Subversion Repositories SmartDukaan

Rev

Rev 14167 | Go to most recent revision | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 14167 Rev 14169
Line 44... Line 44...
44 
 
 44 
 
45 
            if self.count_trials < 3:
 45 
            if self.count_trials < 3:
46 
                return self.read(url)
 46 
                return self.read(url)
47 
 
 47 
 
48 
        self.response_data=response_data
 48 
        self.response_data=response_data
49 
        return self.createSoup(url,redirect_url)
 49 
        return self.createSoup(url)
50
50
51
51
52 
    def createSoup(self,url, redirect_url):
 52 
    def createSoup(self,url):
53 
        print "Creating soup from flipkart data for %s" %(url)
 53 
        print "Creating soup from flipkart data for %s" %(url)
54 
        page=self.response_data.decode("utf-8")
 54 
        page=self.response_data.decode("utf-8")
55 
        self.soup = BeautifulSoup(page,convertEntities=BeautifulSoup.HTML_ENTITIES)
 55 
        self.soup = BeautifulSoup(page,convertEntities=BeautifulSoup.HTML_ENTITIES)
56 
        print "Soup created from flipkart data for %s" %(url)
 56 
        print "Soup created from flipkart data for %s" %(url)
57 
        return self.scrape(self.soup,url)
 57 
        return self.scrape(self.soup)
58 
 
 58 
 
59 
    def scrape(self,soup,url):
 59 
    def scrape(self,soup):
60 
        try:
 60 
        try:
61 
            print "data-config"
 61 
            print "data-config"
62 
            if soup.find('div',{'class':'seller-table-wrap section'}) is None:
 62 
            if soup.find('div',{'class':'seller-table-wrap section'}) is None:
63 
                raise
 63 
                raise
64 
            x = json.loads(soup.find('div',{'class':'seller-table-wrap section'})['data-config'])['dataModel']
 64 
            x = json.loads(soup.find('div',{'class':'seller-table-wrap section'})['data-config'])['dataModel']