import HTMLParser
 class TMP(HTMLParser.HTMLParser):
 	INI = 0
 	DIV = 1
 	OTH = 2
 	AST = 3
 	def __init__(self):
 		HTMLParser.HTMLParser.__init__(self)
 		self.links = [];self.__tmp = [];self.dbuf = '';self.__state = self.INI;self.level = 0
 	def handle_starttag(self, tag, attrs):
 		attrs = dict(attrs);self.level += 1
 		print '%stag: %r, attrs: %r, state: %r' % ('\t'*self.level, tag, attrs, self.__state),
 		if tag == 'div' and self.__state != self.DIV and attrs.get('class') == 'smallbox':
 			print 'useful div',
 			self.__state = self.DIV
 		elif tag == 'img' and self.__state == self.OTH:
 			print 'useful image',
 			self.__tmp.append(attrs['src'])
 		elif tag == 'a' and self.__state == self.DIV:
 			print 'useful link',
 			self.__tmp.append(attrs['href']);self.__state = self.OTH
 		print '..done'
 	def handle_endtag(self, tag):
 		print '%sclose tag: %r' % ('\t'*self.level, tag)
 		self.level -= 1
 		if tag == 'div' and self.__state == self.DIV:
 			if self.dbuf is not '':
 				self.__tmp.append(self.dbuf)
 				self.dbuf = ''
 			self.links.append(self.__tmp);self.__tmp = []
 			self.__state == self.INI
 		elif tag == 'a' and self.__state == self.OTH:
 			self.__state = self.DIV
 			if self.dbuf is not '':
 				self.__tmp.append(self.dbuf)
 				self.dbuf = ''
 	def handle_data(self, data):
 		if self.__state == self.OTH:
 			self.dbuf += data + ' '