Subversion Repositories SmartDukaan

Rev

Rev 20358 | Rev 22736 | Go to most recent revision | Details | Compare with Previous | Last modification | View Log | RSS feed

Rev Author Line No. Line
19645 kshitij.so 1
from dtr.utils.utils import to_java_date, fetchResponseUsingProxy, get_mongo_connection, CATEGORY_MAP
13828 kshitij.so 2
from datetime import datetime
14257 kshitij.so 3
import optparse
14379 kshitij.so 4
import smtplib
5
from email.mime.text import MIMEText
6
from email.mime.multipart import MIMEMultipart
19645 kshitij.so 7
import json
20357 kshitij.so 8
import chardet
21135 kshitij.so 9
from shop2020.utils.EmailAttachmentSender import get_attachment_part
10
from shop2020.utils import EmailAttachmentSender
13754 kshitij.so 11
 
19645 kshitij.so 12
 
13
headers = {
14
           'Browser-Name': 'Chrome',
15
           'User-Agent': 'Mozilla/5.0 (Linux; Android 5.1.1; A0001 Build/LMY48B; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/43.0.2357.121 Mobile Safari/537.36 FKUA/Retail/550900/Android/Mobile (OnePlus/A0001)',
16
           'Host': 'mobileapi.flipkart.net'
17
        }
18
 
13754 kshitij.so 19
con = None
14257 kshitij.so 20
parser = optparse.OptionParser()
21
parser.add_option("-m", "--m", dest="mongoHost",
22
                      default="localhost",
23
                      type="string", help="The HOST where the mongo server is running",
24
                      metavar="mongo_host")
25
 
26
(options, args) = parser.parse_args()
27
 
13754 kshitij.so 28
bestSellers = []
13828 kshitij.so 29
now = datetime.now()
14379 kshitij.so 30
exceptionList = []
13754 kshitij.so 31
 
20357 kshitij.so 32
xstr = lambda s: s or ""
19645 kshitij.so 33
 
13754 kshitij.so 34
class __RankInfo:
35
 
19645 kshitij.so 36
    def __init__(self, identifier, rank, category, title, url):
13754 kshitij.so 37
        self.identifier = identifier
38
        self.rank  = rank
14379 kshitij.so 39
        self.category = category
19645 kshitij.so 40
        self.title = title
41
        self.url = url
13754 kshitij.so 42
 
18284 kshitij.so 43
 
13754 kshitij.so 44
def scrapeBestSellerMobiles():
45
    global bestSellers
19645 kshitij.so 46
    rank = 1
47
    for i in range(0,100,10):
48
        url = "http://mobileapi.flipkart.net/3/discover/getSearch?store=tyy/4io&start=%d&count=10"%(i)
49
        print url
50
        response_data = fetchResponseUsingProxy(url, headers, livePricing=None, proxy=False, flipkart=False)
51
        input_json = json.loads(response_data)
52
        for identifier, data in (input_json['RESPONSE']['product']).iteritems():
20357 kshitij.so 53
            t_title = data['value']['titles']['title']
54
            t_subtitle = data['value']['titles']['subtitle']
55
            title = (xstr(t_title) + " " + xstr(t_subtitle)).strip()
56
            print title
19645 kshitij.so 57
            url =  (data['value']['smartUrl']).replace("http://dl.flipkart.com/dl", "http://www.flipkart.com")
58
            r_info = __RankInfo(identifier, rank, 3, title, url)
13754 kshitij.so 59
            bestSellers.append(r_info)
19645 kshitij.so 60
            rank = rank + 1
61
 
13754 kshitij.so 62
 
63
def scrapeBestSellerTablets():
64
    global bestSellers
19645 kshitij.so 65
    rank = 1
66
    for i in range(0,100,10):
67
        url = "http://mobileapi.flipkart.net/3/discover/getSearch?store=tyy/hry&start=%d&count=10"%(i)
68
        print url
69
        response_data = fetchResponseUsingProxy(url, headers, livePricing=None, proxy=False, flipkart=False)
70
        input_json = json.loads(response_data)
71
        for identifier, data in (input_json['RESPONSE']['product']).iteritems():
20357 kshitij.so 72
            t_title = data['value']['titles']['title']
73
            t_subtitle = data['value']['titles']['subtitle']
74
            title = (xstr(t_title) + " " + xstr(t_subtitle)).strip()
75
            print title
19645 kshitij.so 76
            url =  (data['value']['smartUrl']).replace("http://dl.flipkart.com/dl", "http://www.flipkart.com")
20358 kshitij.so 77
            print url
78
            print identifier
79
            print "=================="
19645 kshitij.so 80
            r_info = __RankInfo(identifier, rank, 5, title, url)
13754 kshitij.so 81
            bestSellers.append(r_info)
19645 kshitij.so 82
            rank = rank + 1
13754 kshitij.so 83
 
19648 kshitij.so 84
def commitBestSellers():
14379 kshitij.so 85
    global exceptionList
13754 kshitij.so 86
    for x in bestSellers:
19648 kshitij.so 87
        col = get_mongo_connection(host=options.mongoHost).Catalog.MasterData.find({'identifier':x.identifier.strip()})
14379 kshitij.so 88
        if len(list(col)) == 0:
89
            exceptionList.append(x)
90
        else:
19648 kshitij.so 91
            get_mongo_connection(host=options.mongoHost).Catalog.MasterData.update({'identifier':x.identifier.strip()}, {'$set' : {'rank':x.rank,'updatedOn':to_java_date(now)}}, multi=True)
13754 kshitij.so 92
 
93
def resetRanks(category):
19647 kshitij.so 94
    get_mongo_connection(host=options.mongoHost).Catalog.MasterData.update({'category_id':category}, {'$set' : {'rank':0,'updatedOn':to_java_date(now)}}, multi=True)
14379 kshitij.so 95
 
96
def sendEmail():
97
    message="""<html>
98
            <body>
99
            <h3>Flipkart Best Sellers not in master</h3>
100
            <table border="1" style="width:100%;">
101
            <thead>
102
            <tr><th>Identifier</th>
19645 kshitij.so 103
            <th>Title</th>
14379 kshitij.so 104
            <th>Category</th>
105
            <th>Rank</th>
19645 kshitij.so 106
            <th>URL</th>
14379 kshitij.so 107
            </tr></thead>
108
            <tbody>"""
109
    for item in exceptionList:
110
        message+="""<tr>
111
        <td style="text-align:center">"""+(item.identifier)+"""</td>
19645 kshitij.so 112
        <td style="text-align:center">"""+(item.title)+"""</td>
113
        <td style="text-align:center">"""+(CATEGORY_MAP.get(item.category))+"""</td>
14379 kshitij.so 114
        <td style="text-align:center">"""+str(item.rank)+"""</td>
19645 kshitij.so 115
        <td style="text-align:center">"""+(item.url)+"""</td>
14379 kshitij.so 116
        </tr>"""
117
    message+="""</tbody></table></body></html>"""
20358 kshitij.so 118
    message = message.decode('ascii', 'ignore')
19645 kshitij.so 119
    #recipients = ['kshitij.sood@saholic.com']
21135 kshitij.so 120
    recipients = ['kshitij.sood@saholic.com','ritesh.chauhan@saholic.com','aishwarya.singh@saholic.com']
121
    EmailAttachmentSender.mail_send_grid("dtr@profitmandi.com","apikey", "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", recipients, "Flipkart Best Sellers",message ,[],[],[])              
122
 
14379 kshitij.so 123
 
13754 kshitij.so 124
def main():
125
    scrapeBestSellerMobiles()
19648 kshitij.so 126
    resetRanks(3)
13754 kshitij.so 127
    scrapeBestSellerTablets()
19648 kshitij.so 128
    resetRanks(5)
129
    commitBestSellers()
14379 kshitij.so 130
    sendEmail()
13754 kshitij.so 131
 
132
if __name__=='__main__':
133
    main()