/*
 * This file is a part of CMSLU -- Content Management System like Unix
 * and is not publicly distributed
 * 
 * Copyright (c) 2008-2009, Tomasz Sowa
 * All rights reserved.
 * 
 */

#include "htmlfilter.h"



void HTMLFilter::Item::Clear()
{
	name[0]  = 0;
	name_len = 0;
	type     = none;
	new_line = false;
}


HTMLFilter::Item::Item()
{
	Clear();
}



void HTMLFilter::Filter(const char * in, std::string & out)
{
	pchar         = in;
	stack_len     = 0;
	out_string    = &out;
	last_new_line = false;
	out_string->clear();

	Read();
}


void HTMLFilter::Filter(const std::string & in, std::string & out)
{
	out.reserve(in.size() * 2 + 1);
	Filter(in.c_str(), out);
}


HTMLFilter::HTMLFilter()
{
	pstack = new Item[CMSLU_HTMLFILTER_STACK_MAXLEN];
	buffer = new char[CMSLU_HTMLFILTER_BUFFER_MAXLEN];

	tab_size         = 2;
	trim_white       = false;
	break_long_lines = false;
}


HTMLFilter::HTMLFilter(const HTMLFilter & f)
{
	// don't need to copy the stack
	pstack = new Item[CMSLU_HTMLFILTER_STACK_MAXLEN];
	buffer = new char[CMSLU_HTMLFILTER_BUFFER_MAXLEN];
}


HTMLFilter & HTMLFilter::operator=(const HTMLFilter & f)
{
	// don't need to copy the stack
	pstack = new Item[CMSLU_HTMLFILTER_STACK_MAXLEN];
	buffer = new char[CMSLU_HTMLFILTER_BUFFER_MAXLEN];

return *this;
}


HTMLFilter::~HTMLFilter()
{
	delete [] pstack;
	delete [] buffer;
}


void HTMLFilter::BreakLongLines(bool break_lines)
{
	break_long_lines = break_lines;
}


void HTMLFilter::TrimWhite(bool trim)
{
	trim_white = trim;
}


void HTMLFilter::InsertTabs(size_t tabsize)
{
	tab_size = tabsize;

	if( tab_size > 1000 )
		tab_size = 1000;
}


HTMLFilter::Item & HTMLFilter::GetItem(size_t i)
{
	if( i >= stack_len )
	{
		empty.Clear();
		return empty;
	}

return pstack[i];
}


HTMLFilter::Item & HTMLFilter::LastItem()
{
	if( stack_len == 0 )
	{
		empty.Clear();
		return empty;
	}

return pstack[stack_len-1];
}


bool HTMLFilter::PushStack()
{
	if( stack_len == CMSLU_HTMLFILTER_STACK_MAXLEN )
		// oops, too many items
		return false;

	pstack[stack_len].Clear();
	stack_len += 1;

return true;
}

void HTMLFilter::PopStack()
{
	if( stack_len == 0 )
		// oops
		return;

	stack_len -= 1;
	pstack[stack_len].Clear();
}


bool HTMLFilter::IsWhite(int c)
{
	// dont use c==10 here

	if( c==' ' || c=='\t' || c==13 )
		return true;

return false;
}


void HTMLFilter::SkipWhite()
{
	while( IsWhite(*pchar) )
		++pchar;
}


void HTMLFilter::SkipWhiteLines()
{
	while( *pchar==10 || IsWhite(*pchar) )
		++pchar;
}


bool HTMLFilter::SkipTagCheck()
{
	pchar += 1;
	SkipWhite();

	if( *pchar == '/' )
	{
		pchar += 1;
		SkipWhite();

		if( IsNameEqual(pchar, LastItem().name, LastItem().name_len) )
		{
			pchar += LastItem().name_len;
			SkipWhite();

			if( *pchar == '>' )
			{
				pchar += 1;
				return true;
			}
		}
	}

return false;
}


void HTMLFilter::SkipNormalText()
{
	while( *pchar!=0 && *pchar!='<' )
		++pchar;
}

void HTMLFilter::CheckNewLine()
{
const char * start = pchar;

	SkipWhite();
	last_new_line = (*pchar==10);

	pchar = start;
}


void HTMLFilter::PutLastTagWithClosingTag()
{
const char * start = pchar;


	while( *pchar != 0 )
	{
		if( *pchar == '<' )
		{
			if( SkipTagCheck() )
			{
				PopStack();
				CheckNewLine();
				break;
			}
		}
		else
		{
			pchar += 1;
		}
	}

	Put(start, pchar);
}



void HTMLFilter::SkipItem()
{
	while( *pchar!=0 && *pchar!='>' )
		++pchar;

	if( *pchar == '>' )
		++pchar;
}


void HTMLFilter::SkipItemCheckXmlSimple()
{
	while( *pchar!=0 )
	{
		while( *pchar!=0 && *pchar!='>' && *pchar!='/')
			++pchar;

		if( *pchar == '/' ) // closing xml tag
		{
			++pchar;
			SkipWhite();

			if( *pchar == '>' )
			{
				++pchar;
				LastItem().type = Item::simple;
				break;
			}
		}
		else
		if( *pchar == '>' )
		{
			++pchar;
			break;
		}
	}
}


bool HTMLFilter::IsValidCharForName(int c)
{
	if( (c>='a' && c<='z') ||
		(c>='A' && c<='Z') ||
		(c>='0' && c<='9') ||
		c=='-' || c=='!' )
		return true;

return false;
}



void HTMLFilter::ReadItemName()
{
size_t i;

	for( i=0 ; IsValidCharForName(*pchar) && i<CMSLU_HTMLFILTER_ITEM_MAXLEN-1 ; ++i )
	{
		LastItem().name[i] = *pchar;
		++pchar;
	}

	LastItem().name[i] = 0;
	LastItem().name_len = i;
}



void HTMLFilter::Put(const char * str, const char * end)
{
	if( str>=end )
		return;

	size_t len = end - str;
	out_string->append(str, len);
}


size_t HTMLFilter::PutTrimFillBuffer(const char * & str, const char * & end)
{
size_t non_whites = 0;
size_t i = 0;
bool was_white;


	for( ; str < end && i<CMSLU_HTMLFILTER_BUFFER_MAXLEN-1 ; ++str )
	{
		was_white = false;

		if( *str==10 || IsWhite(*str) )
		{
			// skipping the whole white string
			for( ; (*str==10 || IsWhite(*str)) && str < end ; ++str );

			was_white = true;
			non_whites = 0;
		}
		else
		{
			non_whites += 1;
		}

		if( (non_whites>60 && break_long_lines) || was_white )
		{
			buffer[i] = ' ';
			i += 1;
			non_whites = 0;
		}

		if( str < end )
		{
			buffer[i] = *str;
			i += 1;
		}
	}

return i;
}


// if there are more than one white characters then they will be changed into one space
// and putting a space between 50 characters (if there were not any other white characters between them)
void HTMLFilter::PutTrim(const char * str, const char * end)
{
size_t buf_len;

	// this buffer must have at least 2 bytes (PutTrimFillBuffer needs it)
	if( !trim_white || CMSLU_HTMLFILTER_BUFFER_MAXLEN < 2 )
	{
		Put(str, end);
		return;
	}

	while( str < end )
	{
		buf_len = PutTrimFillBuffer(str, end);
		Put(buffer, buffer+buf_len);
	}
}


void HTMLFilter::PutOpeningTag(const char * tag)
{
size_t i;

	if( CMSLU_HTMLFILTER_BUFFER_MAXLEN < CMSLU_HTMLFILTER_ITEM_MAXLEN+2 )
		return;

	buffer[0] = '<';

	for(i=1 ; *tag ; ++i, ++tag)
		buffer[i] = *tag;

	buffer[i] = '>';
	i += 1;

	Put(buffer, buffer+i);
}



void HTMLFilter::PutClosingTag(const char * tag)
{
size_t i;

	if( CMSLU_HTMLFILTER_BUFFER_MAXLEN < CMSLU_HTMLFILTER_ITEM_MAXLEN+3 )
		return;

	buffer[0] = '<';
	buffer[1] = '/';

	for(i=2 ; *tag ; ++i, ++tag)
		buffer[i] = *tag;

	buffer[i] = '>';
	i += 1;

	Put(buffer, buffer+i);
}


void HTMLFilter::PutTabs(size_t len)
{
	if( len == 0 )
		return;

	if( len > 20 )
		len = 20;

	// how many spaces do you want
	size_t spaces = (len-1) * tab_size;
	size_t i;

	if( spaces < CMSLU_HTMLFILTER_BUFFER_MAXLEN )
	{
		for(i=0 ; i<spaces ; ++i)
			buffer[i] = ' ';

		Put(buffer, buffer+i);
	}
}


void HTMLFilter::PutNewLine()
{
	if( !trim_white )
		// new line characters will be directly from the source html
		return;

	buffer[0] = 10;
	Put(buffer, buffer+1);
}



bool HTMLFilter::IsOpeningCommentaryTag()
{
static char comm_open[] = "<!--";
size_t comm_open_len = sizeof(comm_open) / sizeof(char) - 1;

	return IsNameEqual(pchar, comm_open, comm_open_len);
}



// skipping the commentary tag if exists
bool HTMLFilter::SkipCommentaryTagIfExists()
{
static char comm_close[] = "-->";
size_t comm_close_len = sizeof(comm_close) / sizeof(char) - 1;

	if( !IsOpeningCommentaryTag() )
		return false;

	pchar += 4; // size of commentary opening tag

	// looking for "-->"
	while( *pchar!=0 && !IsNameEqual(pchar, comm_close, comm_close_len) )
		++pchar;

	if( *pchar!= 0 )
		pchar += comm_close_len;

	CheckNewLine();

return true;
}


void HTMLFilter::ReadNormalText()
{
const char * start = pchar;

	SkipWhiteLines();

	if( *pchar!=0 && (IsOpeningCommentaryTag() || *pchar!='<') && last_new_line )
	{
		if( *pchar == '<' )
			// skipping some white characters before a opening commentary tag
			// (but only a tag in a new line)
			start = pchar;

		PutNewLine();
		PutTabs(stack_len+1);
		last_new_line = false; // in normal text we don't allow a new line character
	}


	while( *pchar != 0 )
	{
		if( !SkipCommentaryTagIfExists() )
		{
			if( *pchar == '<' )
				break;

			pchar += 1;
		}
	}

	PutTrim(start, pchar);
}


void HTMLFilter::PrintItem(const char * start, const char * end)
{
	if( LastItem().type != Item::closing )
	{
		if( last_new_line )
		{
			if( stack_len > 0 )
				PutNewLine();

			if( stack_len > 1 )
				PutTabs(stack_len);
		}

		// closing tags will be printed later
		// (when we check the stack)
	
		Put(start, pchar);
	}
}




bool HTMLFilter::ReadItem()
{
const char * start;

	if( *pchar == 0 )
		return false;

	// we have '<'
	start = pchar;
	pchar += 1;
	SkipWhite();

	if( PushStack() )
	{
		if( *pchar == '/' ) // we have a closing tag
		{	
			pchar += 1;
			SkipWhite();
			LastItem().type = Item::closing;
		}

		ReadItemName();

		if( LastItem().type != Item::closing )
			LastItem().type = (LastItem().name[0] == '!') ? Item::special : Item::opening;

		SkipItemCheckXmlSimple();
		PrintItem(start, pchar);
		CheckNewLine();
		LastItem().new_line = last_new_line;
		return true;
	}

	pchar = start;

return false;
}	


int HTMLFilter::ToLower(int c)
{
	if( c>='A' && c<='Z' )
		return c - 'A' + 'a';

return c;
}


bool HTMLFilter::IsNameEqual(const char * name1, const char * name2)
{
	for( ; *name1!=0 && *name2!=0 ; ++name1, ++name2 )
		if( ToLower(*name1) != ToLower(*name2) )
			return false;

	if( *name1==0 && *name2==0 )
		return true;

return false;
}


// len characters from both strings must be equal
bool HTMLFilter::IsNameEqual(const char * name1, const char * name2, size_t len)
{
	for( ; *name1!=0 && *name2!=0 && len>0 ; ++name1, ++name2, --len )
		if( ToLower(*name1) != ToLower(*name2) )
			return false;

	if( len == 0 )
		return true;

return false;
}

bool HTMLFilter::IsLastTag(const char * name)
{
	const char * tag = LastItem().name;

	return IsNameEqual(name, tag);
}


// checking exceptions for opening tags
void HTMLFilter::CheckExceptions()
{
	if( IsLastTag("meta")	||
		IsLastTag("input")	||
		IsLastTag("br")		||
		IsLastTag("img")		||
		IsLastTag("link")    )
	{
		LastItem().type = Item::simple;
		PopStack();
		return;
	}


	if( IsLastTag("script") || IsLastTag("pre") || IsLastTag("textarea") )
		PutLastTagWithClosingTag();
}





void HTMLFilter::AddForgottenTags()
{
int i;

	if( stack_len < 3 )
		return;

	// we have forgotten to close some tags

	// looking whether there is a matching opening tag
	for(i=int(stack_len)-3 ; i>=0 ; --i)
		if( IsNameEqual(pstack[i].name, pstack[stack_len-1].name) )
			break;

	if( i < 0 )
	{
		// oops, there is no such a tag
		// we don't print the closing and the missing opening tag
		PopStack();
		return;
	}

	for(int z=(int)stack_len-2 ; z>=i ; --z)
	{
		if( pstack[z].new_line )
		{
			PutNewLine();
			PutTabs(z+1);
		}

		PutClosingTag(pstack[z].name);
		pstack[z].Clear();
	}

	last_new_line = pstack[stack_len-1].new_line;

	// invalidate tags
	stack_len = i;
}


void HTMLFilter::CheckClosingTags()
{
	if( stack_len == 0 )
		return;

	// on the stack we have only opening tags
	// but only the last tag is a closing tag

	if( stack_len == 1 )
	{
		// there is only last closing tag
		// we dont print it
		PopStack();
		return;
	}

	// there are more than one tag 
	if( IsNameEqual(pstack[stack_len-1].name, pstack[stack_len-2].name) )
	{
		// last closing tag is from the previous one
		if( pstack[stack_len-2].new_line )
		{
			PutNewLine();
			PutTabs(stack_len-1);
		}

		PutClosingTag(pstack[stack_len-1].name);
		last_new_line = pstack[stack_len-1].new_line;
		PopStack();
		PopStack();
		return;
	}

	AddForgottenTags();
}


void HTMLFilter::PrintRest()
{
const char * start = pchar;

	while( *pchar )
		++pchar;

	if( pchar > start )
		Put(start, pchar);
}


void HTMLFilter::Read()
{
	// white characters at the beginning are skipped
	SkipWhiteLines();

	// it can be some text before the first html tag (we print it)
	ReadNormalText();

	while( *pchar && ReadItem() )
	{
		if( LastItem().type == Item::opening )
		{
			CheckExceptions();
		}
		else
		if( LastItem().type == Item::special || LastItem().type == Item::simple )
		{
			if( stack_len > 1 )
			{
				pstack[stack_len-2].new_line = LastItem().new_line;
			}
			else
			{
				// one new line after a simple or special tag
				// (if the tag has level 0 in the tree - it not means that this is a first tag)
				// for example can be DOCTYPE 
				PutNewLine(); 
			}

			PopStack();
		}
		else
		if( LastItem().type == Item::closing )
		{
			CheckClosingTags();
		}

		ReadNormalText();
	}

	// sometimes ReadItem() can return a false (when there is no space on the stack)
	// we print the rest html without filtering
	PrintRest();
}