Subversion Repositories SmartDukaan

Rev

Rev 14307 | Rev 15154 | Go to most recent revision | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 14307 Rev 14759
Line 1... Line 1...
1 
import urllib2
 1 
import urllib2
2 
from BeautifulSoup import BeautifulSoup, NavigableString
 2 
from BeautifulSoup import BeautifulSoup, NavigableString
3 
import re
 3 
import re
4 
import sys
 4 
from dtr.utils.utils import fetchResponseUsingProxy
5 
 
 5 
 
6 
invalid_tags = ['b', 'i', 'u']
 6 
invalid_tags = ['b', 'i', 'u']
7 
bestSellers = []
 7 
bestSellers = []
8 
 
 8 
 
9 
def strip_tags(html, invalid_tags):
 9 
def strip_tags(html, invalid_tags):
Line 25... Line 25...
25 
class AmazonScraper:
 25 
class AmazonScraper:
26 
    def __init__(self):
 26 
    def __init__(self):
27 
        self.count_trials = 0
 27 
        self.count_trials = 0
28
28
29 
    def read(self, url):
 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 
        response_data = ""
 30 
        response_data = ""
33 
        try:
 31 
        try:
34 
            response = urllib2.urlopen(request)
 - 
 
35 
            response_data = response.read()
 32 
            response_data = fetchResponseUsingProxy(url)
36 
            response.close()
 - 
 
37
- 
 
38 
        except urllib2.HTTPError as e:
 33 
        except Exception as e:
39 
            print 'ERROR: ', e
 34 
            print 'ERROR: ', e
40 
            print 'Retrying'
 35 
            print 'Retrying'
41 
            self.count_trials += 1
 36 
            self.count_trials += 1
42
37
43 
            if self.count_trials < 3:
 38 
            if self.count_trials < 3: