, ) * @license LGPL, see LICENSE_LGPL.txt and the summary below * @copyright (C) 2007 Milian Wolff * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * * This library is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public * License along with this library; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ /** * HTML Parser, see http://sf.net/projects/parseHTML */ require_once dirname(__FILE__) . '/parsehtml/parsehtml.php'; /** * default configuration */ define('MDFY_LINKS_EACH_PARAGRAPH', false); define('MDFY_BODYWIDTH', false); define('MDFY_KEEPHTML', true); /** * HTML to Markdown converter class */ class Markdownify { /** * html parser object * * @var parseHTML */ var $parser; /** * markdown output * * @var string */ var $output; /** * stack with tags which where not converted to html * * @var array */ var $notConverted = array(); /** * skip conversion to markdown * * @var bool */ var $skipConversion = false; /* options */ /** * keep html tags which cannot be converted to markdown * * @var bool */ var $keepHTML = false; /** * wrap output, set to 0 to skip wrapping * * @var int */ var $bodyWidth = 0; /** * minimum body width * * @var int */ var $minBodyWidth = 25; /** * display links after each paragraph * * @var bool */ var $linksAfterEachParagraph = false; /** * constructor, set options, setup parser * * @param bool $linksAfterEachParagraph wether or not to flush stacked links after each paragraph * defaults to false * @param int $bodyWidth wether or not to wrap the output to the given width * defaults to false * @param bool $keepHTML wether to keep non markdownable HTML or to discard it * defaults to true (HTML will be kept) * @return void */ function __construct($linksAfterEachParagraph = MDFY_LINKS_EACH_PARAGRAPH, $bodyWidth = MDFY_BODYWIDTH, $keepHTML = MDFY_KEEPHTML) { $this->linksAfterEachParagraph = $linksAfterEachParagraph; $this->keepHTML = $keepHTML; if ($bodyWidth > $this->minBodyWidth) { $this->bodyWidth = intval($bodyWidth); } else { $this->bodyWidth = false; } $this->parser = new parseHTML; $this->parser->noTagsInCode = true; # we don't have to do this every time $search = array(); $replace = array(); foreach ($this->escapeInText as $s => $r) { array_push($search, '#(?escapeInText = array( 'search' => $search, 'replace' => $replace ); } /** * parse a HTML string * * @param string $html * @return string markdown formatted */ function parseString($html) { $this->parser->html = $html; $this->parse(); return $this->output; } /** * tags with elements which can be handled by markdown * * @var array */ var $isMarkdownable = array( 'p' => array(), 'ul' => array(), 'ol' => array(), 'li' => array(), 'br' => array(), 'blockquote' => array(), 'code' => array(), 'pre' => array(), 'a' => array( 'href' => 'required', 'title' => 'optional', ), 'strong' => array(), 'b' => array(), 'em' => array(), 'i' => array(), 'img' => array( 'src' => 'required', 'alt' => 'optional', 'title' => 'optional', ), 'h1' => array(), 'h2' => array(), 'h3' => array(), 'h4' => array(), 'h5' => array(), 'h6' => array(), 'hr' => array(), ); /** * html tags to be ignored (contents will be parsed) * * @var array */ var $ignore = array( 'html', 'body', ); /** * html tags to be dropped (contents will not be parsed!) * * @var array */ var $drop = array( 'script', 'head', 'style', 'form', 'area', 'object', 'param', 'iframe', ); /** * Markdown indents which could be wrapped * @note: use strings in regex format * * @var array */ var $wrappableIndents = array( '\* ', # ul '\d. ', # ol '\d\d. ', # ol '> ', # blockquote '', # p ); /** * list of chars which have to be escaped in normal text * @note: use strings in regex format * * @var array * * TODO: what's with block chars / sequences at the beginning of a block? */ var $escapeInText = array( '([-*_])([ ]{0,2}\1){2,}' => '\\\\$0|', # hr '\*\*([^*\s]+)\*\*' => '\*\*$1\*\*', # strong '\*([^*\s]+)\*' => '\*$1\*', # em '__(?! |_)(.+)(?!<_| )__' => '\_\_$1\_\_', # em '_(?! |_)(.+)(?!<_| )_' => '\_$1\_', # em '`(.+)`' => '\`$1\`', # code '\[(.+)\](\s*\()' => '\[$1\]$2', # links: [text] (url) => [text\] (url) '\[(.+)\](\s*)\[(.*)\]' => '\[$1\]$2\[$3\]', # links: [text][id] => [text\][id\] ); /** * wether last processed node was a block tag or not * * @var bool */ var $lastWasBlockTag = false; /** * name of last closed tag * * @var string */ var $lastClosedTag = ''; /** * iterate through the nodes and decide what we * shall do with the current node * * @param void * @return void */ function parse() { $this->output = ''; # drop tags $this->parser->html = preg_replace('#<('.implode('|', $this->drop).')[^>]*>.*#sU', '', $this->parser->html); while ($this->parser->nextNode()) { switch ($this->parser->nodeType) { case 'doctype': break; case 'pi': case 'comment': if ($this->keepHTML) { $this->flushLinebreaks(); $this->out($this->parser->node); $this->setLineBreaks(2); } # else drop break; case 'text': $this->handleText(); break; case 'tag': if (in_array($this->parser->tagName, $this->ignore)) { break; } if ($this->parser->isStartTag) { $this->flushLinebreaks(); } if ($this->skipConversion) { $this->isMarkdownable(); # update notConverted $this->handleTagToText(); break; } if (!$this->parser->keepWhitespace && $this->parser->isBlockElement && $this->parser->isStartTag) { $this->parser->html = ltrim($this->parser->html); } if ($this->isMarkdownable()) { if ($this->parser->isBlockElement && $this->parser->isStartTag && !$this->lastWasBlockTag && !empty($this->output)) { if (!empty($this->buffer)) { $str =& $this->buffer[count($this->buffer) -1]; } else { $str =& $this->output; } if (substr($str, -strlen($this->indent)-1) != "\n".$this->indent) { $str .= "\n".$this->indent; } } $func = 'handleTag_'.$this->parser->tagName; $this->$func(); if ($this->linksAfterEachParagraph && $this->parser->isBlockElement && !$this->parser->isStartTag && empty($this->parser->openTags)) { $this->flushStacked(); } if (!$this->parser->isStartTag) { $this->lastClosedTag = $this->parser->tagName; } } else { $this->handleTagToText(); $this->lastClosedTag = ''; } break; default: trigger_error('invalid node type', E_USER_ERROR); break; } $this->lastWasBlockTag = $this->parser->nodeType == 'tag' && $this->parser->isStartTag && $this->parser->isBlockElement; } if (!empty($this->buffer)) { trigger_error('buffer was not flushed, this is a bug. please report!', E_USER_WARNING); while (!empty($this->buffer)) { $this->out($this->unbuffer()); } } ### cleanup $this->output = rtrim(str_replace('&', '&', str_replace('<', '<', str_replace('>', '>', $this->output)))); # end parsing, flush stacked tags $this->flushStacked(); $this->stack = array(); } /** * check if current tag can be converted to Markdown * * @param void * @return bool */ function isMarkdownable() { if (!isset($this->isMarkdownable[$this->parser->tagName])) { # simply not markdownable return false; } if ($this->parser->isStartTag) { $return = true; if ($this->keepHTML) { $diff = array_diff(array_keys($this->parser->tagAttributes), array_keys($this->isMarkdownable[$this->parser->tagName])); if (!empty($diff)) { # non markdownable attributes given $return = false; } } if ($return) { foreach ($this->isMarkdownable[$this->parser->tagName] as $attr => $type) { if ($type == 'required' && !isset($this->parser->tagAttributes[$attr])) { # required markdown attribute not given $return = false; break; } } } if (!$return) { array_push($this->notConverted, $this->parser->tagName.'::'.implode('/', $this->parser->openTags)); } return $return; } else { if (!empty($this->notConverted) && end($this->notConverted) === $this->parser->tagName.'::'.implode('/', $this->parser->openTags)) { array_pop($this->notConverted); return false; } return true; } } /** * output all stacked tags * * @param void * @return void */ function flushStacked() { # links foreach ($this->stack as $tag => $a) { if (!empty($a)) { call_user_func(array(&$this, 'flushStacked_'.$tag)); } } } /** * output link references (e.g. [1]: http://example.com "title"); * * @param void * @return void */ function flushStacked_a() { $out = false; foreach ($this->stack['a'] as $k => $tag) { if (!isset($tag['unstacked'])) { if (!$out) { $out = true; $this->out("\n\n", true); } else { $this->out("\n", true); } $this->out(' ['.$tag['linkID'].']: '.$tag['href'].(isset($tag['title']) ? ' "'.$tag['title'].'"' : ''), true); $tag['unstacked'] = true; $this->stack['a'][$k] = $tag; } } } /** * flush enqued linebreaks * * @param void * @return void */ function flushLinebreaks() { if ($this->lineBreaks && !empty($this->output)) { $this->out(str_repeat("\n".$this->indent, $this->lineBreaks), true); } $this->lineBreaks = 0; } /** * handle non Markdownable tags * * @param void * @return void */ function handleTagToText() { if (!$this->keepHTML) { if (!$this->parser->isStartTag && $this->parser->isBlockElement) { $this->setLineBreaks(2); } } else { # dont convert to markdown inside this tag /** TODO: markdown extra **/ if (!$this->parser->isEmptyTag) { if ($this->parser->isStartTag) { if (!$this->skipConversion) { $this->skipConversion = $this->parser->tagName.'::'.implode('/', $this->parser->openTags); } } else { if ($this->skipConversion == $this->parser->tagName.'::'.implode('/', $this->parser->openTags)) { $this->skipConversion = false; } } } if ($this->parser->isBlockElement) { if ($this->parser->isStartTag) { if (in_array($this->parent(), array('ins', 'del'))) { # looks like ins or del are block elements now $this->out("\n", true); $this->indent(' '); } if ($this->parser->tagName != 'pre') { $this->out($this->parser->node."\n".$this->indent); if (!$this->parser->isEmptyTag) { $this->indent(' '); } else { $this->setLineBreaks(1); } $this->parser->html = ltrim($this->parser->html); } else { # don't indent inside
 tags
            $this->out($this->parser->node);
            static $indent;
            $indent =  $this->indent;
            $this->indent = '';
          }
        } else {
          if (!$this->parser->keepWhitespace) {
            $this->output = rtrim($this->output);
          }
          if ($this->parser->tagName != 'pre') {
            $this->indent('  ');
            $this->out("\n".$this->indent.$this->parser->node);
          } else {
            # reset indentation
            $this->out($this->parser->node);
            static $indent;
            $this->indent = $indent;
          }

          if (in_array($this->parent(), array('ins', 'del'))) {
            # ins or del was block element
            $this->out("\n");
            $this->indent('  ');
          }
          if ($this->parser->tagName == 'li') {
            $this->setLineBreaks(1);
          } else {
            $this->setLineBreaks(2);
          }
        }
      } else {
        $this->out($this->parser->node);
      }
      if (in_array($this->parser->tagName, array('code', 'pre'))) {
        if ($this->parser->isStartTag) {
          $this->buffer();
        } else {
          # add stuff so cleanup just reverses this
          $this->out(str_replace('<', '&lt;', str_replace('>', '&gt;', $this->unbuffer())));
        }
      }
    }
  }
  /**
   * handle plain text
   *
   * @param void
   * @return void
   */
  function handleText() {
    if ($this->hasParent('pre') && strpos($this->parser->node, "\n") !== false) {
      $this->parser->node = str_replace("\n", "\n".$this->indent, $this->parser->node);
    }
    if (!$this->hasParent('code') && !$this->hasParent('pre')) {
      # entity decode
      $this->parser->node = $this->decode($this->parser->node);
      if (!$this->skipConversion) {
        # escape some chars in normal Text
        $this->parser->node = preg_replace($this->escapeInText['search'], $this->escapeInText['replace'], $this->parser->node);
      }
    } else {
      $this->parser->node = str_replace(array('"', '&apos'), array('"', '\''), $this->parser->node);
    }
    $this->out($this->parser->node);
    $this->lastClosedTag = '';
  }
  /**
   * handle  and  tags
   *
   * @param void
   * @return void
   */
  function handleTag_em() {
    $this->out('*', true);
  }
  function handleTag_i() {
    $this->handleTag_em();
  }
  /**
   * handle  and  tags
   *
   * @param void
   * @return void
   */
  function handleTag_strong() {
    $this->out('**', true);
  }
  function handleTag_b() {
    $this->handleTag_strong();
  }
  /**
   * handle 

tags * * @param void * @return void */ function handleTag_h1() { $this->handleHeader(1); } /** * handle

tags * * @param void * @return void */ function handleTag_h2() { $this->handleHeader(2); } /** * handle

tags * * @param void * @return void */ function handleTag_h3() { $this->handleHeader(3); } /** * handle

tags * * @param void * @return void */ function handleTag_h4() { $this->handleHeader(4); } /** * handle

tags * * @param void * @return void */ function handleTag_h5() { $this->handleHeader(5); } /** * handle
tags * * @param void * @return void */ function handleTag_h6() { $this->handleHeader(6); } /** * number of line breaks before next inline output */ var $lineBreaks = 0; /** * handle header tags (

-

) * * @param int $level 1-6 * @return void */ function handleHeader($level) { if ($this->parser->isStartTag) { $this->out(str_repeat('#', $level).' ', true); } else { $this->setLineBreaks(2); } } /** * handle

tags * * @param void * @return void */ function handleTag_p() { if (!$this->parser->isStartTag) { $this->setLineBreaks(2); } } /** * handle tags * * @param void * @return void */ function handleTag_a() { if ($this->parser->isStartTag) { $this->buffer(); if (isset($this->parser->tagAttributes['title'])) { $this->parser->tagAttributes['title'] = $this->decode($this->parser->tagAttributes['title']); } else { $this->parser->tagAttributes['title'] = null; } $this->parser->tagAttributes['href'] = $this->decode(trim($this->parser->tagAttributes['href'])); $this->stack(); } else { $tag = $this->unstack(); $buffer = $this->unbuffer(); if (empty($tag['href']) && empty($tag['title'])) { # empty links... testcase mania, who would possibly do anything like that?! $this->out('['.$buffer.']()', true); return; } if ($buffer == $tag['href'] && empty($tag['title'])) { # $this->out('<'.$buffer.'>', true); return; } $bufferDecoded = $this->decode(trim($buffer)); if (substr($tag['href'], 0, 7) == 'mailto:' && 'mailto:'.$bufferDecoded == $tag['href']) { if (is_null($tag['title'])) { # $this->out('<'.$bufferDecoded.'>', true); return; } # [mail@example.com][1] # ... # [1]: mailto:mail@example.com Title $tag['href'] = 'mailto:'.$bufferDecoded; } if ($this->linksAfterEachParagraph!=='inline'){ # [This link][id] foreach ($this->stack['a'] as $tag2) { if ($tag2['href'] == $tag['href'] && $tag2['title'] === $tag['title']) { $tag['linkID'] = $tag2['linkID']; break; } } if (!isset($tag['linkID'])) { $tag['linkID'] = count($this->stack['a']) + 1; array_push($this->stack['a'], $tag); } $this->out('['.trim($buffer).']['.$tag['linkID'].']', true); } else { # [This link|title](url) if ($tag['title']) $buffer.="|".$tag['title']; $this->out('['.trim($buffer).']('.$tag['href'].')', true); } } } /** * handle tags * * @param void * @return void */ function handleTag_img() { if (!$this->parser->isStartTag) { return; # just to be sure this is really an empty tag... } if (isset($this->parser->tagAttributes['title'])) { $this->parser->tagAttributes['title'] = $this->decode($this->parser->tagAttributes['title']); } else { $this->parser->tagAttributes['title'] = null; } if (isset($this->parser->tagAttributes['alt'])) { $this->parser->tagAttributes['alt'] = $this->decode($this->parser->tagAttributes['alt']); } else { $this->parser->tagAttributes['alt'] = null; } if (empty($this->parser->tagAttributes['src'])) { # support for "empty" images... dunno if this is really needed # but there are some testcases which do that... if (!empty($this->parser->tagAttributes['title'])) { $this->parser->tagAttributes['title'] = ' '.$this->parser->tagAttributes['title'].' '; } $this->out('!['.$this->parser->tagAttributes['alt'].']('.$this->parser->tagAttributes['title'].')', true); return; } else { $this->parser->tagAttributes['src'] = $this->decode($this->parser->tagAttributes['src']); } # [This link][id] $link_id = false; if (!empty($this->stack['a'])) { foreach ($this->stack['a'] as $tag) { if ($tag['href'] == $this->parser->tagAttributes['src'] && $tag['title'] === $this->parser->tagAttributes['title']) { $link_id = $tag['linkID']; break; } } } else { $this->stack['a'] = array(); } if (!$link_id) { $link_id = count($this->stack['a']) + 1; $tag = array( 'href' => $this->parser->tagAttributes['src'], 'linkID' => $link_id, 'title' => $this->parser->tagAttributes['title'] ); array_push($this->stack['a'], $tag); } $this->out('!['.$this->parser->tagAttributes['alt'].']['.$link_id.']', true); } /** * handle tags * * @param void * @return void */ function handleTag_code() { if ($this->hasParent('pre')) { # ignore code blocks inside

      return;
    }
    if ($this->parser->isStartTag) {
      $this->buffer();
    } else {
      $buffer = $this->unbuffer();
      # use as many backticks as needed
      preg_match_all('#`+#', $buffer, $matches);
      if (!empty($matches[0])) {
        rsort($matches[0]);

        $ticks = '`';
        while (true) {
          if (!in_array($ticks, $matches[0])) {
            break;
          }
          $ticks .= '`';
        }
      } else {
        $ticks = '`';
      }
      if ($buffer[0] == '`' || substr($buffer, -1) == '`') {
        $buffer = ' '.$buffer.' ';
      }
      $this->out($ticks.$buffer.$ticks, true);
    }
  }
  /**
   * handle 
 tags
   *
   * @param void
   * @return void
   */
  function handleTag_pre() {
    if ($this->keepHTML && $this->parser->isStartTag) {
      # check if a simple  follows
      if (!preg_match('#^\s*#Us', $this->parser->html)) {
        # this is no standard markdown code block
        $this->handleTagToText();
        return;
      }
    }
    $this->indent('    ');
    if (!$this->parser->isStartTag) {
      $this->setLineBreaks(2);
    } else {
      $this->parser->html = ltrim($this->parser->html);
    }
  }
  /**
   * handle 
tags * * @param void * @return void */ function handleTag_blockquote() { $this->indent('> '); } /** * handle