#  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 3 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.
#
#  You should have received a copy of the GNU General Public License
#  along with this program. If not, see <http://www.gnu.org/licenses>

##################################################
#  Author:      Mostafa Vahedi                   #
#  Date:        06 July 2009                     #
#  Version:     0-11                             # 
#  Application: FarsiTeX to XePersian converter  #
##################################################

import codecs

import sys
import os

ft_numerical = [
chr(0xB9),	# 	Arabic Thoushads Seperator
chr(0xBC)	#	ARABIC DECIMAL SEPARATOR
]


ft_vowels = [
chr(0xB0),	#	ARABIC FATHA
chr(0xB1),	#	ARABIC KASRA
chr(0xB2),	#	ARABIC DAMMA
chr(0xB3),	#	ARABIC FATHATAN
chr(0xB4),	#	ARABIC SHADDA
chr(0xBA),	#	ARABIC LETTER SUPERSCRIPT ALEF
chr(0xBB),	#	ARABIC HAMZA ABOVE
chr(0xC4) 	#	ARABIC SUKUN
]

ft_non_joiners = [
chr(0x8F)	#	ARABIC LETTER HAMZA
]

ft_bidi_joiners_initial = [
chr(0xE4),	#	ARABIC LETTER AIN, initial form
chr(0xE8),	#	ARABIC LETTER GHAIN, initial form
chr(0xFB) 	#	ARABIC LETTER HEH, initial form
]

ft_bidi_joiners_medial = [
chr(0xE3),	#	ARABIC LETTER AIN, medial form
chr(0xE7),	#	ARABIC LETTER GHAIN, medial form
chr(0xFA) 	#	ARABIC LETTER HEH, medial form
]

ft_bidi_joiners_final = [
chr(0xE2),	#	ARABIC LETTER AIN, final form
chr(0xE6),	#	ARABIC LETTER GHAIN, final form
chr(0xFC) 	#	ARABIC LETTER FARSI YEH, final form
]

ft_bidi_joiners_isolated = [
chr(0xE1),	#	ARABIC LETTER AIN, isolated form
chr(0xE5),	#	ARABIC LETTER GHAIN, isolated form
chr(0xFD) 	#	ARABIC LETTER FARSI YEH, isolated form
]

ft_bidi_joiners_initial_medial = [
chr(0x8B),	#	ARABIC TATWEEL
chr(0x8E),	#	ARABIC LETTER YEH WITH HAMZA ABOVE, initial-medial form
chr(0x93),	#	ARABIC LETTER BEH, initial-medial form
chr(0x95),	#	ARABIC LETTER PEH, initial-medial form
chr(0x97),	#	ARABIC LETTER TEH, initial-medial form
chr(0x99),	#	ARABIC LETTER THEH, initial-medial form
chr(0x9B),	#	ARABIC LETTER JEEM, initial-medial form
chr(0x9D),	#	ARABIC LETTER TCHEH, initial-medial form
chr(0x9F),	#	ARABIC LETTER HAH, initial-medial form
chr(0xA1),	#	ARABIC LETTER KHAH, initial-medial form
chr(0xA8),	#	ARABIC LETTER SEEN, initial-medial form
chr(0xAA),	#	ARABIC LETTER SHEEN, initial-medial form
chr(0xAC),	#	ARABIC LETTER SAD, initial-medial form
chr(0xAE),	#	ARABIC LETTER DAD, initial-medial form
chr(0xAF),	#	ARABIC LETTER TAH, initial-medial form
chr(0xE0),	#	ARABIC LETTER ZAH, initial-medial form
chr(0xEA),	#	ARABIC LETTER FEH, initial-medial form
chr(0xEC),	#	ARABIC LETTER QAF, initial-medial form
chr(0xEE),	#	ARABIC LETTER KEHEH, initial-medial form
chr(0xF0),	#	ARABIC LETTER GAF, initial-medial form
chr(0xF3),	#	ARABIC LETTER LAM, initial-medial form
chr(0xF5),	#	ARABIC LETTER MEEM, initial-medial form
chr(0xF7),	#	ARABIC LETTER NOON, initial-medial form
chr(0xFE)	#	ARABIC LETTER FARSI YEH, initial-medial form
]

ft_bidi_joiners_final_isolated = [
chr(0x92),	#	ARABIC LETTER BEH, final-isolated form
chr(0x94),	#	ARABIC LETTER PEH, final-isolated form
chr(0x96),	#	ARABIC LETTER TEH, final-isolated form
chr(0x98),	#	ARABIC LETTER THEH, final-isolated form
chr(0x9A),	#	ARABIC LETTER JEEM, final-isolated form
chr(0x9C),	#	ARABIC LETTER TCHEH, final-isolated form
chr(0x9E),	#	ARABIC LETTER HAH, final-isolated form
chr(0xA0),	#	ARABIC LETTER KHAH, final-isolated form
chr(0xA7),	#	ARABIC LETTER SEEN, final-isolated form
chr(0xA9),	#	ARABIC LETTER SHEEN, final-isolated form
chr(0xAB),	#	ARABIC LETTER SAD, final-isolated form
chr(0xAD),	#	ARABIC LETTER DAD, final-isolated form
chr(0xC1),	#	ARABIC LETTER TAH, final-isolated form
chr(0xC2),	#	ARABIC LETTER ZAH, final-isolated form
chr(0xE9),	#	ARABIC LETTER FEH, final-isolated form
chr(0xEB),	#	ARABIC LETTER QAF, final-isolated form
chr(0xED),	#	ARABIC LETTER KEHEH, final-isolated form
chr(0xEF),	#	ARABIC LETTER GAF, final-isolated form
chr(0xF1),	#	ARABIC LETTER LAM, final-isolated form
chr(0xF4),	#	ARABIC LETTER MEEM, final-isolated form
chr(0xF6),	#	ARABIC LETTER NOON, final-isolated form
chr(0xF9) 	#	ARABIC LETTER HEH, final-isolated form
]

ft_right_joiners_final = [
chr(0x91)	#	ARABIC LETTER ALEF, final form
]

ft_right_joiners_isolated = [
chr(0x8D),	#	ARABIC LETTER ALEF WITH MADDA ABOVE, isolated form
chr(0x90)	#	ARABIC LETTER ALEF, isolated form
]

ft_right_joiners_final_isolated = [
chr(0xA2),	#	ARABIC LETTER DAL
chr(0xA3),	#	ARABIC LETTER THAL
chr(0xA4),	#	ARABIC LETTER REH
chr(0xA5),	#	ARABIC LETTER ZAIN
chr(0xA6),	#	ARABIC LETTER JEH
chr(0xBF),	#	ARABIC LETTER TEH MARBUTAH
chr(0xF2),	#	ARABIC LIGATURE LAM WITH ALEF
chr(0xF8)	#	ARABIC LETTER WAW
]


table_FT_UN = {
chr(0x80) : [u'\u06F0'],	#	EXTENDED ARABIC-INDIC DIGIT ZERO
chr(0x81) : [u'\u06F1'],	#	EXTENDED ARABIC-INDIC DIGIT ONE
chr(0x82) : [u'\u06F2'],	#	EXTENDED ARABIC-INDIC DIGIT TWO
chr(0x83) : [u'\u06F3'],	#	EXTENDED ARABIC-INDIC DIGIT THREE
chr(0x84) : [u'\u06F4'],	#	EXTENDED ARABIC-INDIC DIGIT FOUR
chr(0x85) : [u'\u06F5'],	#	EXTENDED ARABIC-INDIC DIGIT FIVE
chr(0x86) : [u'\u06F6'],	#	EXTENDED ARABIC-INDIC DIGIT SIX
chr(0x87) : [u'\u06F7'],	#	EXTENDED ARABIC-INDIC DIGIT SEVEN
chr(0x88) : [u'\u06F8'],	#	EXTENDED ARABIC-INDIC DIGIT EIGHT
chr(0x89) : [u'\u06F9'],	#	EXTENDED ARABIC-INDIC DIGIT NINE
chr(0x8A) : [u'\u060C'],	#	ARABIC COMMA
chr(0x8B) : [u'\u0640'],	#	ARABIC TATWEEL
chr(0x8C) : [u'\u061F'],	#	ARABIC QUESTION MARK
chr(0x8D) : [u'\u0622'],	#	ARABIC LETTER ALEF WITH MADDA ABOVE, isolated form
chr(0x8E) : [u'\u0626'],	#	ARABIC LETTER YEH WITH HAMZA ABOVE, initial-medial form
chr(0x8F) : [u'\u0621'],	#	ARABIC LETTER HAMZA
chr(0x90) : [u'\u0627'],	#	ARABIC LETTER ALEF, isolated form
chr(0x91) : [u'\u0627'],	#	ARABIC LETTER ALEF, final form
chr(0x92) : [u'\u0628'],	#	ARABIC LETTER BEH, final-isolated form
chr(0x93) : [u'\u0628'],	#	ARABIC LETTER BEH, initial-medial form
chr(0x94) : [u'\u067E'],	#	ARABIC LETTER PEH, final-isolated form
chr(0x95) : [u'\u067E'],	#	ARABIC LETTER PEH, initial-medial form
chr(0x96) : [u'\u062A'],	#	ARABIC LETTER TEH, final-isolated form
chr(0x97) : [u'\u062A'],	#	ARABIC LETTER TEH, initial-medial form
chr(0x98) : [u'\u062B'],	#	ARABIC LETTER THEH, final-isolated form
chr(0x99) : [u'\u062B'],	#	ARABIC LETTER THEH, initial-medial form
chr(0x9A) : [u'\u062C'],	#	ARABIC LETTER JEEM, final-isolated form
chr(0x9B) : [u'\u062C'],	#	ARABIC LETTER JEEM, initial-medial form
chr(0x9C) : [u'\u0686'],	#	ARABIC LETTER TCHEH, final-isolated form
chr(0x9D) : [u'\u0686'],	#	ARABIC LETTER TCHEH, initial-medial form
chr(0x9E) : [u'\u062D'],	#	ARABIC LETTER HAH, final-isolated form
chr(0x9F) : [u'\u062D'],	#	ARABIC LETTER HAH, initial-medial form
chr(0xA0) : [u'\u062E'],	#	ARABIC LETTER KHAH, final-isolated form
chr(0xA1) : [u'\u062E'],	#	ARABIC LETTER KHAH, initial-medial form
chr(0xA2) : [u'\u062F'],	#	ARABIC LETTER DAL
chr(0xA3) : [u'\u0630'],	#	ARABIC LETTER THAL
chr(0xA4) : [u'\u0631'],	#	ARABIC LETTER REH
chr(0xA5) : [u'\u0632'],	#	ARABIC LETTER ZAIN
chr(0xA6) : [u'\u0698'],	#	ARABIC LETTER JEH
chr(0xA7) : [u'\u0633'],	#	ARABIC LETTER SEEN, final-isolated form
chr(0xA8) : [u'\u0633'],	#	ARABIC LETTER SEEN, initial-medial form
chr(0xA9) : [u'\u0634'],	#	ARABIC LETTER SHEEN, final-isolated form
chr(0xAA) : [u'\u0634'],	#	ARABIC LETTER SHEEN, initial-medial form
chr(0xAB) : [u'\u0635'],	#	ARABIC LETTER SAD, final-isolated form
chr(0xAC) : [u'\u0635'],	#	ARABIC LETTER SAD, initial-medial form
chr(0xAD) : [u'\u0636'],	#	ARABIC LETTER DAD, final-isolated form
chr(0xAE) : [u'\u0636'],	#	ARABIC LETTER DAD, initial-medial form
chr(0xAF) : [u'\u0637'],	#	ARABIC LETTER TAH, initial-medial form
chr(0xB0) : [u'\u064E'],	#	ARABIC FATHA
chr(0xB1) : [u'\u0650'],	#	ARABIC KASRA
chr(0xB2) : [u'\u064F'],	#	ARABIC DAMMA
chr(0xB3) : [u'\u064B'],	#	ARABIC FATHATAN
chr(0xB4) : [u'\u0651'],	#	ARABIC SHADDA
chr(0xB5) : [u'\u0023'],	# * #
chr(0xB6) : [u'\u0024'],	# * $
chr(0xB7) : [u'\u0025'],	# * %
chr(0xB8) : [u'\u0026'],	# * &
chr(0xB9) : [u'\u066C'],	# 	Arabic Thoushads Seperator
chr(0xBA) : [u'\u0670'],	#	ARABIC LETTER SUPERSCRIPT ALEF
chr(0xBB) : [u'\u0654'],	#	ARABIC HAMZA ABOVE
chr(0xBC) : [u'\u066B'],	#	ARABIC DECIMAL SEPARATOR
chr(0xBD) : [u'\u0029'],	# * RIGHT PARENTHESIS
chr(0xBE) : [u'\u0028'],	# * LEFT PARENTHESIS
chr(0xBF) : [u'\u0629'],	#	ARABIC LETTER TEH MARBUTAH
chr(0xC0) : [u'\u00BB'],	#	RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK
chr(0xC1) : [u'\u0637'],	#	ARABIC LETTER TAH, final-isolated form
chr(0xC2) : [u'\u0638'],	#	ARABIC LETTER ZAH, final-isolated form
chr(0xC3) : [u'\u00AB'],	#	LEFT-POINTING DOUBLE ANGLE QUOTATION MARK
chr(0xC4) : [u'\u0652'],	#	ARABIC SUKUN
chr(0xC5) : [u'\u002D'],	# * -
chr(0xC6) : [u'\u002E'],	# * FULL STOP
chr(0xC7) : [u'\u002F'],	# * /
chr(0xC8) : [u'\u002A'],	# * *
chr(0xC9) : [u'\u007E'],	# * ~
chr(0xCA) : [u'\u003A'],	# * COLON
chr(0xCB) : [u'\u061B'],	# 	ARABIC SEMICOLON
chr(0xCC) : [u'\u003E'],	# * GREATER-THAN SIGN
chr(0xCD) : [u'\u002B'],	# * +
chr(0xCE) : [u'\u003D'],	# * =
chr(0xCF) : [u'\u003C'],	# * LESS-THAN SIGN
# chr(0xD0) : [u'\u0040'],	# * @
chr(0xD0) : [u''],	        # * @
chr(0xD1) : [u'\u005D'],	# * [
chr(0xD2) : [u'\u005C'],	# * \
chr(0xD3) : [u'\u005B'],	# * ]
chr(0xD4) : [u'\u005E'],	# * ^
chr(0xD5) : [u'\u005F'],	# * _
chr(0xD6) : [u'\u0060'],	# * `
chr(0xD7) : [u'\u007D'],	# * {
chr(0xD8) : [u'\u007C'],	# * |
chr(0xDA) : [u'\u0020'],	# * SPACE
chr(0xDD) : [u'\u0021'],	# * EXCLAMATION MARK
chr(0xDE) : [u'\u007B'],	# * }
chr(0xE0) : [u'\u0638'],	#	ARABIC LETTER ZAH, initial-medial form
chr(0xE1) : [u'\u0639'],	#	ARABIC LETTER AIN, isolated form
chr(0xE2) : [u'\u0639'],	#	ARABIC LETTER AIN, final form
chr(0xE3) : [u'\u0639'],	#	ARABIC LETTER AIN, medial form
chr(0xE4) : [u'\u0639'],	#	ARABIC LETTER AIN, initial form
chr(0xE5) : [u'\u063A'],	#	ARABIC LETTER GHAIN, isolated form
chr(0xE6) : [u'\u063A'],	#	ARABIC LETTER GHAIN, final form
chr(0xE7) : [u'\u063A'],	#	ARABIC LETTER GHAIN, medial form
chr(0xE8) : [u'\u063A'],	#	ARABIC LETTER GHAIN, initial form
chr(0xE9) : [u'\u0641'],	#	ARABIC LETTER FEH, final-isolated form
chr(0xEA) : [u'\u0641'],	#	ARABIC LETTER FEH, initial-medial form
chr(0xEB) : [u'\u0642'],	#	ARABIC LETTER QAF, final-isolated form
chr(0xEC) : [u'\u0642'],	#	ARABIC LETTER QAF, initial-medial form
chr(0xED) : [u'\u06A9'],	#	ARABIC LETTER KEHEH, final-isolated form
chr(0xEE) : [u'\u06A9'],	#	ARABIC LETTER KEHEH, initial-medial form
chr(0xEF) : [u'\u06AF'],	#	ARABIC LETTER GAF, final-isolated form
chr(0xF0) : [u'\u06AF'],	#	ARABIC LETTER GAF, initial-medial form
chr(0xF1) : [u'\u0644'],	#	ARABIC LETTER LAM, final-isolated form
chr(0xF2) : [u'\u0644\u0627'],	#	ARABIC LIGATURE LAM WITH ALEF
chr(0xF3) : [u'\u0644'],	#	ARABIC LETTER LAM, initial-medial form
chr(0xF4) : [u'\u0645'],	#	ARABIC LETTER MEEM, final-isolated form
chr(0xF5) : [u'\u0645'],	#	ARABIC LETTER MEEM, initial-medial form
chr(0xF6) : [u'\u0646'],	#	ARABIC LETTER NOON, final-isolated form
chr(0xF7) : [u'\u0646'],	#	ARABIC LETTER NOON, initial-medial form
chr(0xF8) : [u'\u0648'],	#	ARABIC LETTER WAW
chr(0xF9) : [u'\u0647'],	#	ARABIC LETTER HEH, final-isolated form
chr(0xFA) : [u'\u0647'],	#	ARABIC LETTER HEH, medial form
chr(0xFB) : [u'\u0647'],	#	ARABIC LETTER HEH, initial form
chr(0xFC) : [u'\u06CC'],	#	ARABIC LETTER FARSI YEH, final form
chr(0xFD) : [u'\u06CC'],	#	ARABIC LETTER FARSI YEH, isolated form
chr(0xFE) : [u'\u06CC']		#	ARABIC LETTER FARSI YEH, initial-medial form
}

F_PERCENT_SIGN = chr(0xB7)
F_AT_SIGN = chr(0xD0)
F_SLASH = chr(0xD2)
F_SPACE = chr(0xDA)
F_PRNT_OPEN = chr(0xDE)
F_PRNT_CLOSE = chr(0xD7)

# latex and farsitex commands whose first parameter does not need \lr{...}
commands = [ 
"begin", "end",
"input", "include", "includeonly",
"hspace", "vspace", "hspace*", "vspace*",
"label", "ref", "cite", "bibitem",
"bibliographystyle",
"parbox",
"newenvironment", "newtheorem",
"persianmathdigitsfamily",
"fontfamily", "fontseries", "fontshape",
"rmdefault", "sfdefault", "ttdefault",
"bfdefault", "itdefault", "sldefault", "scdefault",
"pagenumbering", "pagestyle", "thispagestyle",
"setcounter", "stepcounter", "setlength", "addtolength"
]

def ft_is_all_persian_space(next_part):
	l = len(next_part)
	i = 0
	while (i < l):
		if (next_part[i] != F_SPACE):
			return 0
		i += 1
	return 1

def ft_is_numeric(ch):
	if ((ch in ft_numerical) or 
	    ((ch >= chr(0x80)) and (ch <= chr(0x89))) ):
		return 1
	return 0

def ft_can_join_left(ch):
	if ((ch in ft_bidi_joiners_initial) or
	    (ch in ft_bidi_joiners_medial) or
	    (ch in ft_bidi_joiners_final) or
	    (ch in ft_bidi_joiners_isolated) or
	    (ch in ft_bidi_joiners_initial_medial) or
	    (ch in ft_bidi_joiners_final_isolated)):
    		return 1
	return 0

def ft_can_join_right(ch):
	if (ft_can_join_left(ch) or 
	    (ch in ft_right_joiners_final) or
	    (ch in ft_right_joiners_isolated) or
	    (ch in ft_right_joiners_final_isolated)):
		return 1
	return 0

def ft_joining_left(ch):
	if ((ch in ft_bidi_joiners_initial) or 
	    (ch in ft_bidi_joiners_medial) or
	    (ch in ft_bidi_joiners_initial_medial)):
		return 1
	return 0


def ft_joining_right(ch):
	if ((ch in ft_right_joiners_final) or
	    (ch in ft_bidi_joiners_medial) or 
	    (ch in ft_bidi_joiners_final)):
		return 1
	return 0

def ft_not_right_joined(ch):
	if ((ch in ft_bidi_joiners_initial) or
	    (ch in ft_right_joiners_isolated) or
	    (ch in ft_bidi_joiners_isolated)):
		return 1
	return 0

def ft_adjust_shaping(text, i):
	current = text[i]
	u = u''
	try:
		u = table_FT_UN[current][0]
	except KeyError:
		return u''

	#if you don't want shaping remove the following comment
	#return u

	if ((current in ft_vowels) or (ft_is_numeric(current))):
		return u

	#find next non-vowel character on the left
	text_len = len(text)
	next_index = i+1
	while ((next_index < text_len) and (text[next_index] in ft_vowels)):
		next_index += 1

	if (next_index == text_len):
		next = ''
	else:
		next = text[next_index]

	# if current letter is joining from left but next letter is or can not joining
	if (ft_joining_left(current)):
		if (not ft_can_join_right(next)):
			u += u'\u200D' #ZWJ
		elif (ft_not_right_joined(next)):
			u += u'\u200D\u200C' #ZWJ+ZWNJ
	# if current letter can join but next letter is joining from right
	elif (ft_can_join_left(current)):
		if (ft_joining_right(next)):
			u += u'\u200C\u200D' #ZWNJ+ZWJ
		elif (ft_can_join_right(next)):
			u += u'\u200C' #ZWNJ
	return u

def ft_adjust_number(text):
	result = u''
	i = len(text)-1
	while (i >= 0):
		result += ft_adjust_shaping(text, i)
		i -= 1
	return result


def map_ft_unicode(text):
	mapped_text = u''

	i = 0
	while (i < len(text)):
		if (ft_is_numeric(text[i])):
			next_index = i
			while ((next_index+1 < len(text)) and
			       (ft_is_numeric(text[next_index+1]))):
				next_index += 1
			mapped_text += ft_adjust_number(text[i:next_index+1])
			i = next_index+1
			continue

		mapped_text += ft_adjust_shaping(text, i)
		i += 1
	return mapped_text

# Finds next token all of the same language
def ft_next_part(line, i, line_len):
	global global_state
	global recursive
	global filenames
	
	j = i
	language_flag = (line[j]<chr(0x80))
	while ((j<line_len) and ((line[j]<chr(0x80)) == language_flag) ):
		if ( (global_state == 0) and ( (line[j] == '%') or (line[j] == F_PERCENT_SIGN) ) and (line[j-1] != '\\') and (line[j-1] != F_SLASH) ):
			global_state = 1
		elif ((global_state == 0) and ((line[j:j+16] == '\\begin{verbatim}') or (line[j:j+17] == '\\begin{verbatim*}'))):
			global_state = 2
		elif ((global_state == 2) and ((line[j:j+14] == '\\end{verbatim}') or (line[j:j+15] == '\\end{verbatim*}'))):
			global_state = 0
		elif ((global_state == 0) and (line[j:j+6] == '\\verb*')):
			next_index = line.find(line[j+6], j+7)
			j = next_index
		elif ((global_state == 0) and (line[j:j+5] == '\\verb')):
			next_index = line.find(line[j+5], j+6)
			j = next_index
		elif (recursive == 1):
			if ((global_state == 0) and (line[j:j+9] == '\\include{')):
				next_index = line.find('}', j+9)
				filename = line[j+10:next_index-1] + '.ftx'
				if (os.path.exists(filename) and not filename in filenames):
					filenames.append(filename)
			elif ((global_state == 0) and (line[j:j+7] == '\\input{')):
				next_index = line.find('}', j+7)
				filename = line[j+8:next_index-1] + '.ftx'
				if (os.path.exists(filename) and not filename in filenames):
					filenames.append(filename)
			elif ((global_state == 0) and (line[j:j+9] == F_SLASH + 'include' + F_PRNT_OPEN)):
				next_index = line.find(F_PRNT_CLOSE, j+9)
				if (line[j+9]!=F_AT_SIGN):
					filename = line[j+9:next_index] + '.ftx'
				else:
					filename = line[j+10:next_index-1] + '.ftx'
				if (os.path.exists(filename) and not filename in filenames):
					filenames.append(filename)
			elif ((global_state == 0) and (line[j:j+7] == F_SLASH + 'input' + F_PRNT_OPEN)):
				next_index = line.find(F_PRNT_CLOSE, j+7)
				if (line[j+7]!=F_AT_SIGN):
					filename = line[j+7:next_index] + '.ftx'
				else:
					filename = line[j+8:next_index-1] + '.ftx'
				if (os.path.exists(filename) and not filename in filenames):
					filenames.append(filename)
		j += 1
	return j

###############################################
# from here all functions are used to translate
# farsitex commands to xepersian commands
###############################################

def read_size(input,index,last_index):
	dim_index = -1 
	inch_index = input.find(u'in', index)
	if (inch_index != -1):
		dim_index = inch_index
	mm_index = input.find(u'mm', index)
	if (mm_index != -1):
		if (dim_index == -1 or mm_index < index):
			dim_index = mm_index
	cm_index = input.find(u'cm', index)
	if (cm_index != -1):
		if (dim_index == -1 or cm_index < dim_index):
			dim_index = cm_index
	pt_index = input.find(u'pt', index)
	if (pt_index != -1):
		if (dim_index == -1 or pt_index < dim_index):
			dim_index = pt_index
	next_cmd = input.find(u'\\', index)
	if (next_cmd == -1 and dim_index == -1):
		print "Error in parsing \epsfxsize command at " + str(line_number) + "\n"
		return -1
	elif (next_cmd == -1 or (dim_index != -1 and next_cmd > dim_index)):
		epsfxsize = input[index:dim_index+2]
		return dim_index+2
	elif (next_cmd != -1):
		end_cmd = next_cmd+1
		while (end_cmd < last_index and input[end_cmd].isalpha()):
			end_cmd += 1
		return end_cmd
	else:
		print "Error in parsing \epsfxsize command at " + str(line_number) + "\n"
		return -1


latex_options=[u'a4paper', u'a5paper', u'b5paper', 
			   u'letterpaper', u'legalpaper', u'executivepaper', 
			   u'landscape', u'10pt', u'11pt', u'12pt', 
			   u'oneside', u'twoside', u'draft', 
			   u'final', u'titlepage', u'notitlepage',
			   u'onecolumn', u'twocolumn',
			   u'leqno', u'fleqn']

table_eq_packages = {
u'poem'     : u'persianpoem',
u'fmakeidx' : u'makeidx',
u'ffancyhe' : u'fancyhdr',
u'fmultico' : u'multicol'
}

farsitex_ignore_options = [u'persian', u'farsi', u'epsf', u'fgraphix', u'lotusfont']

xepersian_packages = u'\n\\usepackage{url}\n%\\usepackage{fancyvrb}\n\\usepackage{setspace}\\doublespacing\n\\usepackage{graphicx} \n\\usepackage{amssymb}\n'

xepersian_preamble = u'\\settextfont[Scale=1.2]{XB Zar}\n\\setlatintextfont[Scale=1.1]{Times New Roman}\n\\setdigitfont{XB Zar}\n\\setpookfont{XB Kayhan Pook}\n\\setsayehfont{XB Kayhan Sayeh}\n\\defpersianfont\\lotoos[Scale=1.2]{XB Roya}\n\\defpersianfont\elmi[Scale=1.2]{XB Roya} \n\\def\\nazok{\\normalfont\\normalsize}\n\\let\\iranic\\it\n\\let\\siah\\bf\n\\let\\khabide\\sl\n\\def\\siahir{\\siah\\iranic} \n\\def\\siahkh{\\siah\\khabide}\n\\let\\tookhali\\pookfamily\n\\let\\sayedar\\sayehfamily\n\\let\\farsi\\Persian\n\\let\\english\\Latin \n\\let\\farmbox\\mbox\n\\newcommand{\\ftxepmatrix}[1]{\\begin{pmatrix}#1\\end{pmatrix}}\n\\newcommand{\\ftxematrix}[1]{\\begin{matrix}#1\\end{matrix}} \n\\def\\FarsiTeX{\\lr{FarsiTeX}}\n\\def\\فارسی‌تک{\\rl{فارسی‌‌تک}}\n\\def\\InE{\\begingroup\\beginL\\latinfont}\n\\def\\EnE{\\endL\\endgroup} \n\\def\\InF{\\begingroup\\beginR\\persianfont}\n\\def\\EnF{\\endR\\endgroup}\n\\newcommand{\\IE}[1]{\\lr{#1}}\n\\newcommand{\\IP}[1]{\\rl{#1}} \n\\newcommand{\\IF}[1]{\\rl{#1}}\n\\def\\persiandash{\\rl{-}} \n\\def\\DeclareRobustBiCommand#1#2#3#4{\\DeclareRobustCommand#1{\\if@rl{}#4\\else{}#3\\fi}\\let#2=#1} \n\\catcode`\\﷼=3\n\\catcode`‌=11\n\\newcommand\\dotsectionseparator{\\SepMark{.}}\n\\newcommand\\dashsectionseparator{\\SepMark{-}}\n'

thesis_preamble = u'\\university{\\lr{UniversityName}}\n\\city{\\lr{CityName}}\n\\latinuniversity{UniversityName}\n\\latincity{CityName} \n\\let\\englishtitle\\latintitle\n\\let\\englishauthor\\latinauthor\n\\let\\englishdegree\\latindegree\n\\let\\englishthesisdate\\latinthesisdate \n\\let\\englishsupervisor\\latinsupervisor\n\\let\\englishdepartment\\latindepartment\n\\let\\makeenglishtitle\\makelatintitle \n\\let\\englishkeywords\\latinkeywords\n\\newenvironment{englishabstract}{\\begin{latinabstract}}{\\end{latinabstract}}\n'

def is_alpha_numeric_space(input):
	input_len = len(input)
	i = 0
	while (i<input_len):
		if (not (input[i].isalpha() or input[i].isdigit() or input[i]==u'.' or input[i]==u' ') ):
			return 0
		i+=1
	return 1

def is_alpha_numeric(input):
	input_len = len(input)
	i = 0
	while (i<input_len):
		if (not (input[i].isalpha() or input[i].isdigit() or input[i]=='.') ):
			return 0
		i+=1
	return 1
			
def find_eq_cmd(keyword):
	try:
		eq_cmd = table_eq_packages[keyword]
	except KeyError:
		eq_cmd = u''
	return eq_cmd

def convert_persian_to_english(num):
	result = u''
	num_len = len(num)
	index = 0
	while (index < num_len):
		if (ord(num[index]) >= ord(u'۰') and ord(num[index]) <= ord(u'۹')):
			result += unichr(ord(num[index]) - ord(u'۰') + ord(u'0'))
		else:
			result += num[index]
		index += 1
	return result

def generate_farsitex_cmds_file(helper_filename,preamble):
	try:
		of = codecs.open(helper_filename, encoding='utf-8', mode='w')
	except IOError:
		print "Can not open the output file: " + helper_filename
		exit(0)
	of.write(preamble)
	of.close


# \verb|| -> \item[\verb||] or \section{\verb||}
# \kasre{} \alef{} ...
def translate_cmds(output_line):
	global last_epsfxsize
	global last_epsfxsize_line
	global last_epsfysize
	global last_epsfysize_line
	global state
	global filename

	result = u''
	line_len = len(output_line)
	index = 0
	if (state == 1): #\begin{verbatim}
		end_verbatim = output_line.find('\\end{verbatim}')
		if (end_verbatim == -1):
			return output_line
		result += output_line[0:end_verbatim+14]
		index = end_verbatim+14
		state = 0
	elif (state == 2): #\begin{verbatim*}
		end_verbatim = output_line.find('\\end{verbatim*}')
		if (end_verbatim == -1):
			return output_line
		result += output_line[0:end_verbatim+15]
		index = end_verbatim+15
		state = 0
	elif (output_line[0:14] == "\\documentstyle"):
		result += u'\\documentclass'
		#process options
		last_index = output_line.find(u']')
		index = 15
		first_option = 1
		preamble = xepersian_preamble
		packages = xepersian_packages
		xe_document_class = u''
		while (index < last_index):
			next_comma = output_line.find(u',',index,last_index)
			if (next_comma == -1):
				next_comma = last_index
			first_of_option = index
			while (output_line[first_of_option] == u' '):
				first_of_option += 1
			end_of_option = next_comma
			while (output_line[end_of_option-1] == u' '):
				end_of_option -= 1
			option = output_line[first_of_option:end_of_option]
			index = next_comma+1
			eq_cmd = find_eq_cmd(option)
			if (eq_cmd != u''):
				packages += u'\\usepackage{' + eq_cmd + u'}\n'
				continue
			elif (option in farsitex_ignore_options):
				continue
			elif (option == u'sharifth'):
				xe_document_class = u'xepersian-thesis'
				preamble += thesis_preamble
				continue
			elif (not option in latex_options):
				packages += u'\\usepackage{' + option + u'}\n'
				continue
	
			if (first_option):
				result += u'['
			else:
				result += u','
			result += option
			first_option = 0
		#end while
		if (not first_option):
			result += u']'
		# process document style into document class
		index = output_line.find(u'}',last_index)
		document_class = output_line[last_index+2:index]
		if (xe_document_class != u''):
			result += u'{' + xe_document_class + u'}'
		elif (document_class == u'oldarticle'):
			result += u'{article}'
		elif (document_class == u'oldbook'):
			result += u'{book}'
		elif (document_class == u'oldreport'):
			result += u'{report}'
		else:
			result += u'{' + document_class + u'}'
		# I assume that nothing important is after
		# \documentstyle[...]{...}, otherwise it may conflict
		# with our preamble
		if (index != -1):
			result += output_line[index+1:]
		result += packages + u'\\usepackage{xepersian}\n'
		helper_filename = filename + '_farsitex_cmds_xepersian.tex'
		generate_farsitex_cmds_file(helper_filename,preamble)
		result += '\\input{' + helper_filename + '}\n'
		return result
	#end elif "documentstyle"

	while (index < line_len):
		next_index = output_line.find(u'\\', index)
		comment_index = output_line.find(u'%', index)
		if (next_index == -1):
			result += output_line[index:]
			break
		elif (state == 1):
			if (output_line[next_index:next_index+14] == u'\\end{verbatim}'):
				result += output_line[index:next_index+14]
				index = next_index+14
				state = 0
			else:
				result += output_line[index:next_index+1]
				index = next_index+1
		elif (state == 2):
			if (output_line[next_index:next_index+15] == u'\\end{verbatim*}'):
				result += output_line[index:next_index+15]
				index = next_index+15
				state = 0
			else:
				result += output_line[index:next_index+1]
				index = next_index+1
		elif (comment_index != -1 and comment_index < next_index):
			result += output_line[index:]
			break
		elif (output_line[next_index:next_index+14] == u"\\input{amssym}"):
			result += u'\\usepackage{amssymb}'
			index = next_index+14
		elif (output_line[next_index:next_index+12] == u"\\input{epsf}"):
			index = next_index+12
		elif (output_line[next_index:next_index+15] == u"\\includeepspdf{"):
			result += u'\\includegraphics{'
			index = next_index+15
		elif (output_line[next_index:next_index+16] == u"\\begin{verbatim}"):
			result += output_line[index:next_index+16]
			index = next_index+16
			state = 1
		elif (output_line[next_index:next_index+17] == u"\\begin{verbatim*}"):
			result += output_line[index:next_index+17]
			index = next_index+17
			state = 2
		elif (output_line[next_index:next_index+26] == u'\\setlength{\\headrulewidth}'):
			end_cmd = output_line.find('}', next_index+26)
			result += u'\\renewcommand{\\headrulewidth}' + output_line[next_index+26:end_cmd+1]
			index = end_cmd+1
		elif (output_line[next_index:next_index+26] == u'\\setlength{\\footrulewidth}'):
			end_cmd = output_line.find('}', next_index+26)
			result += u'\\renewcommand{\\footrulewidth}' + output_line[next_index+26:end_cmd+1]
			index = end_cmd+1
		elif (output_line[next_index:next_index+10] == u"\\epsffile{"):
			result += output_line[index:next_index]
			result += u'\\includegraphics'
			size_options = u''
			if (line_number - last_epsfxsize_line <= 3):
				size_options = u'width=' + last_epsfxsize
			if (line_number - last_epsfysize_line <= 3):
				if (size_options != u''):
					size_options += u','
				size_options += u'height=' + last_epsfysize
			if (size_options != u''):
				result += u'[' + size_options + u']'
			end_prn = output_line.find(u'.eps}', next_index+9)

			if (end_prn != -1):
				result += output_line[next_index+9:end_prn] + u'}'
				index = end_prn+5
			else:
				end_prn = output_line.find(u'.ps}', next_index+9)
				if (end_prn != -1):
					result += output_line[next_index+9:end_prn] + u'}'
					index = end_prn+4
				else:
					end_prn = output_line.find(u'}', next_index+9)
					result += output_line[next_index+9:end_prn+1]
					index = end_prn+1
		# I assume all the parameter of \epsfxsize comes in one line
		elif (output_line[next_index:next_index+11] == u"\\epsfxsize="):
			end_size = read_size(output_line, next_index+11, line_len)
			if (end_size != -1):
				last_epsfxsize = output_line[next_index+11:end_size]
				index = end_size
			else:
				index = next_index+11
			last_epsfxsize_line = line_number
		# I assume all the parameter of \epsfysize comes in one line
		elif (output_line[next_index:next_index+11] == u"\\epsfysize="):
			end_size = read_size(output_line, next_index+11, line_len)
			if (end_size != -1):
				last_epsfysize = output_line[next_index+11:end_size]
				index = end_size
			else:
				index = next_index+11
			last_epsfysize_line = line_number
		elif (output_line[next_index:next_index+10] == u"\\LR{\\verb*"):
			end_verb = output_line.find(output_line[next_index+10], next_index+11)
			verb_param = output_line[next_index+11:end_verb]
			if (is_alpha_numeric(verb_param)):
				result += output_line[index:next_index]
				result += u'\\lr{\\tt{}' + verb_param
			else:
				result += output_line[index:end_verb+1]
			index = end_verb+1
		elif (output_line[next_index:next_index+9] == u"\\LR{\\verb"):
			end_verb = output_line.find(output_line[next_index+9], next_index+10)
			verb_param = output_line[next_index+10:end_verb]
			if (is_alpha_numeric_space(verb_param)):
				result += output_line[index:next_index]
				result += u'\\lr{\\tt{}' + verb_param
			else:
				result += output_line[index:end_verb+1]
			index = end_verb+1
		elif (output_line[next_index:next_index+6] == u"\\verb*"):
			end_verb = output_line.find(output_line[next_index+6], next_index+7)
			result += output_line[index:end_verb+1]
			index = end_verb+1
		elif (output_line[next_index:next_index+5] == u"\\verb"):
			end_verb = output_line.find(output_line[next_index+5], next_index+6)
			result += output_line[index:end_verb+1]
			index = end_verb+1
		elif (output_line[next_index:next_index+9] == u"\\pmatrix{"):
			result += u'\\ftxepmatrix{'
			index = next_index+9
		elif (output_line[next_index:next_index+8] == u"\\matrix{"):
			result += u'\\ftxematrix{'
			index = next_index+8
		elif (output_line[next_index:next_index+16] == u"\\begin{document}"):
			result += u'\\begin{document}\n%\\VerbatimFootnotes'
			index = next_index+16
		elif (output_line[next_index:next_index+8] == u'\\label {'):
			begin_param = next_index+8
			end_param = output_line.find(u'}', begin_param)
			param = convert_persian_to_english(output_line[begin_param:end_param])
			result += output_line[index:begin_param-2]
			result += u'{' + param + u'}'
			index = end_param+1
		elif (output_line[next_index:next_index+6] == u'\\ref {'):
			begin_param = next_index+6
			end_param = output_line.find(u'}', begin_param)
			param = convert_persian_to_english(output_line[begin_param:end_param])
			result += output_line[index:begin_param-2]
			result += u'{' + param + u'}'
			index = end_param+1
		elif (output_line[next_index:next_index+7] == u'\\label{'):
			begin_param = next_index+7
			end_param = output_line.find(u'}', begin_param)
			param = convert_persian_to_english(output_line[begin_param:end_param])
			result += output_line[index:begin_param]
			result += param + u'}'
			index = end_param+1
		elif (output_line[next_index:next_index+5] == u'\\ref{'):
			begin_param = next_index+5
			end_param = output_line.find(u'}', begin_param)
			param = convert_persian_to_english(output_line[begin_param:end_param])
			result += output_line[index:begin_param]
			result += param + u'}'
			index = end_param+1
		elif (output_line[next_index:next_index+13] == u'\\multicolumn{'):
			begin_param = next_index+13
			end_param = output_line.find(u'}', begin_param)
			param = convert_persian_to_english(output_line[begin_param:end_param])
			result += output_line[index:begin_param]
			result += param + u'}'
			index = end_param+1
		elif (output_line[next_index:next_index+12] == u'\\setcounter{'):
			begin_num = output_line.find(u'{', next_index+12)
			end_num = output_line.find(u'}', begin_num)
			num = convert_persian_to_english(output_line[begin_num+1:end_num])
			result += output_line[index:begin_num+1]
			result += num + u'}'
			index = end_num+1
		elif (output_line[next_index:next_index+14] == u'\\addtocounter{'):
			begin_num = output_line.find(u'{', next_index+14)
			end_num = output_line.find(u'}', begin_num)
			num = convert_persian_to_english(output_line[begin_num+1:end_num])
			result += output_line[index:begin_num+1]
			result += num + u'}'
			index = end_num+1
		else:
			result += output_line[index:next_index+2]
			index = next_index+2
	#end while
	return result

###############################################
# from here all functions are mainly used to
# convert farsitex format to unicode
###############################################

def convert_file(f, of, convert_cmds):
	global state
	global line_number
	global last_epsfysize_line
	global last_epsfxsize_line
	global last_epsfxsize
	global last_epsfysize
	global global_state

	state = 0
	line_number = 0
	last_epsfysize_line = 0
	last_epsfxsize_line = 0
	last_epsfxsize = u''
	last_epsfysize = u''

	for line in f:
		line_number += 1
		print line_number,
		output_line = u''
		line_len = len(line)
		
		# remove new-line characters from end of line
		if (line_len>1 and line[line_len-1] == '\n'):
			line_len-=1
		if (line_len>1 and line[line_len-1] == '\r'):
			line_len-=1
		
		# check line-direction character
		line_direction_rtl = (line[0] == '<')
		if (line[0] != '>') and (line[0] != '<'):
			print "FORMAT ERROR AT LINE: " + str(line_number)
			exit(0)
	
		i = 1
	
		while (i<line_len):
			next_part_index = ft_next_part(line, i, line_len)
			next_part = line[i:next_part_index]
			next_part_latin = (line[i]<chr(0x80))
			
			# see if we should put \lr{...} for the current english expression
			if (global_state == 0):
				if line_direction_rtl and next_part_latin:
					is_command_rtl = (next_part_latin and (i>1) and (line[i-1]==F_SLASH) )
					is_parameter_rtl = (next_part_latin and (i>1) and (next_part_index<line_len) and (line[i-1]==F_AT_SIGN) and (line[next_part_index]==F_AT_SIGN) )
					is_math_rtl = (next_part_latin and (line[i]=="$") and (line[next_part_index-1]=="$") )
					is_verb_parameter = ( (line[i-6:i] == F_SLASH+"verb") and not isalpha(line[i]))
					is_verb = (next_part_latin and (line[i:i+5]=='\\verb' or line[i:i+6]==' \\verb'))
					is_english = (next_part_latin and (line[i:i+8]=='\\english'))
				
					cmd_index = 0
					while cmd_index < len(commands):
						len_cmd = len(commands[cmd_index])+2
						if ( (i > len_cmd) and (line[i-len_cmd:i] == F_SLASH+commands[cmd_index]+F_PRNT_OPEN) ):
							break
						elif ( (i > len_cmd+1) and (line[i-len_cmd-1:i] == F_SLASH+commands[cmd_index]+F_SPACE+F_PRNT_OPEN) ):
							break
						cmd_index += 1
					is_commands_group = cmd_index < len(commands)
					is_documentstyle_cmd = (line_len > 15) and (line[1:15] == F_SLASH+"documentstyle")
	
			if next_part_latin:
				if (global_state == 0):
					# whether we should put a \lr{ command
					if ( line_direction_rtl and not (is_command_rtl or is_parameter_rtl or is_math_rtl or is_commands_group or is_documentstyle_cmd or is_verb_parameter or is_verb or is_english) ):
						output_line += u'\\lr{'
					if ( line_direction_rtl and is_verb):
						output_line += u'\\LR{'
				
				# here is the main place that converting happens
				output_line += next_part.encode( 'utf-8' )
				
				if (global_state == 0):
					# check whether we already used a \lr command: then end it
					if ( line_direction_rtl and not (is_command_rtl or is_parameter_rtl or is_math_rtl or is_commands_group or is_documentstyle_cmd or is_verb_parameter or is_verb or is_english) ):
						output_line += u'}'
					if ( line_direction_rtl and is_verb):
						output_line += u'}'
			else:
				if (global_state == 0):
					# whether we should put a \rl{} command
					if ( not line_direction_rtl and not ft_is_all_persian_space(next_part)):
						output_line += u'\\rl{'
					
				# here is the main place that converting happens
				output_line += map_ft_unicode(next_part)
				
				if (global_state == 0):
					# check whether we already used a \rl command: then end it
					if (not line_direction_rtl and not ft_is_all_persian_space(next_part)):
						output_line += u'}'
					
			i = next_part_index
		# end of while			

		#if there was a % commenting then we can return to normal situation
		if (global_state == 1):
			global_state = 0
		
		# convert some of the FarsiTeX commands to XePersian commands
		# only if it is requested
		if (convert_cmds):
			result = translate_cmds(output_line) 
		else:
			result = output_line
		output_line = result + u'\n'
		# write the processed line
		of.write(output_line)
		# end of line processing
	# end of file processing

def print_usage():
	print 'usage: python ftxe-0-11 [-r] [-s] [-x] [-u] in_filename1 in_filename2'
	print '-r: (DEFAULT) recursively consider files included in the given files'
	print '-s: do not recursively consider files'
	print '-x: (DEFAULT) insert xepersian related commands'
	print '-u: only convert to unicode'

###################################
# Begin of main body of the program
###################################

# global variables
line_number = 0
last_epsfxsize = u''
last_epsfxsize_line = 0
last_epsfysize = u''
last_epsfysize_line = 0
state = 0
global_state = 0
recursive = 1
convert_xepersian = 1
filename = ''

if len(sys.argv) <= 1:
	print_usage()
	exit(0)

#find options
options_index = 1
while (options_index < len(sys.argv) and sys.argv[options_index][0]=='-'):
	if (sys.argv[options_index]=='-s'):
		recursive = 0
	elif (sys.argv[options_index]=='-u'):
		convert_xepersian = 0
	options_index += 1

filenames = []
while (options_index < len(sys.argv)):
	filenames.append(sys.argv[options_index])
	options_index += 1

if (len(filenames) == 0):
	print 'error: no input filename is specified!'
	print_usage()
	exit(0)
	
index = 0
while (index < len(filenames)):
	filename = filenames[index]
	index += 1

	outfile = ''
	if (filename[-4:] != '.tex'):
		outfile = filename[0:-3] + 'tex'
	else: 
		outfile = filename + '.tex'

	print '\n\nConverting "' + filename + '" into "' + outfile + '"'
	try:
		f = open(filename, 'r')
	except IOError:
		print "Can not open the input file: " + filename
		exit(0)

	try:
		of = codecs.open(outfile, encoding='utf-8', mode='w')
	except IOError:
		print "Can not open the output file: " + outfile
		exit(0)

	convert_file(f, of, convert_xepersian)

	of.close()
	f.close()