Subversion Repositories SmartDukaan

Rev

Rev 14157 | Go to most recent revision | Details | Last modification | View Log | RSS feed

Rev Author Line No. Line
13829 kshitij.so 1
import urllib2
2
from BeautifulSoup import BeautifulSoup, NavigableString
3
import re
4
import sys
5
 
6
invalid_tags = ['b', 'i', 'u']
7
bestSellers = []
8
 
9
def strip_tags(html, invalid_tags):
10
    soup = BeautifulSoup(html,convertEntities=BeautifulSoup.HTML_ENTITIES)
11
 
12
    for tag in soup.findAll(True):
13
        if tag.name in invalid_tags:
14
            s = ""
15
 
16
            for c in tag.contents:
17
                if not isinstance(c, NavigableString):
18
                    c = strip_tags(unicode(c), invalid_tags)
19
                s += unicode(c)
20
 
21
            tag.replaceWith(s)
22
 
23
    return soup
24
 
25
class AmazonScraper:
26
    def __init__(self):
27
        self.count_trials = 0
28
 
29
    def read(self, url):
30
        request = urllib2.Request(url)
31
        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')
32
        opener = urllib2.build_opener()
33
        response_data = ""
34
        try:
35
            response_data = opener.open(request).read()
36
 
37
        except urllib2.HTTPError as e:
38
            print 'ERROR: ', e
39
            print 'Retrying'
40
            self.count_trials += 1
41
 
42
            if self.count_trials < 3:
43
                return self.read(url)
44
 
45
        self.response_data=response_data
46
        return self.createData()
47
 
48
    def createData(self):
49
        self.soup = strip_tags(self.response_data,invalid_tags)
50
        self.response_data =None
51
        return self.scrape(self.soup)
52
 
53
 
54
    def scrape(self,soup):
55
        sellerData = soup.findAll("div" , {"class" : "a-row a-spacing-mini olpOffer"})
56
        for data in sellerData:
57
            print "sellerData****"
58
            price = data.find('span', attrs={'class' : re.compile('.*olpOfferPrice*')}).find('span').text
59
            print "Unit cost= ",float(price.replace("Rs.","").replace(",",""))
60
            unitCost = float(price.replace("Rs.","").replace(",",""))
61
            shippingCost = data.find('p', attrs={'class' : re.compile('.*olpShippingInfo*')}).find('span').text
62
            if "FREE" in shippingCost:
63
                print "shippingCost=0"
64
                shippingCost = 0
65
            else:
66
                try:
67
                    print "shippingCost= ",float(shippingCost.replace("+Rs.","").replace("Delivery","").replace(",",""))
68
                    shippingCost = float(shippingCost.replace("+Rs.","").replace("Delivery","").replace(",",""))
69
                except:
70
                    shippingCost = 0.0
71
            try:
72
                return (unitCost + shippingCost)
73
            except:
74
                return  unitCost
75
 
76
 
77
if __name__ == '__main__':
78
    scraper = AmazonScraper()
79
    print scraper.read('http://www.amazon.in/gp/offer-listing/B00J60MFTO/ref=olp_sort_ps')
80