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 use Wikimedia\Rdbms\ResultWrapper;
  26
  27 /**
  28  * @ingroup SpecialPage
  29  */
  30 class MediaStatisticsPage extends QueryPage {
  31         protected $totalCount = 0, $totalBytes = 0;
  32         /**
  33         * @var integer $totalPerType Combined file size of all files in a section
  34         */
  35         protected $totalPerType = 0;
  36         /**
  37         * @var integer $totalSize Combined file size of all files
  38         */
  39         protected $totalSize = 0;
  40
  41         function __construct( $name = 'MediaStatistics' ) {
  42                 parent::__construct( $name );
  43                 // Generally speaking there is only a small number of file types,
  44                 // so just show all of them.
  45                 $this->limit = 5000;
  46                 $this->shownavigation = false;
  47         }
  48
  49         public function isExpensive() {
  50                 return true;
  51         }
  52
  53         /**
  54          * Query to do.
  55          *
  56          * This abuses the query cache table by storing mime types as "titles".
  57          *
  58          * This will store entries like [[Media:BITMAP;image/jpeg;200;20000]]
  59          * where the form is Media type;mime type;count;bytes.
  60          *
  61          * This relies on the behaviour that when value is tied, the order things
  62          * come out of querycache table is the order they went in. Which is hacky.
  63          * However, other special pages like Special:Deadendpages and
  64          * Special:BrokenRedirects also rely on this.
  65          */
  66         public function getQueryInfo() {
  67                 $dbr = wfGetDB( DB_REPLICA );
  68                 $fakeTitle = $dbr->buildConcat( [
  69                         'img_media_type',
  70                         $dbr->addQuotes( ';' ),
  71                         'img_major_mime',
  72                         $dbr->addQuotes( '/' ),
  73                         'img_minor_mime',
  74                         $dbr->addQuotes( ';' ),
  75                         'COUNT(*)',
  76                         $dbr->addQuotes( ';' ),
  77                         'SUM( img_size )'
  78                 ] );
  79                 return [
  80                         'tables' => [ 'image' ],
  81                         'fields' => [
  82                                 'title' => $fakeTitle,
  83                                 'namespace' => NS_MEDIA, /* needs to be something */
  84                                 'value' => '1'
  85                         ],
  86                         'conds' => [
  87                                 // WMF has a random null row in the db
  88                                 'img_media_type IS NOT NULL'
  89                         ],
  90                         'options' => [
  91                                 'GROUP BY' => [
  92                                         'img_media_type',
  93                                         'img_major_mime',
  94                                         'img_minor_mime',
  95                                 ]
  96                         ]
  97                 ];
  98         }
  99
 100         /**
 101          * How to sort the results
 102          *
 103          * It's important that img_media_type come first, otherwise the
 104          * tables will be fragmented.
 105          * @return Array Fields to sort by
 106          */
 107         function getOrderFields() {
 108                 return [ 'img_media_type', 'count(*)', 'img_major_mime', 'img_minor_mime' ];
 109         }
 110
 111         /**
 112          * Output the results of the query.
 113          *
 114          * @param OutputPage $out
 115          * @param Skin $skin (deprecated presumably)
 116          * @param IDatabase $dbr
 117          * @param ResultWrapper $res Results from query
 118          * @param int $num Number of results
 119          * @param int $offset Paging offset (Should always be 0 in our case)
 120          */
 121         protected function outputResults( $out, $skin, $dbr, $res, $num, $offset ) {
 122                 $prevMediaType = null;
 123                 foreach ( $res as $row ) {
 124                         $mediaStats = $this->splitFakeTitle( $row->title );
 125                         if ( count( $mediaStats ) < 4 ) {
 126                                 continue;
 127                         }
 128                         list( $mediaType, $mime, $totalCount, $totalBytes ) = $mediaStats;
 129                         if ( $prevMediaType !== $mediaType ) {
 130                                 if ( $prevMediaType !== null ) {
 131                                         // We're not at beginning, so we have to
 132                                         // close the previous table.
 133                                         $this->outputTableEnd();
 134                                 }
 135                                 $this->outputMediaType( $mediaType );
 136                                 $this->totalPerType = 0;
 137                                 $this->outputTableStart( $mediaType );
 138                                 $prevMediaType = $mediaType;
 139                         }
 140                         $this->outputTableRow( $mime, intval( $totalCount ), intval( $totalBytes ) );
 141                 }
 142                 if ( $prevMediaType !== null ) {
 143                         $this->outputTableEnd();
 144                         // add total size of all files
 145                         $this->outputMediaType( 'total' );
 146                         $this->getOutput()->addWikiText(
 147                                 $this->msg( 'mediastatistics-allbytes' )
 148                                         ->numParams( $this->totalSize )
 149                                         ->sizeParams( $this->totalSize )
 150                                         ->text()
 151                         );
 152                 }
 153         }
 154
 155         /**
 156          * Output closing </table>
 157          */
 158         protected function outputTableEnd() {
 159                 $this->getOutput()->addHTML( Html::closeElement( 'table' ) );
 160                 $this->getOutput()->addWikiText(
 161                                 $this->msg( 'mediastatistics-bytespertype' )
 162                                         ->numParams( $this->totalPerType )
 163                                         ->sizeParams( $this->totalPerType )
 164                                         ->numParams( $this->makePercentPretty( $this->totalPerType / $this->totalBytes ) )
 165                                         ->text()
 166                 );
 167                 $this->totalSize += $this->totalPerType;
 168         }
 169
 170         /**
 171          * Output a row of the stats table
 172          *
 173          * @param string $mime mime type (e.g. image/jpeg)
 174          * @param int $count Number of images of this type
 175          * @param int $totalBytes Total space for images of this type
 176          */
 177         protected function outputTableRow( $mime, $count, $bytes ) {
 178                 $mimeSearch = SpecialPage::getTitleFor( 'MIMEsearch', $mime );
 179                 $linkRenderer = $this->getLinkRenderer();
 180                 $row = Html::rawElement(
 181                         'td',
 182                         [],
 183                         $linkRenderer->makeLink( $mimeSearch, $mime )
 184                 );
 185                 $row .= Html::element(
 186                         'td',
 187                         [],
 188                         $this->getExtensionList( $mime )
 189                 );
 190                 $row .= Html::rawElement(
 191                         'td',
 192                         // Make sure js sorts it in numeric order
 193                         [ 'data-sort-value' => $count ],
 194                         $this->msg( 'mediastatistics-nfiles' )
 195                                 ->numParams( $count )
 196                                 /** @todo Check to be sure this really should have number formatting */
 197                                 ->numParams( $this->makePercentPretty( $count / $this->totalCount ) )
 198                                 ->parse()
 199                 );
 200                 $row .= Html::rawElement(
 201                         'td',
 202                         // Make sure js sorts it in numeric order
 203                         [ 'data-sort-value' =>  $bytes ],
 204                         $this->msg( 'mediastatistics-nbytes' )
 205                                 ->numParams( $bytes )
 206                                 ->sizeParams( $bytes )
 207                                 /** @todo Check to be sure this really should have number formatting */
 208                                 ->numParams( $this->makePercentPretty( $bytes / $this->totalBytes ) )
 209                                 ->parse()
 210                 );
 211                 $this->totalPerType += $bytes;
 212                 $this->getOutput()->addHTML( Html::rawElement( 'tr', [], $row ) );
 213         }
 214
 215         /**
 216          * @param float $decimal A decimal percentage (ie for 12.3%, this would be 0.123)
 217          * @return String The percentage formatted so that 3 significant digits are shown.
 218          */
 219         protected function makePercentPretty( $decimal ) {
 220                 $decimal *= 100;
 221                 // Always show three useful digits
 222                 if ( $decimal == 0 ) {
 223                         return '0';
 224                 }
 225                 if ( $decimal >= 100 ) {
 226                         return '100';
 227                 }
 228                 $percent = sprintf( "%." . max( 0, 2 - floor( log10( $decimal ) ) ) . "f", $decimal );
 229                 // Then remove any trailing 0's
 230                 return preg_replace( '/\.?0*$/', '', $percent );
 231         }
 232
 233         /**
 234          * Given a mime type, return a comma separated list of allowed extensions.
 235          *
 236          * @param string $mime mime type
 237          * @return string Comma separated list of allowed extensions (e.g. ".ogg, .oga")
 238          */
 239         private function getExtensionList( $mime ) {
 240                 $exts = MimeMagic::singleton()->getExtensionsForType( $mime );
 241                 if ( $exts === null ) {
 242                         return '';
 243                 }
 244                 $extArray = explode( ' ', $exts );
 245                 $extArray = array_unique( $extArray );
 246                 foreach ( $extArray as &$ext ) {
 247                         $ext = '.' . $ext;
 248                 }
 249
 250                 return $this->getLanguage()->commaList( $extArray );
 251         }
 252
 253         /**
 254          * Output the start of the table
 255          *
 256          * Including opening <table>, and first <tr> with column headers.
 257          */
 258         protected function outputTableStart( $mediaType ) {
 259                 $this->getOutput()->addHTML(
 260                         Html::openElement(
 261                                 'table',
 262                                 [ 'class' => [
 263                                         'mw-mediastats-table',
 264                                         'mw-mediastats-table-' . strtolower( $mediaType ),
 265                                         'sortable',
 266                                         'wikitable'
 267                                 ] ]
 268                         )
 269                 );
 270                 $this->getOutput()->addHTML( $this->getTableHeaderRow() );
 271         }
 272
 273         /**
 274          * Get (not output) the header row for the table
 275          *
 276          * @return String the header row of the able
 277          */
 278         protected function getTableHeaderRow() {
 279                 $headers = [ 'mimetype', 'extensions', 'count', 'totalbytes' ];
 280                 $ths = '';
 281                 foreach ( $headers as $header ) {
 282                         $ths .= Html::rawElement(
 283                                 'th',
 284                                 [],
 285                                 // for grep:
 286                                 // mediastatistics-table-mimetype, mediastatistics-table-extensions
 287                                 // tatistics-table-count, mediastatistics-table-totalbytes
 288                                 $this->msg( 'mediastatistics-table-' . $header )->parse()
 289                         );
 290                 }
 291                 return Html::rawElement( 'tr', [], $ths );
 292         }
 293
 294         /**
 295          * Output a header for a new media type section
 296          *
 297          * @param string $mediaType A media type (e.g. from the MEDIATYPE_xxx constants)
 298          */
 299         protected function outputMediaType( $mediaType ) {
 300                 $this->getOutput()->addHTML(
 301                         Html::element(
 302                                 'h2',
 303                                 [ 'class' => [
 304                                         'mw-mediastats-mediatype',
 305                                         'mw-mediastats-mediatype-' . strtolower( $mediaType )
 306                                 ] ],
 307                                 // for grep
 308                                 // mediastatistics-header-unknown, mediastatistics-header-bitmap,
 309                                 // mediastatistics-header-drawing, mediastatistics-header-audio,
 310                                 // mediastatistics-header-video, mediastatistics-header-multimedia,
 311                                 // mediastatistics-header-office, mediastatistics-header-text,
 312                                 // mediastatistics-header-executable, mediastatistics-header-archive,
 313                                 $this->msg( 'mediastatistics-header-' . strtolower( $mediaType ) )->text()
 314                         )
 315                 );
 316                 /** @todo Possibly could add a message here explaining what the different types are.
 317                  *    not sure if it is needed though.
 318                  */
 319         }
 320
 321         /**
 322          * parse the fake title format that this special page abuses querycache with.
 323          *
 324          * @param string $fakeTitle A string formatted as <media type>;<mime type>;<count>;<bytes>
 325          * @return array The constituant parts of $fakeTitle
 326          */
 327         private function splitFakeTitle( $fakeTitle ) {
 328                 return explode( ';', $fakeTitle, 4 );
 329         }
 330
 331         /**
 332          * What group to put the page in
 333          * @return string
 334          */
 335         protected function getGroupName() {
 336                 return 'media';
 337         }
 338
 339         /**
 340          * This method isn't used, since we override outputResults, but
 341          * we need to implement since abstract in parent class.
 342          *
 343          * @param Skin $skin
 344          * @param stdClass $result Result row
 345          * @return bool|string|void
 346          * @throws MWException
 347          */
 348         public function formatResult( $skin, $result ) {
 349                 throw new MWException( "unimplemented" );
 350         }
 351
 352         /**
 353          * Initialize total values so we can figure out percentages later.
 354          *
 355          * @param IDatabase $dbr
 356          * @param ResultWrapper $res
 357          */
 358         public function preprocessResults( $dbr, $res ) {
 359                 $this->executeLBFromResultWrapper( $res );
 360                 $this->totalCount = $this->totalBytes = 0;
 361                 foreach ( $res as $row ) {
 362                         $mediaStats = $this->splitFakeTitle( $row->title );
 363                         $this->totalCount += isset( $mediaStats[2] ) ? $mediaStats[2] : 0;
 364                         $this->totalBytes += isset( $mediaStats[3] ) ? $mediaStats[3] : 0;
 365                 }
 366                 $res->seek( 0 );
 367         }
 368 }