includes/api/ApiFormatBase.php

   1 <?php
   2 /**
   3  *
   4  *
   5  * Created on Sep 19, 2006
   6  *
   7  * Copyright © 2006 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 2 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License along
  20  * with this program; if not, write to the Free Software Foundation, Inc.,
  21  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22  * http://www.gnu.org/copyleft/gpl.html
  23  *
  24  * @file
  25  */
  26
  27 /**
  28  * This is the abstract base class for API formatters.
  29  *
  30  * @ingroup API
  31  */
  32 abstract class ApiFormatBase extends ApiBase {
  33         private $mIsHtml, $mFormat, $mUnescapeAmps, $mHelp, $mCleared;
  34         private $mBufferResult = false, $mBuffer, $mDisabled = false;
  35
  36         /**
  37          * If $format ends with 'fm', pretty-print the output in HTML.
  38          * @param ApiMain $main
  39          * @param string $format Format name
  40          */
  41         public function __construct( ApiMain $main, $format ) {
  42                 parent::__construct( $main, $format );
  43
  44                 $this->mIsHtml = ( substr( $format, -2, 2 ) === 'fm' ); // ends with 'fm'
  45                 if ( $this->mIsHtml ) {
  46                         $this->mFormat = substr( $format, 0, -2 ); // remove ending 'fm'
  47                 } else {
  48                         $this->mFormat = $format;
  49                 }
  50                 $this->mFormat = strtoupper( $this->mFormat );
  51                 $this->mCleared = false;
  52         }
  53
  54         /**
  55          * Overriding class returns the MIME type that should be sent to the client.
  56          * This method is not called if getIsHtml() returns true.
  57          * @return string
  58          */
  59         abstract public function getMimeType();
  60
  61         /**
  62          * Whether this formatter needs raw data such as _element tags
  63          * @return bool
  64          */
  65         public function getNeedsRawData() {
  66                 return false;
  67         }
  68
  69         /**
  70          * Get the internal format name
  71          * @return string
  72          */
  73         public function getFormat() {
  74                 return $this->mFormat;
  75         }
  76
  77         /**
  78          * Specify whether or not sequences like &amp;quot; should be unescaped
  79          * to &quot; . This should only be set to true for the help message
  80          * when rendered in the default (xmlfm) format. This is a temporary
  81          * special-case fix that should be removed once the help has been
  82          * reworked to use a fully HTML interface.
  83          *
  84          * @deprecated since 1.25
  85          * @param bool $b Whether or not ampersands should be escaped.
  86          */
  87         public function setUnescapeAmps( $b ) {
  88                 wfDeprecated( __METHOD__, '1.25' );
  89                 $this->mUnescapeAmps = $b;
  90         }
  91
  92         /**
  93          * Returns true when the HTML pretty-printer should be used.
  94          * The default implementation assumes that formats ending with 'fm'
  95          * should be formatted in HTML.
  96          * @return bool
  97          */
  98         public function getIsHtml() {
  99                 return $this->mIsHtml;
 100         }
 101
 102         /**
 103          * Whether this formatter can format the help message in a nice way.
 104          * By default, this returns the same as getIsHtml().
 105          * When action=help is set explicitly, the help will always be shown
 106          * @deprecated since 1.25
 107          * @return bool
 108          */
 109         public function getWantsHelp() {
 110                 wfDeprecated( __METHOD__, '1.25' );
 111                 return $this->getIsHtml();
 112         }
 113
 114         /**
 115          * Disable the formatter completely. This causes calls to initPrinter(),
 116          * printText() and closePrinter() to be ignored.
 117          */
 118         public function disable() {
 119                 $this->mDisabled = true;
 120         }
 121
 122         public function isDisabled() {
 123                 return $this->mDisabled;
 124         }
 125
 126         /**
 127          * Whether this formatter can handle printing API errors. If this returns
 128          * false, then on API errors the default printer will be instantiated.
 129          * @since 1.23
 130          * @return bool
 131          */
 132         public function canPrintErrors() {
 133                 return true;
 134         }
 135
 136         /**
 137          * Initialize the printer function and prepare the output headers, etc.
 138          * This method must be the first outputting method during execution.
 139          * A human-targeted notice about available formats is printed for the HTML-based output,
 140          * except for help screens (caused by either an error in the API parameters,
 141          * the calling of action=help, or requesting the root script api.php).
 142          * @param bool $unused Always false since 1.25
 143          */
 144         function initPrinter( $unused ) {
 145                 if ( $this->mDisabled ) {
 146                         return;
 147                 }
 148                 $isHtml = $this->getIsHtml();
 149                 $mime = $isHtml ? 'text/html' : $this->getMimeType();
 150                 $script = wfScript( 'api' );
 151
 152                 // Some printers (ex. Feed) do their own header settings,
 153                 // in which case $mime will be set to null
 154                 if ( is_null( $mime ) ) {
 155                         return; // skip any initialization
 156                 }
 157
 158                 $this->getMain()->getRequest()->response()->header( "Content-Type: $mime; charset=utf-8" );
 159
 160                 //Set X-Frame-Options API results (bug 39180)
 161                 $apiFrameOptions = $this->getConfig()->get( 'ApiFrameOptions' );
 162                 if ( $apiFrameOptions ) {
 163                         $this->getMain()->getRequest()->response()->header( "X-Frame-Options: $apiFrameOptions" );
 164                 }
 165
 166                 if ( $isHtml ) {
 167 ?>
 168 <!DOCTYPE HTML>
 169 <html>
 170 <head>
 171 <?php
 172                         if ( $this->mUnescapeAmps ) {
 173 ?>      <title>MediaWiki API</title>
 174 <?php
 175                         } else {
 176 ?>      <title>MediaWiki API Result</title>
 177 <?php
 178                         }
 179 // @codingStandardsIgnoreStart Exclude long line from CodeSniffer checks
 180 ?>
 181 </head>
 182 <body>
 183 <br />
 184 <small>
 185 You are looking at the HTML representation of the <?php echo $this->mFormat; ?> format.<br />
 186 HTML is good for debugging, but is unsuitable for application use.<br />
 187 Specify the format parameter to change the output format.<br />
 188 To see the non HTML representation of the <?php echo $this->mFormat; ?> format, set format=<?php echo strtolower( $this->mFormat ); ?>.<br />
 189 See the <a href='https://www.mediawiki.org/wiki/API'>complete documentation</a>, or
 190 <a href='<?php echo $script; ?>'>API help</a> for more information.
 191 </small>
 192 <pre style='white-space: pre-wrap;'>
 193 <?php
 194 // @codingStandardsIgnoreEnd
 195                 }
 196         }
 197
 198         /**
 199          * Finish printing. Closes HTML tags.
 200          */
 201         public function closePrinter() {
 202                 if ( $this->mDisabled ) {
 203                         return;
 204                 }
 205                 if ( $this->getIsHtml() ) {
 206 ?>
 207
 208 </pre>
 209 </body>
 210 </html>
 211 <?php
 212                 }
 213         }
 214
 215         /**
 216          * The main format printing function. Call it to output the result
 217          * string to the user. This function will automatically output HTML
 218          * when format name ends in 'fm'.
 219          * @param string $text
 220          */
 221         public function printText( $text ) {
 222                 if ( $this->mDisabled ) {
 223                         return;
 224                 }
 225                 if ( $this->mBufferResult ) {
 226                         $this->mBuffer = $text;
 227                 } elseif ( $this->getIsHtml() ) {
 228                         echo $this->formatHTML( $text );
 229                 } else {
 230                         // For non-HTML output, clear all errors that might have been
 231                         // displayed if display_errors=On
 232                         // Do this only once, of course
 233                         if ( !$this->mCleared ) {
 234                                 ob_clean();
 235                                 $this->mCleared = true;
 236                         }
 237                         echo $text;
 238                 }
 239         }
 240
 241         /**
 242          * Get the contents of the buffer.
 243          * @return string
 244          */
 245         public function getBuffer() {
 246                 return $this->mBuffer;
 247         }
 248
 249         /**
 250          * Set the flag to buffer the result instead of printing it.
 251          * @param bool $value
 252          */
 253         public function setBufferResult( $value ) {
 254                 $this->mBufferResult = $value;
 255         }
 256
 257         /**
 258          * Sets whether the pretty-printer should format *bold*
 259          * @deprecated since 1.25
 260          * @param bool $help
 261          */
 262         public function setHelp( $help = true ) {
 263                 wfDeprecated( __METHOD__, '1.25' );
 264                 $this->mHelp = $help;
 265         }
 266
 267         /**
 268          * Pretty-print various elements in HTML format, such as xml tags and
 269          * URLs. This method also escapes characters like <
 270          * @param string $text
 271          * @return string
 272          */
 273         protected function formatHTML( $text ) {
 274                 // Escape everything first for full coverage
 275                 $text = htmlspecialchars( $text );
 276
 277                 if ( $this->mFormat === 'XML' || $this->mFormat === 'WDDX' ) {
 278                         // encode all comments or tags as safe blue strings
 279                         $text = str_replace( '&lt;', '<span style="color:blue;">&lt;', $text );
 280                         $text = str_replace( '&gt;', '&gt;</span>', $text );
 281                 }
 282
 283                 // identify requests to api.php
 284                 $text = preg_replace( '#^(\s*)(api\.php\?[^ <\n\t]+)$#m', '\1<a href="\2">\2</a>', $text );
 285                 if ( $this->mHelp ) {
 286                         // make lines inside * bold
 287                         $text = preg_replace( '#^(\s*)(\*[^<>\n]+\*)(\s*)$#m', '$1<b>$2</b>$3', $text );
 288                 }
 289
 290                 // Armor links (bug 61362)
 291                 $masked = array();
 292                 $text = preg_replace_callback( '#<a .*?</a>#', function ( $matches ) use ( &$masked ) {
 293                         $sha = sha1( $matches[0] );
 294                         $masked[$sha] = $matches[0];
 295                         return "<$sha>";
 296                 }, $text );
 297
 298                 // identify URLs
 299                 $protos = wfUrlProtocolsWithoutProtRel();
 300                 // This regex hacks around bug 13218 (&quot; included in the URL)
 301                 $text = preg_replace(
 302                         "#(((?i)$protos).*?)(&quot;)?([ \\'\"<>\n]|&lt;|&gt;|&quot;)#",
 303                         '<a href="\\1">\\1</a>\\3\\4',
 304                         $text
 305                 );
 306
 307                 // Unarmor links
 308                 $text = preg_replace_callback( '#<([0-9a-f]{40})>#', function ( $matches ) use ( &$masked ) {
 309                         $sha = $matches[1];
 310                         return isset( $masked[$sha] ) ? $masked[$sha] : $matches[0];
 311                 }, $text );
 312
 313                 /**
 314                  * Temporary fix for bad links in help messages. As a special case,
 315                  * XML-escaped metachars are de-escaped one level in the help message
 316                  * for legibility. Should be removed once we have completed a fully-HTML
 317                  * version of the help message.
 318                  */
 319                 if ( $this->mUnescapeAmps ) {
 320                         $text = preg_replace( '/&amp;(amp|quot|lt|gt);/', '&\1;', $text );
 321                 }
 322
 323                 return $text;
 324         }
 325
 326         public function getExamples() {
 327                 return array(
 328                         'api.php?action=query&meta=siteinfo&siprop=namespaces&format=' . $this->getModuleName()
 329                                 => "Format the query result in the {$this->getModuleName()} format",
 330                 );
 331         }
 332
 333         public function getHelpUrls() {
 334                 return 'https://www.mediawiki.org/wiki/API:Data_formats';
 335         }
 336
 337         public function getDescription() {
 338                 return $this->getIsHtml() ? ' (pretty-print in HTML)' : '';
 339         }
 340
 341         /**
 342          * To avoid code duplication with the deprecation of dbg, dump, txt, wddx,
 343          * and yaml, this method is added to do the necessary work. It should be
 344          * removed when those deprecated formats are removed.
 345          */
 346         protected function markDeprecated() {
 347                 $fm = $this->getIsHtml() ? 'fm' : '';
 348                 $name = $this->getModuleName();
 349                 $this->logFeatureUsage( "format=$name" );
 350                 $this->setWarning( "format=$name has been deprecated. Please use format=json$fm instead." );
 351         }
 352 }