Subversion Repositories SmartDukaan

import urllib

import urllib2

from BeautifulSoup import BeautifulSoup

import pymongo

import re

from dtr.utils.utils import to_java_date

from datetime import datetime

import optparse

con = None

parser = optparse.OptionParser()

parser.add_option("-m", "--m", dest="mongoHost",

                      default="localhost",

                      type="string", help="The HOST where the mongo server is running",

                      metavar="mongo_host")

(options, args) = parser.parse_args()

bestSellers = []

now = datetime.now()

mobUrl = "http://www.saholic.com/mobile-phone/10006"

tabUrl = "http://www.saholic.com/all-tablets/10010"

class __RankInfo:

    def __init__(self, identifier, rank):

        self.identifier = identifier

        self.rank  = rank

def get_mongo_connection(host=options.mongoHost, port=27017):

    global con

    if con is None:

        print "Establishing connection %s host and port %d" %(host,port)

        try:

            con = pymongo.MongoClient(host, port)

        except Exception, e:

            print e

            return None

    return con

def getSoupObject(url,data):

    print "Getting soup object for"

    global RETRY_COUNT

    RETRY_COUNT = 1 

    while RETRY_COUNT < 10:

        try:

            soup = None

            request = urllib2.Request(url,data)

            request.add_header('Accept','text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8')

            request.add_header('Accept-Language','en-US,en;q=0.8,hi;q=0.6')

            request.add_header('Connection','keep-alive')

            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')

            response = urllib2.urlopen(request)   

            response_data = response.read()

            response.close()

            try:

                page=response_data.decode("utf-8")

                soup = BeautifulSoup(page,convertEntities=BeautifulSoup.HTML_ENTITIES)

            except:

                soup = BeautifulSoup(response_data,convertEntities=BeautifulSoup.HTML_ENTITIES)

            if soup is None:

                raise

            return soup

        except Exception as e:

            print e

            print "Retrying"

            RETRY_COUNT = RETRY_COUNT + 1

def scrapeBestSellerMobiles():

    global bestSellers

    rank = 1

    for i in range(1,6):

        query_args = { 'fq':'F_50028:In Stock', 'page':i }

        data = urllib.urlencode(query_args)

        soup = getSoupObject(mobUrl,data)

        titleTags = soup.findAll('div',{'class':'title'})

        for titleTag in titleTags:

            print titleTag.find('a').text

            identifier = getSaholicIdentifier(titleTag.find('a')['href'])

            r_info = __RankInfo(str(identifier), rank)

            bestSellers.append(r_info)

            rank = rank + 1

def scrapeBestSellerTablets():

    global bestSellers

    bestSellers = [] 

    rank = 1

    for i in range(1,6):

        query_args = { 'fq':'F_50028:In Stock', 'page':i }

        data = urllib.urlencode(query_args)

        soup = getSoupObject(tabUrl, data)

        titleTags = soup.findAll('div',{'class':'title'})

        for titleTag in titleTags:

            print titleTag.find('a').text

            identifier = getSaholicIdentifier(titleTag.find('a')['href'])

            r_info = __RankInfo(str(identifier), rank)

            bestSellers.append(r_info)

            rank = rank + 1

def getSaholicIdentifier(url):

    return url[url.rfind('-')+len('-'):]

def resetRanks(category):

    oldRankedItems = get_mongo_connection().Catalog.MasterData.find({'rank':{'$gt':0},'source_id':4,'category':category})

    for item in oldRankedItems:

        print item['_id']

        get_mongo_connection().Catalog.MasterData.update({'_id':item['_id']}, {'$set' : {'rank':0,'updatedOn':to_java_date(now)}}, multi=True)

def commitBestSellers():

    print "Rank",

    print '\t',

    print 'Identifier'

    for x in bestSellers:

        print x.rank,

        print '\t',

        print x.identifier,

        col = get_mongo_connection().Catalog.MasterData.find({'identifier':x.identifier})

        print "count sku",

        print '\t',

        print len(list(col))

        get_mongo_connection().Catalog.MasterData.update({'identifier':x.identifier,'source_id':4}, {'$set' : {'rank':x.rank,'updatedOn':to_java_date(now)}}, multi=True)

def main():

    scrapeBestSellerMobiles()

    if len(bestSellers) > 0:

        resetRanks('Mobiles')

        commitBestSellers()

    scrapeBestSellerTablets()

    if len(bestSellers) > 0:

        resetRanks('Tablets')

        commitBestSellers()

if __name__=='__main__':

    main()