path: root/library/markdownify/parsehtml/parsehtml.php

<?php
/**
 * parseHTML is a HTML parser which works with PHP 4 and above.
 * It tries to handle invalid HTML to some degree.
 *
 * @version 1.0 beta
 * @author Milian Wolff (mail@milianw.de, http://milianw.de)
 * @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
 */
class parseHTML {
  /**
   * tags which are always empty (<br /> etc.)
   *
   * @var array<string>
   */
  var $emptyTags = array(
    'br',
    'hr',
    'input',
    'img',
    'area',
    'link',
    'meta',
    'param',
  );
  /**
   * tags with preformatted text
   * whitespaces wont be touched in them
   *
   * @var array<string>
   */
  var $preformattedTags = array(
    'script',
    'style',
    'pre',
    'code',
  );
  /**
   * supress HTML tags inside preformatted tags (see above)
   *
   * @var bool
   */
  var $noTagsInCode = false;
  /**
   * html to be parsed
   *
   * @var string
   */
  var $html = '';
  /**
   * node type:
   *
   * - tag (see isStartTag)
   * - text (includes cdata)
   * - comment
   * - doctype
   * - pi (processing instruction)
   *
   * @var string
   */
  var $nodeType = '';
  /**
   * current node content, i.e. either a
   * simple string (text node), or something like
   * <tag attrib="value"...>
   *
   * @var string
   */
  var $node = '';
  /**
   * wether current node is an opening tag (<a>) or not (</a>)
   * set to NULL if current node is not a tag
   * NOTE: empty tags (<br />) set this to true as well!
   *
   * @var bool | null
   */
  var $isStartTag = null;
  /**
   * wether current node is an empty tag (<br />) or not (<a></a>)
   *
   * @var bool | null
   */
  var $isEmptyTag = null;
  /**
   * tag name
   *
   * @var string | null
   */
  var $tagName = '';
  /**
   * attributes of current tag
   *
   * @var array (attribName=>value) | null
   */
  var $tagAttributes = null;
  /**
   * wether the current tag is a block element
   *
   * @var bool | null
   */
  var $isBlockElement = null;

  /**
   * keep whitespace
   *
   * @var int
   */
  var $keepWhitespace = 0;
  /**
   * list of open tags
   * count this to get current depth
   *
   * @var array
   */
  var $openTags = array();
  /**
   * list of block elements
   *
   * @var array
   * TODO: what shall we do with <del> and <ins> ?!
   */
  var $blockElements = array (
    # tag name => <bool> is block
    # block elements
    'address' => true,
    'blockquote' => true,
    'center' => true,
    'del' => true,
    'dir' => true,
    'div' => true,
    'dl' => true,
    'fieldset' => true,
    'form' => true,
    'h1' => true,
    'h2' => true,
    'h3' => true,
    'h4' => true,
    'h5' => true,
    'h6' => true,
    'hr' => true,
    'ins' => true,
    'isindex' => true,
    'menu' => true,
    'noframes' => true,
    'noscript' => true,
    'ol' => true,
    'p' => true,
    'pre' => true,
    'table' => true,
    'ul' => true,
    # set table elements and list items to block as well
    'thead' => true,
    'tbody' => true,
    'tfoot' => true,
    'td' => true,
    'tr' => true,
    'th' => true,
    'li' => true,
    'dd' => true,
    'dt' => true,
    # header items and html / body as well
    'html' => true,
    'body' => true,
    'head' => true,
    'meta' => true,
    'link' => true,
    'style' => true,
    'title' => true,
    # unfancy media tags, when indented should be rendered as block
    'map' => true,
    'object' => true,
    'param' => true,
    'embed' => true,
    'area' => true,
    # inline elements
    'a' => false,
    'abbr' => false,
    'acronym' => false,
    'applet' => false,
    'b' => false,
    'basefont' => false,
    'bdo' => false,
    'big' => false,
    'br' => false,
    'button' => false,
    'cite' => false,
    'code' => false,
    'del' => false,
    'dfn' => false,
    'em' => false,
    'font' => false,
    'i' => false,
    'img' => false,
    'ins' => false,
    'input' => false,
    'iframe' => false,
    'kbd' => false,
    'label' => false,
    'q' => false,
    'samp' => false,
    'script' => false,
    'select' => false,
    'small' => false,
    'span' => false,
    'strong' => false,
    'sub' => false,
    'sup' => false,
    'textarea' => false,
    'tt' => false,
    'var' => false,
  );
  /**
   * get next node, set $this->html prior!
   *
   * @param void
   * @return bool
   */
  function nextNode() {
    if (empty($this->html)) {
      # we are done with parsing the html string
      return false;
    }
    static $skipWhitespace = true;
    if ($this->isStartTag && !$this->isEmptyTag) {
      array_push($this->openTags, $this->tagName);
      if (in_array($this->tagName, $this->preformattedTags)) {
        # dont truncate whitespaces for <code> or <pre> contents
        $this->keepWhitespace++;
      }
    }

    if ($this->html[0] == '<') {
      $token = substr($this->html, 0, 9);
      if (substr($token, 0, 2) == '<?') {
        # xml prolog or other pi's
        /** TODO **/
        #trigger_error('this might need some work', E_USER_NOTICE);
        $pos = strpos($this->html, '>');
        $this->setNode('pi', $pos + 1);
        return true;
      }
      if (substr($token, 0, 4) == '<!--') {
        # comment
        $pos = strpos($this->html, '-->');
        if ($pos === false) {
          # could not find a closing -->, use next gt instead
          # this is firefox' behaviour
          $pos = strpos($this->html, '>') + 1;
        } else {
          $pos += 3;
        }
        $this->setNode('comment', $pos);

        $skipWhitespace = true;
        return true;
      }
      if ($token == '<!DOCTYPE') {
        # doctype
        $this->setNode('doctype', strpos($this->html, '>')+1);

        $skipWhitespace = true;
        return true;
      }
      if ($token == '<![CDATA[') {
        # cdata, use text node

        # remove leading <![CDATA[
        $this->html = substr($this->html, 9);

        $this->setNode('text', strpos($this->html, ']]>')+3);

        # remove trailing ]]> and trim
        $this->node = substr($this->node, 0, -3);
        $this->handleWhitespaces();

        $skipWhitespace = true;
        return true;
      }
      if ($this->parseTag()) {
        # seems to be a tag
        # handle whitespaces
        if ($this->isBlockElement) {
          $skipWhitespace = true;
        } else {
          $skipWhitespace = false;
        }
        return true;
      }
    }
    if ($this->keepWhitespace) {
      $skipWhitespace = false;
    }
    # when we get here it seems to be a text node
    $pos = strpos($this->html, '<');
    if ($pos === false) {
      $pos = strlen($this->html);
    }
    $this->setNode('text', $pos);
    $this->handleWhitespaces();
    if ($skipWhitespace && $this->node == ' ') {
      return $this->nextNode();
    }
    $skipWhitespace = false;
    return true;
  }
  /**
   * parse tag, set tag name and attributes, see if it's a closing tag and so forth...
   *
   * @param void
   * @return bool
   */
  function parseTag() {
    static $a_ord, $z_ord, $special_ords;
    if (!isset($a_ord)) {
      $a_ord = ord('a');
      $z_ord = ord('z');
      $special_ords = array(
        ord(':'), // for xml:lang
        ord('-'), // for http-equiv
      );
    }

    $tagName = '';

    $pos = 1;
    $isStartTag = $this->html[$pos] != '/';
    if (!$isStartTag) {
      $pos++;
    }
    # get tagName
    while (isset($this->html[$pos])) {
      $pos_ord = ord(strtolower($this->html[$pos]));
      if (($pos_ord >= $a_ord && $pos_ord <= $z_ord) || (!empty($tagName) && is_numeric($this->html[$pos]))) {
        $tagName .= $this->html[$pos];
        $pos++;
      } else {
        $pos--;
        break;
      }
    }

    $tagName = strtolower($tagName);
    if (empty($tagName) || !isset($this->blockElements[$tagName])) {
      # something went wrong => invalid tag
      $this->invalidTag();
      return false;
    }
    if ($this->noTagsInCode && end($this->openTags) == 'code' && !($tagName == 'code' && !$isStartTag)) {
      # we supress all HTML tags inside code tags
      $this->invalidTag();
      return false;
    }

    # get tag attributes
    /** TODO: in html 4 attributes do not need to be quoted **/
    $isEmptyTag = false;
    $attributes = array();
    $currAttrib = '';
    while (isset($this->html[$pos+1])) {
      $pos++;
      # close tag
      if ($this->html[$pos] == '>' || $this->html[$pos].$this->html[$pos+1] == '/>') {
        if ($this->html[$pos] == '/') {
          $isEmptyTag = true;
          $pos++;
        }
        break;
      }

      $pos_ord = ord(strtolower($this->html[$pos]));
      if ( ($pos_ord >= $a_ord && $pos_ord <= $z_ord) || in_array($pos_ord, $special_ords)) {
        # attribute name
        $currAttrib .= $this->html[$pos];
      } elseif (in_array($this->html[$pos], array(' ', "\t", "\n"))) {
        # drop whitespace
      } elseif (in_array($this->html[$pos].$this->html[$pos+1], array('="', "='"))) {
        # get attribute value
        $pos++;
        $await = $this->html[$pos]; # single or double quote
        $pos++;
        $value = '';
        while (isset($this->html[$pos]) && $this->html[$pos] != $await) {
          $value .= $this->html[$pos];
          $pos++;
        }
        $attributes[$currAttrib] = $value;
        $currAttrib = '';
      } else {
        $this->invalidTag();
        return false;
      }
    }
    if ($this->html[$pos] != '>') {
      $this->invalidTag();
      return false;
    }

    if (!empty($currAttrib)) {
      # html 4 allows something like <option selected> instead of <option selected="selected">
      $attributes[$currAttrib] = $currAttrib;
    }
    if (!$isStartTag) {
      if (!empty($attributes) || $tagName != end($this->openTags)) {
        # end tags must not contain any attributes
        # or maybe we did not expect a different tag to be closed
        $this->invalidTag();
        return false;
      }
      array_pop($this->openTags);
      if (in_array($tagName, $this->preformattedTags)) {
        $this->keepWhitespace--;
      }
    }
    $pos++;
    $this->node = substr($this->html, 0, $pos);
    $this->html = substr($this->html, $pos);
    $this->tagName = $tagName;
    $this->tagAttributes = $attributes;
    $this->isStartTag = $isStartTag;
    $this->isEmptyTag = $isEmptyTag || in_array($tagName, $this->emptyTags);
    if ($this->isEmptyTag) {
      # might be not well formed
      $this->node = preg_replace('# */? *>$#', ' />', $this->node);
    }
    $this->nodeType = 'tag';
    $this->isBlockElement = $this->blockElements[$tagName];
    return true;
  }
  /**
   * handle invalid tags
   *
   * @param void
   * @return void
   */
  function invalidTag() {
    $this->html = substr_replace($this->html, '&lt;', 0, 1);
  }
  /**
   * update all vars and make $this->html shorter
   *
   * @param string $type see description for $this->nodeType
   * @param int $pos to which position shall we cut?
   * @return void
   */
  function setNode($type, $pos) {
    if ($this->nodeType == 'tag') {
      # set tag specific vars to null
      # $type == tag should not be called here
      # see this::parseTag() for more
      $this->tagName = null;
      $this->tagAttributes = null;
      $this->isStartTag = null;
      $this->isEmptyTag = null;
      $this->isBlockElement = null;

    }
    $this->nodeType = $type;
    $this->node = substr($this->html, 0, $pos);
    $this->html = substr($this->html, $pos);
  }
  /**
   * check if $this->html begins with $str
   *
   * @param string $str
   * @return bool
   */
  function match($str) {
    return substr($this->html, 0, strlen($str)) == $str;
  }
  /**
   * truncate whitespaces
   *
   * @param void
   * @return void
   */
  function handleWhitespaces() {
    if ($this->keepWhitespace) {
      # <pre> or <code> before...
      return;
    }
    # truncate multiple whitespaces to a single one
    $this->node = preg_replace('#\s+#s', ' ', $this->node);
  }
  /**
   * normalize self::node
   *
   * @param void
   * @return void
   */
  function normalizeNode() {
    $this->node = '<';
    if (!$this->isStartTag) {
      $this->node .= '/'.$this->tagName.'>';
      return;
    }
    $this->node .= $this->tagName;
    foreach ($this->tagAttributes as $name => $value) {
      $this->node .= ' '.$name.'="'.str_replace('"', '&quot;', $value).'"';
    }
    if ($this->isEmptyTag) {
      $this->node .= ' /';
    }
    $this->node .= '>';
  }
}

/**
 * indent a HTML string properly
 *
 * @param string $html
 * @param string $indent optional
 * @return string
 */
function indentHTML($html, $indent = "  ", $noTagsInCode = false) {
  $parser = new parseHTML;
  $parser->noTagsInCode = $noTagsInCode;
  $parser->html = $html;
  $html = '';
  $last = true; # last tag was block elem
  $indent_a = array();
  while($parser->nextNode()) {
    if ($parser->nodeType == 'tag') {
      $parser->normalizeNode();
    }
    if ($parser->nodeType == 'tag' && $parser->isBlockElement) {
      $isPreOrCode = in_array($parser->tagName, array('code', 'pre'));
      if (!$parser->keepWhitespace && !$last && !$isPreOrCode) {
        $html = rtrim($html)."\n";
      }
      if ($parser->isStartTag) {
        $html .= implode($indent_a);
        if (!$parser->isEmptyTag) {
          array_push($indent_a, $indent);
        }
      } else {
        array_pop($indent_a);
        if (!$isPreOrCode) {
          $html .= implode($indent_a);
        }
      }
      $html .= $parser->node;
      if (!$parser->keepWhitespace && !($isPreOrCode && $parser->isStartTag)) {
        $html .= "\n";
      }
      $last = true;
    } else {
      if ($parser->nodeType == 'tag' && $parser->tagName == 'br') {
        $html .= $parser->node."\n";
        $last = true;
        continue;
      } elseif ($last && !$parser->keepWhitespace) {
        $html .= implode($indent_a);
        $parser->node = ltrim($parser->node);
      }
      $html .= $parser->node;

      if (in_array($parser->nodeType, array('comment', 'pi', 'doctype'))) {
        $html .= "\n";
      } else {
        $last = false;
      }
    }
  }
  return $html;
}
/*
# testcase / example
error_reporting(E_ALL);

$html = '<p>Simple block on one line:</p>

<div>foo</div>

<p>And nested without indentation:</p>

<div>
<div>
<div>
foo
</div>
<div style=">"/>
</div>
<div>bar</div>
</div>

<p>And with attributes:</p>

<div>
    <div id="foo">
    </div>
</div>

<p>This was broken in 1.0.2b7:</p>

<div class="inlinepage">
<div class="toggleableend">
foo
</div>
</div>';
#$html = '<a href="asdfasdf"       title=\'asdf\' foo="bar">asdf</a>';
echo indentHTML($html);
die();
*/