/*****************************************************************
 * Word HTML Cleaner
 * Copyright (C) 2005 Connor McKay
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version 2
 * of the License, or (at your option) any later version.
 *
 * This program 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 General Public License for more details.
*****************************************************************/

//remplacement characters
var rchars = [["–","—", "‘", "’", "“", "”", ' '], ["-", "-", "'", "'", '"', '"', ' ']];

//html entities translation array
var hents = new Array();
hents['¡'] = '¡';
hents['¢'] = '¢';
hents['£'] = '£';
hents['¤'] = '¤';
hents['¥'] = '¥';
hents['¦'] = '¦';
hents['§'] = '§';
hents['¨'] = '¨';
hents['©'] = '©';
hents['ª'] = 'ª';
hents['«'] = '«';
hents['¬'] = '¬';
hents['­'] = '­';
hents['®'] = '®';
hents['¯'] = '¯';
hents['°'] = '°';
hents['±'] = '±';
hents['²'] = '²';
hents['³'] = '³';
hents['´'] = '´';
hents['µ'] = 'µ';
hents['¶'] = '¶';
hents['·'] = '·';
hents['¸'] = '¸';
hents['¹'] = '¹';
hents['º'] = 'º';
hents['»'] = '»';
hents['¼'] = '¼';
hents['½'] = '½';
hents['¾'] = '¾';
hents['¿'] = '¿';
hents['À'] = 'À';
hents['Á'] = 'Á';
hents['Â'] = 'Â';
hents['Ã'] = 'Ã';
hents['Ä'] = 'Ä';
hents['Å'] = 'Å';
hents['Æ'] = 'Æ';
hents['Ç'] = 'Ç';
hents['È'] = 'È';
hents['É'] = 'É';
hents['Ê'] = 'Ê';
hents['Ë'] = 'Ë';
hents['Ì'] = 'Ì';
hents['Í'] = 'Í';
hents['Î'] = 'Î';
hents['Ï'] = 'Ï';
hents['Ð'] = 'Ð';
hents['Ñ'] = 'Ñ';
hents['Ò'] = 'Ò';
hents['Ó'] = 'Ó';
hents['Ô'] = 'Ô';
hents['Õ'] = 'Õ';
hents['Ö'] = 'Ö';
hents['×'] = '×';
hents['Ø'] = 'Ø';
hents['Ù'] = 'Ù';
hents['Ú'] = 'Ú';
hents['Û'] = 'Û';
hents['Ü'] = 'Ü';
hents['Ý'] = 'Ý';
hents['Þ'] = 'Þ';
hents['ß'] = 'ß';
hents['à'] = 'à';
hents['á'] = 'á';
hents['â'] = 'â';
hents['ã'] = 'ã';
hents['ä'] = 'ä';
hents['å'] = 'å';
hents['æ'] = 'æ';
hents['ç'] = 'ç';
hents['è'] = 'è';
hents['é'] = 'é';
hents['ê'] = 'ê';
hents['ë'] = 'ë';
hents['ì'] = 'ì';
hents['í'] = 'í';
hents['î'] = 'î';
hents['ï'] = 'ï';
hents['ð'] = 'ð';
hents['ñ'] = 'ñ';
hents['ò'] = 'ò';
hents['ó'] = 'ó';
hents['ô'] = 'ô';
hents['õ'] = 'õ';
hents['ö'] = 'ö';
hents['÷'] = '÷';
hents['ø'] = 'ø';
hents['ù'] = 'ù';
hents['ú'] = 'ú';
hents['û'] = 'û';
hents['ü'] = 'ü';
hents['ý'] = 'ý';
hents['þ'] = 'þ';
hents['ÿ'] = 'ÿ';
hents['"'] = '"';
hents['<'] = '&lt;';
hents['>'] = '&gt;';

//allowed tags
var tags = ['p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'h7', 'h8', 'ul', 'ol', 'li', 'u', 'i', 'b', 'a', 'table', 'tr', 'th', 'td', 'img', 'em', 'strong', 'br'];

//tags which should be removed when empty
var rempty = ['p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'h7', 'h8', 'ul', 'ol', 'li', 'u', 'i', 'b', 'a', 'table', 'tr', 'em', 'strong'];

//allowed atributes for tags
var aattr = new Array();
aattr['a'] = ['href', 'name'];
aattr['table'] = ['border'];
aattr['th'] = ['colspan', 'rowspan'];
aattr['td'] = ['colspan', 'rowspan'];
aattr['img'] = ['src', 'width', 'height', 'alt'];

//tags who's content should be deleted
var dctags = ['head'];

//Quote characters
var quotes = ["'", '"'];

//tags which are displayed as a block
var btags = ['p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'h7', 'h8', 'ul', 'ol', 'table', 'tr', 'th', 'td', 'br'];

//d = data, o = out, c = character, n = next character
//in and out variables
var d = '';
var o = '';

//pass in the data to clean

function cleanHTML (data)
{
	d = data;
	o = '';
	var i;

	//Replace all whitespace characters with spaces
	d = d.replace(/(\s|&nbsp;)+/g, ' ');
	//replace weird word characters
	for (i = 0; i < rchars[0].length; i++)
		d = d.replace(new RegExp(rchars[0][i], 'g'), rchars[1][i]);
	
	//initialize flags
	//what the next character is expected to be
	var expected = '';
	//tag text
	var tag = '';
	//tag name
	var tagname = '';
	//what type of tag it is, start, end, or single
	var tagtype = 'start';
	//attribute text
	var attribute = '';
	//attribute name
	var attributen = '';
	//if the attribute has had an equals sign
	var attributeequals = false;
	//if attribute has quotes, and what they are
	var attributequotes = '';
	
	var c = '';
	var n = '';
	
	/*Parser format:
	The parser is divided into three parts:
	The first section is for when the current type of character is known
	The second is for when it is an unknown character in a tag
	The third is for anything outside of a tag
	*/
	
	//editing pass
	for (i = 0; i < d.length; i++)
	{
		//current character
		c = getc(i);
		//next character
		n = getc(i+1);
		
		//***Section for when the current character is known
		
		//if the tagname is expected
		if (expected == 'tagname')
		{
			tagname += c.toLowerCase();
			//lookahead for end of tag name
			if (n == ' ' || n == '>' || n == '/')
			{
				tag += tagname;
				expected = 'tag';
			}
		}
		//if an attribute name is expected
		else if (expected == 'attributen')
		{
			attributen += c.toLowerCase();
			//lookahead for end of attribute name
			if (n == ' ' || n == '>' || n == '/' || n == '=')
			{
				attribute += attributen;
				//check to see if its an attribute without an assigned value
				//determines whether there is anything but spaces between the attribute name and the next equals sign
				if (endOfAttr(i))
				{
					//if the attribute is allowed, add it to the output
					if (ae(attributen, aattr[tagname]))
						tag += attribute;
					
					attribute = '';
					attributen = '';
					attributeequals = false;
					attributequotes = '';
				}
				expected = 'tag';
			}
		}
		//if an attribute value is expected
		else if (expected == 'attributev')
		{
			attribute += c;
			
			//lookahead for end of value
			if ((c == attributequotes) || ((n == ' ' || n == '/' || n == '>') && !attributequotes))
			{
				//if the attribute is allowed, add it to the output
				if (ae(attributen, aattr[tagname]))
					tag += attribute;
				
				attribute = '';
				attributen = '';
				attributeequals = false;
				attributequotes = '';
				
				expected = 'tag';
			}
		}
		
		//***Section for when the character is unknown but it is inside of a tag
		
		else if (expected == 'tag')
		{
			//if its a space
			if (c == ' ')
				tag += c;
			//if its a slash after the tagname, signalling a single tag.
			else if (c == '/' && tagname)
			{
				tag += c;
				tagtype = 'single';
			}
			//if its a slash before the tagname, signalling its an end tag
			else if (c == '/')
			{
				tag += c;
				tagtype = 'end';
			}
			//if its the end of a tag
			else if (c == '>')
			{
				tag += c;
				//if the tag is allowed, add it to the output
				if (ae(tagname, tags))
					o += tag;
				
				//if its a start tag
				if (tagtype == 'start')
				{
					//if the tag is supposed to have its contents deleted
					if (ae(tagname, dctags))
					{
						//if there is an end tag, skip to it in order to delete the tags contents
						if (-1 != (endpos = d.indexOf('</' + tagname, i)))
						{
							//have to make it one less because i gets incremented at the end of the loop
							i = endpos-1;
						}
						//if there isn't an end tag, then it was probably a non-compliant single tag
					}
				}
				
				tag = '';
				tagname = '';
				tagtype = 'start';
				expected = '';
			}
			//if its an attribute name
			else if (tagname && !attributen)
			{
				attributen += c.toLowerCase();
				expected = 'attributen';
				//lookahead for end of attribute name, in case its a one character attribute name
				if (n == ' ' || n == '>' || n == '/' || n == '=')
				{
					attribute += attributen;
					//check to see if its an attribute without an assigned value
					//determines whether there is anything but spaces between the attribute name and the next equals sign
					if (endOfAttr(i))
					{
						//if the attribute is allowed, add it to the output
						if (ae(attributen, attributen))
							tag += attribute;
						
						attribute = '';
						attributen = '';
						attributeequals = false;
						attributequotes = '';
					}
					expected = 'tag';
				}
			}
			//if its a start quote for an attribute value
			else if (ae(c, quotes) && attributeequals)
			{
				attribute += c;
				attributequotes = c;
				expected = 'attributev';
			}
			//if its an attribute value
			else if (attributeequals)
			{
				attribute += c;
				expected = 'attributev';
				
				//lookahead for end of value, in case its only one character
				if ((c == attributequotes) || ((n == ' ' || n == '/' || n == '>') && !attributequotes))
				{
					//if the attribute is allowed, add it to the output
					if (ae(attributen, attributen))
						tag += attribute;
					
					attribute = '';
					attributen = '';
					attributeequals = false;
					attributequotes = '';
					
					expected = 'tag';
				}
			}
			//if its an attribute equals
			else if (c == '=' && attributen)
			{
				attribute += c;
				attributeequals = true;
			}
			//if its the tagname
			else
			{
				tagname += c.toLowerCase();
				expected = 'tagname';
				
				//lookahead for end of tag name, in case its a one character tag name
				if (n == ' ' || n == '>' || n == '/')
				{
					tag += tagname;
					expected = 'tag';
				}
			}
		}
		//if nothing is expected
		else
		{
			//if its the start of a tag
			if (c == '<')
			{
				tag = c;
				expected = 'tag';
			}
			//anything else
			else
				o += htmlentities(c);
		}
	}
	
	//beautifying regexs
	//remove duplicate spaces
	o = o.replace(/\s+/g, ' ');
	//remove unneeded spaces in tags
	o = o.replace(/\s>/g, '>');
	//remove empty tags
	//this loops until there is no change from running the regex
	var remptys = rempty.join('|');
	var oo = o;
	while ((o = o.replace(new RegExp("\\s?<(" + remptys + ")>\s*<\\/\\1>", 'gi'), '')) != oo)
		oo = o;
	//make block tags regex string
	var btagss = btags.join('|');
	//add newlines after block tags
	o = o.replace(new RegExp("\\s?</(" + btagss+ ")>", 'gi'), "</$1>\n");
	//remove spaces before block tags
	o = o.replace(new RegExp("\\s<(" + btagss + ")", 'gi'), "<$1");
	
	//fix lists
	o = o.replace(/((<p.*>\s*(&middot;|&#9642;) .*<\/p.*>\n)+)/gi, "<ul>\n$1</ul>\n");//make ul for dot lists
	o = o.replace(/((<p.*>\s*\d+\S*\. .*<\/p.*>\n)+)/gi, "<ol>\n$1</ol>\n");//make ol for numerical lists
	o = o.replace(/((<p.*>\s*[a-z]+\S*\. .*<\/p.*>\n)+)/gi, "<ol style=\"list-style-type: lower-latin;\">\n$1</ol>\n");//make ol for latin lists
	o = o.replace(/<p(.*)>\s*(&middot;|&#9642;|\d+(\S*)\.|[a-z]+\S*\.) (.*)<\/p(.*)>\n/gi, "\t<li$1>$3$4</li$5>\n");//make li
	
	//extend outer lists around the nesting lists
	o = o.replace(/<\/(ul|ol|ol style="list-style-type: lower-latin;")>\n(<(?:ul|ol|ol style="list-style-type: lower-latin;")>[\s\S]*<\/(?:ul|ol|ol style="list-style-type: lower-latin;")>)\n(?!<(ul|ol|ol style="list-style-type: lower-latin;")>)/g, "</$1>\n$2\n<$1>\n</$1>\n");
	
	//nesting lists
	o = o.replace(/<\/li>\s+<\/ol>\s+<ul>([\s\S]*?)<\/ul>\s+<ol>/g, "\n<ul>$1</ul></li>");//ul in ol
	o = o.replace(/<\/li>\s+<\/ol>\s+<ol style="list-style-type: lower-latin;">([\s\S]*?)<\/ol>\s+<ol>/g, "\n<ol style=\"list-style-type: lower-latin;\">$1</ol></li>");//latin in ol
	o = o.replace(/<\/li>\s+<\/ul>\s+<ol>([\s\S]*?)<\/ol>\s+<ul>/g, "\n<ol>$1</ol></li>");//ol in ul
	o = o.replace(/<\/li>\s+<\/ul>\s+<ol style="list-style-type: lower-latin;">([\s\S]*?)<\/ol>\s+<ul>/g, "\n<ol style=\"list-style-type: lower-latin;\">$1</ol></li>");//latin in ul
	o = o.replace(/<\/li>\s+<\/ol>\s+<ol style="list-style-type: lower-latin;">([\s\S]*?)<\/ol>\s+<ol>/g, "\n<ol style=\"list-style-type: lower-latin;\">$1</ol></li>");//ul in latin
	o = o.replace(/<\/li>\s+<\/ul>\s+<ol style="list-style-type: lower-latin;">([\s\S]*?)<\/ol>\s+<ul>/g, "\n<ol style=\"list-style-type: lower-latin;\">$1</ol></li>");//ul in latin
	//remove empty tags. this is needed a second time to delete empty lists that were created to fix nesting, but weren't needed
	o = o.replace(new RegExp("\\s?<(" + remptys + ")>\s*<\\/\\1>", 'gi'), '');
	
	return o;
}

//array equals
//loops through all the elements of an array to see if any of them equal the test.
function ae (needle, haystack)
{
	if (typeof(haystack) == 'object')
		for (var i = 0; i < haystack.length; i++)
			if (needle == haystack[i])
				return true;
	
	return false;
}

//get character
//return specified character from d
function getc (i)
{
	return d.charAt(i);
}

//end of attr
//determines if their is anything but spaces between the current character, and the next equals sign
function endOfAttr (i)
{
	var between = d.substring(i+1, d.indexOf('=', i+1));
	if (between.replace(/\s+/g, ''))
		return true;
	else
		return false;
}

function htmlentities (character)
{
	if (hents[character])
		return hents[character];
	else
		return character;
}