'value' => '1'
  83                         ],
  84                         'conds' => [
  85                                 // WMF has a random null row in the db
  86                                 'img_media_type IS NOT NULL'
  87                         ],
  88                         'options' => [
  89                                 'GROUP BY' => [
  90                                         'img_media_type',
  91                                         'img_major_mime',
  92                                         'img_minor_mime',
  93                                 ]
  94                         ]
  95                 ];
  96         }
  97
  98         /**
  99          * How to sort the results
 100          *
 101          * It's important that img_media_type come first, otherwise the
 102          * tables will be fragmented.
 103          * @return Array Fields to sort by
 104          */
 105         function getOrderFields() {
 106                 return [ 'img_media_type', 'count(*)', 'img_major_mime', 'img_minor_mime' ];
 107         }
 108
 109         /**
 110          * Output the results of the query.
 111          *
 112          * @param OutputPage $out
 113          * @param Skin $skin (deprecated presumably)
 114          * @param IDatabase $dbr
 115          * @param ResultWrapper $res Results from query
 116          * @param int $num Number of results
 117          * @param int $offset Paging offset (Should always be 0 in our case)
 118          */
 119         protected function outputResults( $out, $skin, $dbr, $res, $num, $offset ) {
 120                 $prevMediaType = null;
 121                 foreach ( $res as $row ) {
 122                         $mediaStats = $this->splitFakeTitle( $row->title );
 123                         if ( count( $mediaStats ) < 4 ) {
 124                                 continue;
 125                         }
 126                         list( $mediaType, $mime, $totalCount, $totalBytes ) = $mediaStats;
 127                         if ( $prevMediaType !== $mediaType ) {
 128                                 if ( $prevMediaType !== null ) {
 129                                         // We're not at beginning, so we have to
 130                                         // close the previous table.
 131                                         $this->outputTableEnd();
 132                                 }
 133                                 $this->outputMediaType( $mediaType );
 134                                 $this->totalPerType = 0;
 135                                 $this->outputTableStart( $mediaType );
 136                                 $prevMediaType = $mediaType;
 137                         }
 138                         $this->outputTableRow( $mime, intval( $totalCount ), intval( $totalBytes ) );
 139                 }
 140                 if ( $prevMediaType !== null ) {
 141                         $this->outputTableEnd();
 142                         // add total size of all files
 143                         $this->outputMediaType( 'total' );
 144                         $this->getOutput()->addWikiText(
 145                                 $this->msg( 'mediastatistics-allbytes' )
 146                                         ->numParams( $this->totalSize )
 147                                         ->sizeParams( $this->totalSize )
 148                                         ->text()
 149                         );
 150                 }
 151         }
 152
 153         /**
 154          * Output closing </table>
 155          */
 156         protected function outputTableEnd() {
 157                 $this->getOutput()->addHTML( Html::closeElement( 'table' ) );
 158                 $this->getOutput()->addWikiText(
 159                                 $this->msg( 'mediastatistics-bytespertype' )
 160                                         ->numParams( $this->totalPerType )
 161                                         ->sizeParams( $this->totalPerType )
 162                                         ->numParams( $this->makePercentPretty( $this->totalPerType / $this->totalBytes ) )
 163                                         ->text()
 164                 );
 165                 $this->totalSize += $this->totalPerType;
 166         }
 167
 168         /**
 169          * Output a row of the stats table
 170          *
 171          * @param string $mime mime type (e.g. image/jpeg)
 172          * @param int $count Number of images of this type
 173          * @param int $totalBytes Total space for images of this type
 174          */
 175         protected function outputTableRow( $mime, $count, $bytes ) {
 176                 $mimeSearch = SpecialPage::getTitleFor( 'MIMEsearch', $mime );
 177                 $row = Html::rawElement(
 178                         'td',
 179                         [],
 180                         Linker::link( $mimeSearch, htmlspecialchars( $mime ) )
 181                 );
 182                 $row .= Html::element(
 183                         'td',
 184                         [],
 185                         $this->getExtensionList( $mime )
 186                 );
 187                 $row .= Html::rawElement(
 188                         'td',
 189                         // Make sure js sorts it in numeric order
 190                         [ 'data-sort-value' => $count ],
 191                         $this->msg( 'mediastatistics-nfiles' )
 192                                 ->numParams( $count )
 193                                 /** @todo Check to be sure this really should have number formatting */
 194                                 ->numParams( $this->makePercentPretty( $count / $this->totalCount ) )
 195                                 ->parse()
 196                 );
 197                 $row .= Html::rawElement(
 198                         'td',
 199                         // Make sure js sorts it in numeric order
 200                         [ 'data-sort-value' =>  $bytes ],
 201                         $this->msg( 'mediastatistics-nbytes' )
 202                                 ->numParams( $bytes )
 203                                 ->sizeParams( $bytes )
 204                                 /** @todo Check to be sure this really should have number formatting */
 205                                 ->numParams( $this->makePercentPretty( $bytes / $this->totalBytes ) )
 206                                 ->parse()
 207                 );
 208                 $this->totalPerType += $bytes;
 209                 $this->getOutput()->addHTML( Html::rawElement( 'tr', [], $row ) );
 210         }
 211
 212         /**
 213          * @param float $decimal A decimal percentage (ie for 12.3%, this would be 0.123)
 214          * @return String The percentage formatted so that 3 significant digits are shown.
 215          */
 216         protected function makePercentPretty( $decimal ) {
 217                 $decimal *= 100;
 218                 // Always show three useful digits
 219                 if ( $decimal == 0 ) {
 220                         return '0';
 221                 }
 222                 if ( $decimal >= 100 ) {
 223                         return '100';
 224                 }
 225                 $percent = sprintf( "%." . max( 0, 2 - floor( log10( $decimal ) ) ) . "f", $decimal );
 226                 // Then remove any trailing 0's
 227                 return preg_replace( '/\.?0*$/', '', $percent );
 228         }
 229
 230         /**
 231          * Given a mime type, return a comma separated list of allowed extensions.
 232          *
 233          * @param string $mime mime type
 234          * @return string Comma separated list of allowed extensions (e.g. ".ogg, .oga")
 235          */
 236         private function getExtensionList( $mime ) {
 237                 $exts = MimeMagic::singleton()->getExtensionsForType( $mime );
 238                 if ( $exts === null ) {
 239                         return '';
 240                 }
 241                 $extArray = explode( ' ', $exts );
 242                 $extArray = array_unique( $extArray );
 243                 foreach ( $extArray as &$ext ) {
 244                         $ext = '.' . $ext;
 245                 }
 246
 247                 return $this->getLanguage()->commaList( $extArray );
 248         }
 249
 250         /**
 251          * Output the start of the table
 252          *
 253          * Including opening <table>, and first <tr> with column headers.
 254          */
 255         protected function outputTableStart( $mediaType ) {
 256                 $this->getOutput()->addHTML(
 257                         Html::openElement(
 258                                 'table',
 259                                 [ 'class' => [
 260                                         'mw-mediastats-table',
 261                                         'mw-mediastats-table-' . strtolower( $mediaType ),
 262                                         'sortable',
 263                                         'wikitable'
 264                                 ] ]
 265                         )
 266                 );
 267                 $this->getOutput()->addHTML( $this->getTableHeaderRow() );
 268         }
 269
 270         /**
 271          * Get (not output) the header row for the table
 272          *
 273          * @return String the header row of the able
 274          */
 275         protected function getTableHeaderRow() {
 276                 $headers = [ 'mimetype', 'extensions', 'count', 'totalbytes' ];
 277                 $ths = '';
 278                 foreach ( $headers as $header ) {
 279                         $ths .= Html::rawElement(
 280                                 'th',
 281                                 [],
 282                                 // for grep:
 283                                 // mediastatistics-table-mimetype, mediastatistics-table-extensions
 284                                 // tatistics-table-count, mediastatistics-table-totalbytes
 285                                 $this->msg( 'mediastatistics-table-' . $header )->parse()
 286                         );
 287                 }
 288                 return Html::rawElement( 'tr', [], $ths );
 289         }
 290
 291         /**
 292          * Output a header for a new media type section
 293          *
 294          * @param string $mediaType A media type (e.g. from the MEDIATYPE_xxx constants)
 295          */
 296         protected function outputMediaType( $mediaType ) {
 297                 $this->getOutput()->addHTML(
 298                         Html::element(
 299                                 'h2',
 300                                 [ 'class' => [
 301                                         'mw-mediastats-mediatype',
 302                                         'mw-mediastats-mediatype-' . strtolower( $mediaType )
 303                                 ] ],
 304                                 // for grep
 305                                 // mediastatistics-header-unknown, mediastatistics-header-bitmap,
 306                                 // mediastatistics-header-drawing, mediastatistics-header-audio,
 307                                 // mediastatistics-header-video, mediastatistics-header-multimedia,
 308                                 // mediastatistics-header-office, mediastatistics-header-text,
 309                                 // mediastatistics-header-executable, mediastatistics-header-archive,
 310                                 $this->msg( 'mediastatistics-header-' . strtolower( $mediaType ) )->text()
 311                         )
 312                 );
 313                 /** @todo Possibly could add a message here explaining what the different types are.
 314                  *    not sure if it is needed though.
 315                  */
 316         }
 317
 318         /**
 319          * parse the fake title format that this special page abuses querycache with.
 320          *
 321          * @param string $fakeTitle A string formatted as <media type>;<mime type>;<count>;<bytes>
 322          * @return array The constituant parts of $fakeTitle
 323          */
 324         private function splitFakeTitle( $fakeTitle ) {
 325                 return explode( ';', $fakeTitle, 4 );
 326         }
 327
 328         /**
 329          * What group to put the page in
 330          * @return string
 331          */
 332         protected function getGroupName() {
 333                 return 'media';
 334         }
 335
 336         /**
 337          * This method isn't used, since we override outputResults, but
 338          * we need to implement since abstract in parent class.
 339          *
 340          * @param Skin $skin
 341          * @param stdObject $result Result row
 342          * @return bool|string|void
 343          * @throws MWException
 344          */
 345         public function formatResult( $skin, $result ) {
 346                 throw new MWException( "unimplemented" );
 347         }
 348
 349         /**
 350          * Initialize total values so we can figure out percentages later.
 351          *
 352          * @param IDatabase $dbr
 353          * @param ResultWrapper $res
 354          */
 355         public function preprocessResults( $dbr, $res ) {
 356                 $this->executeLBFromResultWrapper( $res );
 357                 $this->totalCount = $this->totalBytes = 0;
 358                 foreach ( $res as $row ) {
 359                         $mediaStats = $this->splitFakeTitle( $row->title );
 360                         $this->totalCount += isset( $mediaStats[2] ) ? $mediaStats[2] : 0;
 361                         $this->totalBytes += isset( $mediaStats[3] ) ? $mediaStats[3] : 0;
 362                 }
 363                 $res->seek( 0 );
 364         }
 365 }