[Sommaire] [Imprimer]

   1  <?php
   2  
   3  /*************************************************************************
   4   *                                                                       *
   5   * class.html2text.inc                                                   *
   6   *                                                                       *
   7   *************************************************************************
   8   *                                                                       *
   9   * Converts HTML to formatted plain text                                 *
  10   *                                                                       *
  11   * Copyright (c) 2003 Jon Abernathy <jon@chuggnutt.com>                  *
  12   * All rights reserved.                                                  *
  13   *                                                                       *
  14   * This script is free software; you can redistribute it and/or modify   *
  15   * it under the terms of the GNU General Public License as published by  *
  16   * the Free Software Foundation; either version 2 of the License, or     *
  17   * (at your option) any later version.                                   *
  18   *                                                                       *
  19   * The GNU General Public License can be found at                        *
  20   * http://www.gnu.org/copyleft/gpl.html.                                 *
  21   *                                                                       *
  22   * This script is distributed in the hope that it will be useful,        *
  23   * but WITHOUT ANY WARRANTY; without even the implied warranty of        *
  24   * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the          *
  25   * GNU General Public License for more details.                          *
  26   *                                                                       *
  27   * Author(s): Jon Abernathy <jon@chuggnutt.com>                          *
  28   *                                                                       *
  29   * Last modified: 11/07/03                                               *
  30   *                                                                       *
  31   *************************************************************************/
  32  
  33  
  34  /**
  35   *  Takes HTML and converts it to formatted, plain text.
  36   *
  37   *  Thanks to Alexander Krug (http://www.krugar.de/) to pointing out and
  38   *  correcting an error in the regexp search array. Fixed 7/30/03.
  39   *
  40   *  Updated set_html() function's file reading mechanism, 9/25/03.
  41   *
  42   *  Thanks to Joss Sanglier (http://www.dancingbear.co.uk/) for adding
  43   *  several more HTML entity codes to the $search and $replace arrays.
  44   *  Updated 11/7/03.
  45   *
  46   *  @author Jon Abernathy <jon@chuggnutt.com>
  47   *  @version 0.4
  48   *  @since PHP 4.0.2
  49   */
  50  class html2text
  51  {
  52  
  53      /**
  54       *  Contains the HTML content to convert.
  55       *
  56       *  @var string $html
  57       *  @access public
  58       */
  59      var $html;
  60  
  61      /**
  62       *  Contains the converted, formatted text.
  63       *
  64       *  @var string $text
  65       *  @access public
  66       */
  67      var $text;
  68  
  69      /**
  70       *  Maximum width of the formatted text, in columns.
  71       *
  72       *  @var integer $width
  73       *  @access public
  74       */
  75      var $width = 70;
  76  
  77      /**
  78       *  List of preg* regular expression patterns to search for,
  79       *  used in conjunction with $replace.
  80       *
  81       *  @var array $search
  82       *  @access public
  83       *  @see $replace
  84       */
  85      var $search = array(
  86          "/\r/",                                  // Non-legal carriage return
  87          "/[\n\t]+/",                             // Newlines and tabs
  88          '/<script[^>]*>.*?<\/script>/i',         // <script>s -- which strip_tags supposedly has problems with
  89          //'/<!-- .* -->/',                         // Comments -- which strip_tags might have problem a with
  90          '/<h[123][^>]*>(.+?)<\/h[123]>/ie',      // H1 - H3
  91          '/<h[456][^>]*>(.+?)<\/h[456]>/ie',      // H4 - H6
  92          '/<p[^>]*>/i',                           // <P>
  93          '/<br[^>]*>/i',                          // <br>
  94          '/<b[^>]*>(.+?)<\/b>/ie',                // <b>
  95          '/<i[^>]*>(.+?)<\/i>/i',                 // <i>
  96          '/(<ul[^>]*>|<\/ul>)/i',                 // <ul> and </ul>
  97          '/<li[^>]*>/i',                          // <li>
  98          '/<\/li[^>]*>/i',                             // </li>
  99          '/<a href="([^"]+)"[^>]*>(.+?)<\/a>/ie', // <a href="">
 100          '/<hr[^>]*>/i',                          // <hr>
 101          '/(<table[^>]*>|<\/table>)/i',           // <table> and </table>
 102          '/(<tr[^>]*>|<\/tr>)/i',                 // <tr> and </tr>
 103          '/<td[^>]*>(.+?)<\/td>/i',               // <td> and </td>
 104          '/&nbsp;/i',
 105          '/&quot;/i',
 106          '/&gt;/i',
 107          '/&lt;/i',
 108          '/&amp;/i',
 109          '/&copy;/i',
 110          '/&trade;/i',
 111          '/&#8220;/',
 112          '/&#8221;/',
 113          '/&#8211;/',
 114          '/&#8217;/',
 115          '/&#38;/',
 116          '/&#169;/',
 117          '/&#8482;/',
 118          '/&#151;/',
 119          '/&#147;/',
 120          '/&#148;/',
 121          '/&#149;/',
 122          '/&reg;/i'
 123      );
 124  
 125      /**
 126       *  List of pattern replacements corresponding to patterns searched.
 127       *
 128       *  @var array $replace
 129       *  @access public
 130       *  @see $search
 131       */
 132      var $replace = array(
 133          '',                                     // Non-legal carriage return
 134          ' ',                                    // Newlines and tabs
 135          '',                                     // <script>s -- which strip_tags supposedly has problems with
 136          //'',                                     // Comments -- which strip_tags might have problem a with
 137          "strtoupper(\"\n\n\\1\n\n\")",          // H1 - H3
 138          "ucwords(\"\n\n\\1\n\n\")",             // H4 - H6
 139          "\n\n\t",                               // <P>
 140          "\n",                                   // <br>
 141          'strtoupper("\\1")',                    // <b>
 142          '_\\1_',                                // <i>
 143          "\n\n",                                 // <ul> and </ul>
 144          "\t*",                                  // <li>
 145          "\n",                                  // </li>
 146          '$this->_build_link_list($link_count++, "\\1", "\\2")',
 147                                                  // <a href="">
 148          "\n-------------------------\n",        // <hr>
 149          "\n\n",                                 // <table> and </table>
 150          "\n",                                   // <tr> and </tr>
 151          "\t\t\\1\n",                            // <td> and </td>
 152          ' ',
 153          '"',
 154          '>',
 155          '<',
 156          '&',
 157          '(c)',
 158          '(tm)',
 159          '"',
 160          '"',
 161          '-',
 162          "'",
 163          '&',
 164          '(c)',
 165          '(tm)',
 166          '--',
 167          '"',
 168          '"',
 169          '*',
 170          '(R)'
 171      );
 172  
 173      /**
 174       *  Indicates whether content in the $html variable has been converted yet.
 175       *
 176       *  @var boolean $converted
 177       *  @access private
 178       *  @see $html, $text
 179       */
 180      var $_converted = false;
 181  
 182      /**
 183       *  Contains URL addresses from links to be rendered in plain text.
 184       *
 185       *  @var string $link_list
 186       *  @access private
 187       *  @see _build_link_list()
 188       */
 189      var $_link_list;
 190  
 191      /**
 192       *  Constructor.
 193       *
 194       *  If the HTML source string (or file) is supplied, the class
 195       *  will instantiate with that source propagated, all that has
 196       *  to be done it to call get_text().
 197       *
 198       *  @param string $source HTML content
 199       *  @param boolean $from_file Indicates $source is a file to pull content from
 200       *  @access public
 201       *  @return void
 202       */
 203      function html2text( $source = '', $from_file = false )
 204      {
 205          if ( !empty($source) ) {
 206              $this->set_html($source, $from_file);
 207          }
 208      }
 209  
 210      /**
 211       *  Loads source HTML into memory, either from $source string or a file.
 212       *
 213       *  @param string $source HTML content
 214       *  @param boolean $from_file Indicates $source is a file to pull content from
 215       *  @access public
 216       *  @return void
 217       */
 218      function set_html( $source, $from_file = false )
 219      {
 220          $this->html = $source;
 221  
 222          if ( $from_file && file_exists($source) ) {
 223              $fp = fopen($source, 'r');
 224              $this->html = fread($fp, filesize($source));
 225              fclose($fp);
 226          }
 227  
 228          $this->_converted = false;
 229      }
 230  
 231      /**
 232       *  Returns the text, converted from HTML.
 233       *
 234       *  @access public
 235       *  @return string
 236       */
 237      function get_text()
 238      {
 239          if ( !$this->_converted ) {
 240              $this->_convert();
 241          }
 242  
 243          return $this->text;
 244      }
 245  
 246      /**
 247       *  Prints the text, converted from HTML.
 248       *
 249       *  @access public
 250       *  @return void
 251       */
 252      function print_text()
 253      {
 254          print $this->get_text();
 255      }
 256  
 257      /**
 258       *  Alias to print_text(), operates identically.
 259       *
 260       *  @access public
 261       *  @return void
 262       *  @see print_text()
 263       */
 264      function p()
 265      {
 266          print $this->get_text();
 267      }
 268  
 269      /**
 270       *  Workhorse function that does actual conversion.
 271       *
 272       *  First performs custom tag replacement specified by $search and
 273       *  $replace arrays. Then strips any remaining HTML tags, reduces whitespace
 274       *  and newlines to a readable format, and word wraps the text to
 275       *  $width characters.
 276       *
 277       *  @access private
 278       *  @return void
 279       */
 280      function _convert()
 281      {
 282          // Variables used for building the link list
 283          $link_count = 1;
 284          $this->_link_list = '';
 285  
 286          $text = trim(stripslashes($this->html));
 287  
 288          // Run our defined search-and-replace
 289          $text = preg_replace($this->search, $this->replace, trim($text));
 290  
 291          // Strip any other HTML tags
 292          $text = strip_tags($text);
 293  
 294          // Bring down number of empty lines to 2 max
 295          $text = preg_replace("/\n[[:space:]]+\n/", "\n\n", $text);
 296          $text = preg_replace("/[\n]{3,}/", "\n\n", $text);
 297  
 298          // Add link list
 299          if ( !empty($this->_link_list) ) {
 300              $text .= "\n\n".__('Links:')."\n------\n" . $this->_link_list;
 301          }
 302  
 303          // Wrap the text to a readable format
 304          // for PHP versions >= 4.0.2. Default width is 75
 305          $text = wordwrap($text, $this->width);
 306  
 307          $this->text = $text;
 308  
 309          $this->_converted = true;
 310      }
 311  
 312      /**
 313       *  Helper function called by preg_replace() on link replacement.
 314       *
 315       *  Maintains an internal list of links to be displayed at the end of the
 316       *  text, with numeric indices to the original point in the text they
 317       *  appeared.
 318       *
 319       *  @param integer $link_count Counter tracking current link number
 320       *  @param string $link URL of the link
 321       *  @param string $display Part of the text to associate number with
 322       *  @access private
 323       *  @return string
 324       */
 325      function _build_link_list($link_count, $link, $display)
 326      {
 327          $this->_link_list .= "[$link_count] $link\n";
 328  
 329          return $display . '[' . $link_count . ']';
 330      }
 331  
 332  }
 333  
 334  ?>