#!/usr/bin/python
# encoding=utf8
import smtplib
import time
import sys
from pageimports import page, page1, page3

reload(sys)
sys.setdefaultencoding('utf8')




data = []
comparedata = []
differentdata = []

def get_links(textmatter):
	textmatter = textmatter.encode('utf')
	start_deadmark = textmatter.find('class="deadmark"></span><a href="')
	startlink = textmatter.find('http',start_deadmark)
	endlink = textmatter.find('"', startlink)

	#textwala
	starttext = textmatter.find('>', endlink)
	endtext = textmatter.find('<',starttext)
	textoutput =  textmatter[starttext+1:endtext]

	linkoutput = textmatter[startlink:endlink]
	return textoutput, linkoutput, endtext


def scrape_all_data(textmatter,datalist):
	'''TEXTMATTER SHOULD BE STRING - DATALIST SHOULD BE LIST
	'''
	x = 0
	while x <=30:
		#if len(textmatter) >0:
		texto, linko, endtext = get_links(textmatter)
		textmatter = textmatter[endtext:]
		datalist.append([texto, linko])
		x +=1



		#else:
		#	print 'else break'
		#	break


def compare(textmatter, pagetest, data, comparedata, differentdata):
	if len(data) > 0:
		scrape_all_data(pagetest,comparedata)
		# comparisoin goes here
		# compare data with comparedate
		same = True
		for each in comparedata:
			milgaya = False
			for every in data:
				if each[0] in every:
					milgaya = True
			if milgaya == False:
				differentdata.append(each)
				same = False
					#print differentdata
					#print j # changed data from the comparedata for notification
		if same:
			comparedata = []
			print 'no new data on the website'
		# if everything is same then comparedata = []
		# if it is different
		else:
			a = ' '.join(differentdata[0])
			print a
			fromaddr = 'usefulemail70@gmail.com'
			toaddrs = 'abdussamad.m7@gmail.com' #enter your email address
			msg =  """
			The website has been updated,
			Heading of news and link =  %s
			""" % (''.join(a))

			username = 'usefulemail70@gmail.com'
			password = 'myeasypassword'

			server = smtplib.SMTP('smtp.gmail.com:587')
			server.starttls()
			server.login(username,password)
			server.sendmail(fromaddr,toaddrs,msg)
			server.quit()
			print 'new data on the website:'
			print '-----------------------'
			print differentdata
			data = comparedata
			differentdata = []
			comparedata = []

		# then find the difference
		# then mail the difference
		# then data = comparedata
		# then comparedata = []
	else:
		scrape_all_data(textmatter,data)
		print 'data has been added for the first time'


def timer(page,page1, data, comparedata, differentdata):
	min= 2
	while min>0:
		compare(page,page1, data, comparedata, differentdata)
		time.sleep(10) # every 10 secs just for testing later will make it 300 secs
		min-=1


print timer(page,page1, data, comparedata, differentdata)