Subversion Repositories SmartDukaan

Rev

Go to most recent revision | Blame | Compare with Previous | Last modification | View Log | RSS feed

'''
Created on 20-May-2010

@author: gaurav
'''

from scrapy.spider import BaseSpider
from scrapy.selector import HtmlXPathSelector
from scrapy.http import Request

from demo.items import DemoItem
from scrapy.contrib.spidermiddleware import referer
from scrapy.http.headers import Headers
from scrapy.http.request.form import FormRequest
from scrapy.log import msg
from scrapy.http.response import Response

from datastore import DataAccessor
from datastore.DataAccessor import DataHelper


class mobilestore_spider(BaseSpider):
   
    def __init__(self): 
       self.domain_name = "mobilestore"
       str1 = "http://www.themobilestore.in/sitemap.xml"
       self.start_urls.append(str1)
    
    
    def start_requests(self):
        da = DataHelper()
        da.add_supplier(self.domain_name, "www.themobilestore.in")
        listreq = []
        for url1 in self.start_urls:
            request = Request(url = str(url1), callback=self.parse)
            request.headers.setdefault("Referer", "http://www.mobilestore.in")
            listreq.append(request)
        return listreq
    
    def parse(self, response):
        da = DataHelper()
        hxs = HtmlXPathSelector(response)
        phone_urls = hxs.select('//url/loc/text()')
        ct = 1
        for i in phone_urls:
            site = i.extract()
            pos1 = pos2 = pos3 = 0
            temp =""
            pos1 = site.rfind('/')
            if pos1 != -1:
                pos2 = site.rfind('/',0,pos1-1)
            if pos2 != -1:    
                pos3 = site.rfind('/',0,pos2-1)
            if pos3 > 0:
                temp = site[pos3+1:pos2]
            if temp == "Mobile_Phones":
                #print str(ct) + " " + site
                #print "\n"
                #ct = ct +1    
                da.add_mobstoreurl(site)           
SPIDER = mobilestore_spider()