Rev 147 | Blame | Compare with Previous | Last modification | View Log | RSS feed
'''Created on 11-May-2010@author: gaurav'''from scrapy.spider import BaseSpiderfrom scrapy.selector import HtmlXPathSelectorfrom scrapy.http import Requestfrom demo.items import DemoItemfrom scrapy.contrib.spidermiddleware import refererfrom scrapy.http.headers import Headersfrom scrapy.http.request.form import FormRequestfrom scrapy.log import msgfrom scrapy.http.response import Responsefrom datastore import DataAccessorfrom datastore.DataAccessor import DataHelperclass vendor_links(BaseSpider):domain_name = "vendors"start_urls = ["http://www.infibeam.com/Mobiles/"]def start_requests(self):request = Request(url = "http://www.infibeam.com/Mobiles/", callback=self.parse)request.headers.setdefault("Referer", "www.google.com/search")return [request]def parse(self, response):str1 = "http://www.infibeam.com"hxs = HtmlXPathSelector(response)vendor_info = hxs.select('//td[h3="Mobiles"]/ul/li')print len(vendor_info)items = []for i in vendor_info:item = {}item['name'] = i.select('.//a/@title')[0].extract()item['site'] = i.select('.//a/@href')[0].extract()items.append(item)#da = DataHelper()for item in items:str2 = str1 + str(item['site'])#da.add_vendor(item['name'], str2)print item['name']print str2SPIDER = vendor_links()