css_parser2.py~

#!/usr/bin/python
# ------------------------------------------------------------
# css_parser.py
#
# tokenizer for css
# ------------------------------------------------------------
import ply.lex as lex
import ply.yacc as yacc
import sys
import logging
DEBUG=False
#List of pstates
pstates=['NORMAL',
		'RULE']

pstate='NORMAL'
inComment=False
actualRule=None
actualSelector=None
PASE=0
# List of token names that ARE NOT reserved words.
tokens=[
	'START_COMMENT',
	'END_COMMENT',
	'DOLLAR',
	'AT',
	'EQUALS',
	'RCURL',
	'LCURL',
	'COLON',
	'SEMICOLON',
	'STRING'
]

#List of reserved words.
reserved={
	'private': 'PRIVATE'
}
#The final list of tokens made up from plain tokens and reserved words.
tokens = tokens + list(reserved.values())
#List of ignored characters
t_ignore = ' \t\r'

#Token value assignation
t_LCURL=r'\{'
t_RCURL=r'\}'
t_EQUALS=r'='
t_DOLLAR=r'[$]'
t_AT=r'@'
t_START_COMMENT=r'[/][*]'
t_END_COMMENT=r'[*][/]'
t_COLON=':'
t_SEMICOLON=r'[;]'



def t_STRING(t):
	'([A-Za-z0-9 ]|[,_>.#"%()]|[-]|\'|[[]|[]])+'
	#Si no es una palabra reservada devuelve String
	t.type=reserved.get(t.value,'STRING')
	return t
	
# Error handling rule
def t_error(t):
    print "Illegal character '%s'" % t.value[0]
    t.lexer.skip(1)
# NewLine handling rule
def t_newline(t):
	r'\n+'
	t.lexer.lineno+=t.value.count("\n")
#######################################################PARSER
lex.lex()

#dictionary of constants
constants={}
#directory of plain rules
plain_rules=[]
#directory of directionable rules
directionable_rules={}

def p_line(t):
	'''line : constant_definition
			| css_property
			| selector
			| rule_invocation
			| comment
			| void_line
			| character'''
	t[0]="hola"

def p_constant_definition(t):
	'constant_definition : DOLLAR STRING EQUALS STRING'
	console("constant_definition",t)
	if(pstate=='NORMAL'):
		constants[t[2]]=t[4]
		print constants
	else:
		error("Constant definition in a "+pstate+ "pstate")
	
def p_constant_invocation(t):
	'constant_invocation : DOLLAR STRING'
	if(PASE==1):	
		console("constant_invocation",t)
		if(pstate=='RULE'):
			t[0]=constants[t[2]]
		elif(pstate=='NORMAL'):
			error("Constant invocation outside a rule.")
	
def p_rule_invocation(t):
	'''rule_invocation : AT STRING
						| AT STRING LCURL'''
	if(PASE==0):
		if(pstate=='NORMAL'):
			actualRule=t[2]
			directionable_rules[actualRule]=[]
			print("new_rule")
	elif(PASE==1):
		if(pstate=='RULE'):
			logMsg("rule_invocation")

		
def p_selector(t):
	'''selector : STRING LCURL
				| STRING COLON STRING LCURL'''
	if(len(t)==3):
		actualSelector=t[1]
	else:
		actualSelector=str(t[1])+":"+str(t[3])
	
	if(PASE==0):
		plain_rules.append([actualSelector,[]])
	logMsg("selector: "+actualSelector)
 
def getAsInDic(tup,key):
	for a in tup:
		if(a[0]=="key"):
			return a[1]
	raise KeyError(str(a[0])+" no encontrado en"+str(tup))
	
def p_css_property(t):
	'''css_property : STRING COLON STRING SEMICOLON
					| STRING COLON constant_invocation SEMICOLON'''
	if(PASE==0):
		print("CSS PROP: "+str(actualRule)+" "+str(actualSelector))
		try:
			print "CSS prop:"+str(directionable_rules[actualRule])
			directionable_rules[actualRule].append([t[1],t[3]])
			print "CSS prop:"+str(directionable_rules[actualRule])		
			print 
		except KeyError:
			print "Key Error directionable: "+str(actualRule)
		try:
			getAsInDic(plain_rules,actualSelector).append([t[1],t[3]])
		except KeyError:
			pass
	console("property:",t)
	
def p_character(t):
	'''character : START_COMMENT
				| END_COMMENT
				| RCURL'''
	if(pstate=='RULE'):
		if(t[1]=="}"):
			actualRule=None
			actualSelector=None

	
	console("character:",t)

def p_comment(t):
	'comment : START_COMMENT STRING END_COMMENT'
	logMsg("comment: "+str(t[2]))
	
def p_void_line(t):
	'void_line : '
	#console("void", ["void_line"]);
	pass

def p_error(t):
	logMsg("error:"+str(t))

def logMsg(msg):
	print msg

def console(name,t):
	if(DEBUG):
		print name+":"+str(list(t))

##################################################LOGGER
logging.basicConfig(
	level=logging.DEBUG,
	filename="log.txt",
	filemode="w",
	format="%(filename)10s:%(lineno)4d:%(message)s"
)
log=logging.getLogger()
##################################################END LOGGER
print "\t\tCONSTANTS"
print constants
print "\t\tPLAIN RULES"
print plain_rules
print "\t\tDIRECTIONABLE RULES"
print directionable_rules
print "\n\n\n"
#######################################################PARSE
def parse(line):
	return yacc.yacc().parse(str(line),debug=log)