| 13831 |
kshitij.so |
1 |
import urllib2
|
|
|
2 |
from BeautifulSoup import BeautifulSoup
|
|
|
3 |
import re
|
|
|
4 |
from sys import exit
|
|
|
5 |
|
| 14176 |
kshitij.so |
6 |
headers = {
|
|
|
7 |
'User-agent':'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/535.11 (KHTML, like Gecko) Chrome/17.0.963.56 Safari/535.11',
|
|
|
8 |
'Accept' : 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
|
|
|
9 |
'Accept-Language' : 'en-US,en;q=0.8',
|
|
|
10 |
'Accept-Charset' : 'ISO-8859-1,utf-8;q=0.7,*;q=0.3',
|
|
|
11 |
'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; s_ppv=36',
|
|
|
12 |
'Host':'www.flipkart.com'
|
|
|
13 |
}
|
|
|
14 |
|
| 13831 |
kshitij.so |
15 |
class FlipkartScraper:
|
|
|
16 |
def __init__(self):
|
|
|
17 |
self.count_trials = 0
|
|
|
18 |
self.redirectCount = 0
|
|
|
19 |
|
|
|
20 |
def read(self, url):
|
|
|
21 |
try:
|
| 14541 |
kshitij.so |
22 |
proxy = urllib2.ProxyHandler({'http': 'http://192.161.163.60:8800'})
|
|
|
23 |
opener = urllib2.build_opener(proxy)
|
|
|
24 |
urllib2.install_opener(opener)
|
|
|
25 |
response = urllib2.urlopen(url)
|
| 13831 |
kshitij.so |
26 |
response_data = response.read()
|
| 14157 |
kshitij.so |
27 |
response.close()
|
| 14541 |
kshitij.so |
28 |
# try:
|
|
|
29 |
# response = urllib2.urlopen(request)
|
|
|
30 |
# response_data = response.read()
|
|
|
31 |
# response.close()
|
|
|
32 |
# print "Fetched response from flipkart for %s" %(url)
|
| 13831 |
kshitij.so |
33 |
|
|
|
34 |
except Exception as e:
|
|
|
35 |
print 'ERROR: ', e
|
|
|
36 |
print 'Retrying'
|
|
|
37 |
self.count_trials += 1
|
|
|
38 |
|
|
|
39 |
if self.count_trials < 3:
|
|
|
40 |
return self.read(url)
|
|
|
41 |
|
|
|
42 |
self.response_data=response_data
|
| 14168 |
kshitij.so |
43 |
return self.createSoup(url)
|
| 13831 |
kshitij.so |
44 |
|
| 14168 |
kshitij.so |
45 |
# def scrapeRedirectedPage(self,soup,redirect_url):
|
|
|
46 |
# print soup
|
|
|
47 |
# print redirect_url
|
|
|
48 |
# t = soup.find("div" , {"class" : "seller-table fk-user-select-none line"})
|
|
|
49 |
# print t
|
|
|
50 |
# table_rows = t.findAll("tr" , {"class" : re.compile('t-row.*')})
|
|
|
51 |
# print table_rows
|
|
|
52 |
# for x in table_rows:
|
|
|
53 |
# print x
|
|
|
54 |
#
|
|
|
55 |
def createSoup(self, url):
|
| 13831 |
kshitij.so |
56 |
print "Creating soup from flipkart data for %s" %(url)
|
|
|
57 |
page=self.response_data.decode("utf-8")
|
|
|
58 |
self.soup = BeautifulSoup(page,convertEntities=BeautifulSoup.HTML_ENTITIES)
|
|
|
59 |
page = None
|
|
|
60 |
self.response_data = None
|
|
|
61 |
print "Soup created from flipkart data for %s" %(url)
|
| 14168 |
kshitij.so |
62 |
return self.scrape(self.soup,url)
|
| 13831 |
kshitij.so |
63 |
|
|
|
64 |
def scrape(self,soup,url):
|
|
|
65 |
print "Inside json creator for %s" %(url)
|
|
|
66 |
info = []
|
|
|
67 |
oddSeller = soup.findAll("div" , {"class" : "line seller-item odd "})
|
|
|
68 |
for data in oddSeller:
|
|
|
69 |
temp={}
|
|
|
70 |
price = data.find('span', attrs={'class' : re.compile('pxs-final-price.*')}).string.strip('Rs.').strip()
|
|
|
71 |
temp['sellingPrice']=float(price)
|
|
|
72 |
for metrics in data.find("div",{"class":"fk-text-right"}):
|
|
|
73 |
try:
|
|
|
74 |
metric = metrics.findAll('input', {'type': 'submit'})
|
|
|
75 |
except AttributeError:
|
|
|
76 |
continue
|
|
|
77 |
dataMetrics = metric[0]['data-listing-metrics']
|
|
|
78 |
dataMetric = dataMetrics.split(';')
|
|
|
79 |
temp['sellingPriceMetric'] = float(dataMetric[1])
|
| 13935 |
kshitij.so |
80 |
try:
|
|
|
81 |
temp['shippingFee'] = float(dataMetric[2])
|
|
|
82 |
except:
|
|
|
83 |
temp['shippingFee'] = 0.0
|
|
|
84 |
temp['sellingPrice'] = temp['sellingPrice'] + temp['shippingFee']
|
| 13831 |
kshitij.so |
85 |
info.append(temp)
|
|
|
86 |
evenSeller = soup.findAll("div" , {"class" : "line seller-item even "})
|
|
|
87 |
for data in evenSeller:
|
|
|
88 |
temp={}
|
|
|
89 |
price = data.find('span', attrs={'class' : re.compile('pxs-final-price.*')}).string.strip('Rs.')
|
|
|
90 |
temp['sellingPrice']=float(price)
|
|
|
91 |
for metrics in data.find("div",{"class":"fk-text-right"}):
|
|
|
92 |
try:
|
|
|
93 |
metric = metrics.findAll('input', {'type': 'submit'})
|
|
|
94 |
except AttributeError:
|
|
|
95 |
continue
|
|
|
96 |
dataMetrics = metric[0]['data-listing-metrics']
|
|
|
97 |
dataMetric = dataMetrics.split(';')
|
|
|
98 |
temp['sellingPriceMetric'] = float(dataMetric[1])
|
| 13935 |
kshitij.so |
99 |
try:
|
|
|
100 |
temp['shippingFee'] = float(dataMetric[2])
|
|
|
101 |
except:
|
|
|
102 |
temp['shippingFee'] = 0.0
|
|
|
103 |
temp['sellingPrice'] = temp['sellingPrice'] + temp['shippingFee']
|
| 13831 |
kshitij.so |
104 |
info.append(temp)
|
|
|
105 |
print info
|
|
|
106 |
print "Returning Json response from flipkart for %s" %(url)
|
|
|
107 |
return info
|
|
|
108 |
|
|
|
109 |
if __name__ == '__main__':
|
|
|
110 |
scraper = FlipkartScraper()
|
|
|
111 |
scraper.read('http://www.flipkart.com/ps/MOBDZB3Q8WJNKVHG')
|