Subversion Repositories SmartDukaan

Rev

Rev 4198 | Rev 4203 | Go to most recent revision | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 4198 Rev 4199
Line 64... Line 64...
64 
            else:
 64 
            else:
65 
                return pagination_links[-1]('a')[0]['href'].strip()
 65 
                return pagination_links[-1]('a')[0]['href'].strip()
66 
        except KeyError:
 66 
        except KeyError:
67 
            print pagination_links
 67 
            print pagination_links
68 
 
 68 
 
- 
 
 69 
    def getDataFromProductPage(self, url):
- 
 
 70 
        html = BaseScraper.read(self, url)
- 
 
 71 
        soup = BeautifulSoup(html)
- 
 
 72 
        name = soup.find('h1', {'itemprop': 'name'}).string.strip()
- 
 
 73 
        price = soup.find('span',{'id': 'fk-mprod-our-id'}).contents[2]
- 
 
 74 
        in_stock = soup.find('div', {'id': 'fk-stock-info-id'}).string.strip()
- 
 
 75
- 
 
 76 
        data = {
- 
 
 77 
            "product_url": str(url),
- 
 
 78 
            "source": "flipkart",
- 
 
 79 
            "price": price,
- 
 
 80 
            "in_stock": 1 if in_stock == 'In Stock.' else 0,
- 
 
 81 
            "name": name
- 
 
 82 
        }
- 
 
 83 
        return data
- 
 
 84 
 
- 
 
 85 
 
69 
if __name__ == '__main__':
 86 
if __name__ == '__main__':
70 
    scraper = HS18Scraper()
 87 
    scraper = HS18Scraper()
71 
    scraper.setUrl('http://www.homeshop18.com//ipads-2f-tablets/categoryid:8937/search:*/start:32/')
 88 
    scraper.setUrl('http://www.homeshop18.com//ipads-2f-tablets/categoryid:8937/search:*/start:32/')
72 
    scraper.scrape()
 89 
    scraper.scrape()
73 
    products = scraper.getPhones()
 90 
    products = scraper.getPhones()