Subversion Repositories SmartDukaan

Rev

Rev 12190 | Rev 12199 | Go to most recent revision | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 12190 Rev 12191
Line 1... Line 1...
1 
import urllib2
 1 
import urllib2
2 
from BeautifulSoup import BeautifulSoup
 2 
from BeautifulSoup import BeautifulSoup
3 
import re
 3 
import re
4 
from sys import exit
 4 
from sys import exit
5 
from datetime import datetime
 - 
 
6 
 
 5 
 
7 
class FlipkartScraper:
 6 
class FlipkartScraper:
8 
    def __init__(self):
 7 
    def __init__(self):
9 
        self.count_trials = 0
 8 
        self.count_trials = 0
10
9
Line 13... Line 12...
13 
        request.add_header('User-Agent', 'Mozilla/5.0 (X11; Linux i686) AppleWebKit/535.1 (KHTML, like Gecko) Chrome/13.0.782.218 Safari/535.1')
 12 
        request.add_header('User-Agent', 'Mozilla/5.0 (X11; Linux i686) AppleWebKit/535.1 (KHTML, like Gecko) Chrome/13.0.782.218 Safari/535.1')
14 
        opener = urllib2.build_opener()
 13 
        opener = urllib2.build_opener()
15 
        response_data = ""
 14 
        response_data = ""
16 
        try:
 15 
        try:
17 
            response_data = opener.open(request).read()
 16 
            response_data = opener.open(request).read()
18 
            print "Response fetched ",datetime.now()
 - 
 
19
17
20 
        except urllib2.HTTPError as e:
 18 
        except urllib2.HTTPError as e:
21 
            print 'ERROR: ', e
 19 
            print 'ERROR: ', e
22 
            print 'Retrying'
 20 
            print 'Retrying'
23 
            self.count_trials += 1
 21 
            self.count_trials += 1
Line 110... Line 108...
110 
                if not temp.has_key('shippingTime'):
 108 
                if not temp.has_key('shippingTime'):
111 
                    print "Populating shipping time from metrics"
 109 
                    print "Populating shipping time from metrics"
112 
                    temp['shippingTime'] = dataMetric[3]
 110 
                    temp['shippingTime'] = dataMetric[3]
113 
                temp['sellerScore'] = int(dataMetric[4])
 111 
                temp['sellerScore'] = int(dataMetric[4])
114 
                info.append(temp)
 112 
                info.append(temp)
115 
        print "Returning response ",datetime.now()
 - 
 
116 
        return info
 113 
        return info
117 
 
 114 
 
118 
if __name__ == '__main__':
 115 
if __name__ == '__main__':
119 
    print datetime.now()
 - 
 
120 
    scraper = FlipkartScraper()
 116 
    scraper = FlipkartScraper()
121 
    scraper.read('http://www.flipkart.com/ps/ACCDSN84XXH5P9WG')
 117 
    scraper.read('http://www.flipkart.com/ps/ACCDSN84XXH5P9WG')
122 
    print scraper.createData()
 118 
    print scraper.createData()
123 
    print datetime.now()
 - 
 
124
119
125 

            120