Subversion Repositories SmartDukaan

Rev

Rev 14122 | Rev 14366 | Go to most recent revision | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 14122 Rev 14333
Line 64... Line 64...
64
        page=self.response_data.decode("utf-8")
64
        page=self.response_data.decode("utf-8")
65
        self.soup = BeautifulSoup(page,convertEntities=BeautifulSoup.HTML_ENTITIES)
65
        self.soup = BeautifulSoup(page,convertEntities=BeautifulSoup.HTML_ENTITIES)
66
        page = None
66
        page = None
67
        self.response_data = None
67
        self.response_data = None
68
        print "Soup created from flipkart data for %s" %(url)
68
        print "Soup created from flipkart data for %s" %(url)
69
        if (url==redirect_url):
69
        print redirect_url
70
            return self.scrape(self.soup,url)
70
        return self.scrape(self.soup,url)
71
        else:
-
 
72
            print self.redirectCount
-
 
73
            self.redirectCount+=1
-
 
74
            if self.redirectCount >5:
-
 
75
                raise
-
 
76
            return self.read(url)
-
 
77
    
71
    
78
    def scrape(self,soup,url):
72
    def scrape(self,soup,url):
79
        try:
73
        try:
80
            print "data-config"
74
            print "data-config"
81
            if soup.find('div',{'class':'seller-table-wrap section'}) is None:
75
            if soup.find('div',{'class':'seller-table-wrap section'}) is None:
Line 95... Line 89...
95
        
89
        
96
        
90
        
97
 
91
 
98
if __name__ == '__main__':
92
if __name__ == '__main__':
99
    scraper = FlipkartProductPageScraper()
93
    scraper = FlipkartProductPageScraper()
100
    print scraper.read('http://www.flipkart.com/lenovo-yoga-2-tablet-android-8-inch/p/itmeyqkz6gxnkkmn?pid=TABEYQKXHQZQNDJG')
94
    print scraper.read('http://www.flipkart.com/samsung-galaxy-e7/p/itme3un6wdygtqtr?pid=MOBE3UN6NYKGPZPM')