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_SLAVE );
  66                 $fakeTitle = $dbr->buildConcat( array(
  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 array(
  78                         'tables' => array( 'image' ),
  79                         'fields' => array(
  80                                 'title' => $fakeTitle,
  81                                 'namespace' => NS_MEDIA, /* needs to be something */
  82                                 'value' => '1'
  83                         ),
  84                         'conds' => array(
  85                                 // WMF has a random null row in the db
  86                                 'img_media_type IS NOT NULL'
  87                         ),
  88                         'options' => array(
  89                                 'GROUP BY' => array(
  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 array( 'img_media_type', 'count(*)', 'img_major_mime', 'img_minor_mime' );
 107         }
 108
 109         /**
 110          * Output the results of the query.
 111          *
 112          * @param $out OutputPage
 113          * @param $skin Skin (deprecated presumably)
 114          * @param $dbr IDatabase
 115          * @param $res ResultWrapper Results from query
 116          * @param $num integer Number of results
 117          * @param $offset integer 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                                         ->text()
 148                         );
 149                 }
 150         }
 151
 152         /**
 153          * Output closing </table>
 154          */
 155         protected function outputTableEnd() {
 156                 $this->getOutput()->addHtml( Html::closeElement( 'table' ) );
 157                 $this->getOutput()->addWikiText(
 158                                 $this->msg( 'mediastatistics-bytespertype' )
 159                                         ->numParams( $this->totalPerType )
 160                                         ->text()
 161                 );
 162                 $this->totalSize += $this->totalPerType;
 163         }
 164
 165         /**
 166          * Output a row of the stats table
 167          *
 168          * @param $mime String mime type (e.g. image/jpeg)
 169          * @param $count integer Number of images of this type
 170          * @param $totalBytes integer Total space for images of this type
 171          */
 172         protected function outputTableRow( $mime, $count, $bytes ) {
 173                 $mimeSearch = SpecialPage::getTitleFor( 'MIMEsearch', $mime );
 174                 $row = Html::rawElement(
 175                         'td',
 176                         array(),
 177                         Linker::link( $mimeSearch, htmlspecialchars( $mime ) )
 178                 );
 179                 $row .= Html::element(
 180                         'td',
 181                         array(),
 182                         $this->getExtensionList( $mime )
 183                 );
 184                 $row .= Html::rawElement(
 185                         'td',
 186                         // Make sure js sorts it in numeric order
 187                         array( 'data-sort-value' => $count ),
 188                         $this->msg( 'mediastatistics-nfiles' )
 189                                 ->numParams( $count )
 190                                 /** @todo Check to be sure this really should have number formatting */
 191                                 ->numParams( $this->makePercentPretty( $count / $this->totalCount ) )
 192                                 ->parse()
 193                 );
 194                 $row .= Html::rawElement(
 195                         'td',
 196                         // Make sure js sorts it in numeric order
 197                         array( 'data-sort-value' =>  $bytes ),
 198                         $this->msg( 'mediastatistics-nbytes' )
 199                                 ->numParams( $bytes )
 200                                 ->sizeParams( $bytes )
 201                                 /** @todo Check to be sure this really should have number formatting */
 202                                 ->numParams( $this->makePercentPretty( $bytes / $this->totalBytes ) )
 203                                 ->parse()
 204                 );
 205                 $this->totalPerType += $bytes;
 206                 $this->getOutput()->addHTML( Html::rawElement( 'tr', array(), $row ) );
 207         }
 208
 209         /**
 210          * @param float $decimal A decimal percentage (ie for 12.3%, this would be 0.123)
 211          * @return String The percentage formatted so that 3 significant digits are shown.
 212          */
 213         protected function makePercentPretty( $decimal ) {
 214                 $decimal *= 100;
 215                 // Always show three useful digits
 216                 if ( $decimal == 0 ) {
 217                         return '0';
 218                 }
 219                 if ( $decimal >= 100 ) {
 220                         return '100';
 221                 }
 222                 $percent = sprintf( "%." . max( 0, 2 - floor( log10( $decimal ) ) ) . "f", $decimal );
 223                 // Then remove any trailing 0's
 224                 return preg_replace( '/\.?0*$/', '', $percent );
 225         }
 226
 227         /**
 228          * Given a mime type, return a comma separated list of allowed extensions.
 229          *
 230          * @param $mime String mime type
 231          * @return String Comma separated list of allowed extensions (e.g. ".ogg, .oga")
 232          */
 233         private function getExtensionList( $mime ) {
 234                 $exts = MimeMagic::singleton()->getExtensionsForType( $mime );
 235                 if ( $exts === null ) {
 236                         return '';
 237                 }
 238                 $extArray = explode( ' ', $exts );
 239                 $extArray = array_unique( $extArray );
 240                 foreach ( $extArray as &$ext ) {
 241                         $ext = '.' . $ext;
 242                 }
 243
 244                 return $this->getLanguage()->commaList( $extArray );
 245         }
 246
 247         /**
 248          * Output the start of the table
 249          *
 250          * Including opening <table>, and first <tr> with column headers.
 251          */
 252         protected function outputTableStart( $mediaType ) {
 253                 $this->getOutput()->addHTML(
 254                         Html::openElement(
 255                                 'table',
 256                                 array( 'class' => array(
 257                                         'mw-mediastats-table',
 258                                         'mw-mediastats-table-' . strtolower( $mediaType ),
 259                                         'sortable',
 260                                         'wikitable'
 261                                 ) )
 262                         )
 263                 );
 264                 $this->getOutput()->addHTML( $this->getTableHeaderRow() );
 265         }
 266
 267         /**
 268          * Get (not output) the header row for the table
 269          *
 270          * @return String the header row of the able
 271          */
 272         protected function getTableHeaderRow() {
 273                 $headers = array( 'mimetype', 'extensions', 'count', 'totalbytes' );
 274                 $ths = '';
 275                 foreach ( $headers as $header ) {
 276                         $ths .= Html::rawElement(
 277                                 'th',
 278                                 array(),
 279                                 // for grep:
 280                                 // mediastatistics-table-mimetype, mediastatistics-table-extensions
 281                                 // tatistics-table-count, mediastatistics-table-totalbytes
 282                                 $this->msg( 'mediastatistics-table-' . $header )->parse()
 283                         );
 284                 }
 285                 return Html::rawElement( 'tr', array(), $ths );
 286         }
 287
 288         /**
 289          * Output a header for a new media type section
 290          *
 291          * @param $mediaType string A media type (e.g. from the MEDIATYPE_xxx constants)
 292          */
 293         protected function outputMediaType( $mediaType ) {
 294                 $this->getOutput()->addHTML(
 295                         Html::element(
 296                                 'h2',
 297                                 array( 'class' => array(
 298                                         'mw-mediastats-mediatype',
 299                                         'mw-mediastats-mediatype-' . strtolower( $mediaType )
 300                                 ) ),
 301                                 // for grep
 302                                 // mediastatistics-header-unknown, mediastatistics-header-bitmap,
 303                                 // mediastatistics-header-drawing, mediastatistics-header-audio,
 304                                 // mediastatistics-header-video, mediastatistics-header-multimedia,
 305                                 // mediastatistics-header-office, mediastatistics-header-text,
 306                                 // mediastatistics-header-executable, mediastatistics-header-archive,
 307                                 $this->msg( 'mediastatistics-header-' . strtolower( $mediaType ) )->text()
 308                         )
 309                 );
 310                 /** @todo Possibly could add a message here explaining what the different types are.
 311                  *    not sure if it is needed though.
 312                  */
 313         }
 314
 315         /**
 316          * parse the fake title format that this special page abuses querycache with.
 317          *
 318          * @param $fakeTitle String A string formatted as <media type>;<mime type>;<count>;<bytes>
 319          * @return Array The constituant parts of $fakeTitle
 320          */
 321         private function splitFakeTitle( $fakeTitle ) {
 322                 return explode( ';', $fakeTitle, 4 );
 323         }
 324
 325         /**
 326          * What group to put the page in
 327          * @return string
 328          */
 329         protected function getGroupName() {
 330                 return 'media';
 331         }
 332
 333         /**
 334          * This method isn't used, since we override outputResults, but
 335          * we need to implement since abstract in parent class.
 336          *
 337          * @param $skin Skin
 338          * @param $result stdObject Result row
 339          * @return bool|string|void
 340          * @throws MWException
 341          */
 342         public function formatResult( $skin, $result ) {
 343                 throw new MWException( "unimplemented" );
 344         }
 345
 346         /**
 347          * Initialize total values so we can figure out percentages later.
 348          *
 349          * @param $dbr IDatabase
 350          * @param $res ResultWrapper
 351          */
 352         public function preprocessResults( $dbr, $res ) {
 353                 $this->totalCount = $this->totalBytes = 0;
 354                 foreach ( $res as $row ) {
 355                         $mediaStats = $this->splitFakeTitle( $row->title );
 356                         $this->totalCount += isset( $mediaStats[2] ) ? $mediaStats[2] : 0;
 357                         $this->totalBytes += isset( $mediaStats[3] ) ? $mediaStats[3] : 0;
 358                 }
 359                 $res->seek( 0 );
 360         }
 361 }