Subversion Repositories SmartDukaan

Rev

Rev 152 | Rev 239 | Go to most recent revision | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 152 Rev 169
Line 19... Line 19...
19 
from datastore import DataAccessor
 19 
from datastore import DataAccessor
20 
from datastore.DataAccessor import DataHelper
 20 
from datastore.DataAccessor import DataHelper
21 
 
 21 
 
22 
 
 22 
 
23 
class vendor_links(BaseSpider):
 23 
class vendor_links(BaseSpider):
24 
    domain_name = "univercellvendors"
 24 
    domain_name = "univercell"
25 
    start_urls = [
 25 
    start_urls = [
26 
          "http://www.univercell.in/mobiles/populateStore.action"
 26 
          "http://www.univercell.in/mobiles/populateStore.action"
27 
    ]
 27 
    ]
28
28
29 
    def start_requests(self):
 29 
    def start_requests(self):
- 
 
 30 
        da = DataHelper()
- 
 
 31 
        da.add_supplier(self.domain_name, "www.univercell.in")
30 
        request = Request(url = "http://www.univercell.in/mobiles/populateStore.action", callback=self.parse)
 32 
        request = Request(url = "http://www.univercell.in/mobiles/populateStore.action", callback=self.parse)
31 
        request.headers.setdefault("Referer", "www.google.com/search")
 33 
        request.headers.setdefault("Referer", "www.google.com/search")
32 
        return [request]
 34 
        return [request]
33
35
34 
    def parse(self, response):
 36 
    def parse(self, response):
35 
        str1 = "http://www.univercell.in"
 37 
        str1 = "http://www.univercell.in"
36 
        hxs = HtmlXPathSelector(response)
 38 
        hxs = HtmlXPathSelector(response)
37 
        vendor_info = hxs.select('//div[@id="mobilesTab"]/table/tr[1]/td/table/tr')
 39 
        vendor_info = hxs.select('//div[@id="mobilesTab"]/table/tr[1]/td/table/tr')
38 
        print len(vendor_info)
 40 
        #print len(vendor_info)
39 
        items = []
 41 
        items = []
40 
        for i in vendor_info:
 42 
        for i in vendor_info:
41 
            item = {}
 43 
            item = {}
42 
            item['name'] = i.select('.//a/text()')[0].extract()
 44 
            item['name'] = i.select('.//a/text()')[0].extract()
43 
            temp = i.select('.//a/@href')[0].extract()
 45 
            temp = i.select('.//a/@href')[0].extract()