Subversion Repositories SmartDukaan

Rev

Rev 14167 | Go to most recent revision | Details | Compare with Previous | Last modification | View Log | RSS feed

Rev Author Line No. Line
14167 kshitij.so 1
import urllib2
2
from BeautifulSoup import BeautifulSoup
3
import re
4
from sys import exit
5
import json
6
 
7
headers = { 
8
            'User-agent':'Mozilla/5.0 (X11; Linux i686) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/37.0.2062.120 Safari/537.36',
9
            'Accept' : 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',      
10
            'Accept-Language' : 'en-US,en;q=0.8',                     
11
            'Accept-Charset' : 'ISO-8859-1,utf-8;q=0.7,*;q=0.3',
12
            'Cookie':'T=TI141257426738726661427143281839817329423126740566618323641725716448; __sonar=7237334677420142002; __gads=ID=c8b82101a0e4f451:T=1412574724:S=ALNI_MbPMbEOZj2nAGjM54z8ZHFMqwTOTQ; FK-CMP-DATA=; SN=2.VI11FB3FB6ED9D4693A796AB8C965B3417.SI802C325AC43444858830E870C4FD3324.VS141257426735693951472.1412576209; VID=2.VI11FB3FB6ED9D4693A796AB8C965B3417.1412576209.VS141257426735693951472; NSID=2.SI802C325AC43444858830E870C4FD3324.1412576209.VI11FB3FB6ED9D4693A796AB8C965B3417; __utma=19769839.709301254.1412574234.1412574234.1412574234.1; __utmb=19769839.23.10.1412574234; __utmc=19769839; __utmz=19769839.1412574234.1.1.utmcsr=(direct)|utmccn=(direct)|utmcmd=(none); s_cc=true; gpv_pn=SellerListing%3AMobile%3AMicromax%20Canvas%20Fire%20A093; gpv_pn_t=no%20value; s_sq=%5B%5BB%5D%5D; pincode=110011; s_ppv=36',
13
            'Connection':'keep-alive'
14
        }
15
 
16
class FlipkartProductPageScraper:
17
    def __init__(self):
18
        self.count_trials = 0
19
        self.redirectCount = 0
20
 
21
    def read(self, url):
22
        request = urllib2.Request(url)
23
        request.add_header('Accept','text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8')
24
        request.add_header('Accept-Language','en-US,en;q=0.8,hi;q=0.6')
25
        request.add_header('Connection','keep-alive')
26
        request.add_header('Cookie','T=TI141257426738726661427143281839817329423126740566618323641725716448; __sonar=7237334677420142002; __gads=ID=c8b82101a0e4f451:T=1412574724:S=ALNI_MbPMbEOZj2nAGjM54z8ZHFMqwTOTQ; FK-CMP-DATA=; SN=2.VI11FB3FB6ED9D4693A796AB8C965B3417.SI802C325AC43444858830E870C4FD3324.VS141257426735693951472.1412576209; VID=2.VI11FB3FB6ED9D4693A796AB8C965B3417.1412576209.VS141257426735693951472; NSID=2.SI802C325AC43444858830E870C4FD3324.1412576209.VI11FB3FB6ED9D4693A796AB8C965B3417; __utma=19769839.709301254.1412574234.1412574234.1412574234.1; __utmb=19769839.23.10.1412574234; __utmc=19769839; __utmz=19769839.1412574234.1.1.utmcsr=(direct)|utmccn=(direct)|utmcmd=(none); s_cc=true; gpv_pn=SellerListing%3AMobile%3AMicromax%20Canvas%20Fire%20A093; gpv_pn_t=no%20value; s_sq=%5B%5BB%5D%5D; pincode=110011; s_ppv=36')
27
        request.add_header('Cache-Control','max-age=0')
28
        request.add_header('Host','www.flipkart.com')
29
        request.add_header('User-Agent','Mozilla/5.0 (X11; Linux i686) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/37.0.2062.120 Safari/537.36')
30
        response_data = ""
31
        redirect_url = ""
32
        try:
33
            req = urllib2.Request(url,headers=headers)
34
            response = urllib2.urlopen(req)
35
            response_data = response.read()
36
            response.close()
37
            print "Fetched response from flipkart for %s" %(url)
38
            redirect_url = response.url
39
 
40
        except Exception as e:
41
            print 'ERROR: ', e
42
            print 'Retrying'
43
            self.count_trials += 1
44
 
45
            if self.count_trials < 3:
46
                return self.read(url)
47
 
48
        self.response_data=response_data
14169 kshitij.so 49
        return self.createSoup(url)
14167 kshitij.so 50
 
51
 
14169 kshitij.so 52
    def createSoup(self,url):
14167 kshitij.so 53
        print "Creating soup from flipkart data for %s" %(url)
54
        page=self.response_data.decode("utf-8")
55
        self.soup = BeautifulSoup(page,convertEntities=BeautifulSoup.HTML_ENTITIES)
56
        print "Soup created from flipkart data for %s" %(url)
14169 kshitij.so 57
        return self.scrape(self.soup)
14167 kshitij.so 58
 
14169 kshitij.so 59
    def scrape(self,soup):
14167 kshitij.so 60
        try:
61
            print "data-config"
62
            if soup.find('div',{'class':'seller-table-wrap section'}) is None:
63
                raise
64
            x = json.loads(soup.find('div',{'class':'seller-table-wrap section'})['data-config'])['dataModel']
65
            lines = sorted(x, key=lambda k: k['priceInfo'].get('sellingPrice', 0), reverse=False)
66
            sellingPrice =  float(lines[0]['priceInfo']['sellingPrice'])
67
            try:
68
                offerText = lines[0]['offerInfo']['listingOffers'][0]['description']
69
            except:
70
                offerText = ""
71
            return {'lowestSp':sellingPrice,'inStock':1}
72
        except:
73
            return {'lowestSp':0,'inStock':0}
74
 
75
if __name__ == '__main__':
76
    scraper = FlipkartProductPageScraper()
77
    print scraper.read('http://www.flipkart.com/axe-denim-deodorant-spray-150-ml/p/itmdtbxyjfzfkppn?pid=DEOD8DHEW5NPTMNR')