Subversion Repositories SmartDukaan

Rev

Rev 20357 | Rev 21135 | 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
13754 kshitij.so 9
 
19645 kshitij.so 10
 
11
 
12
headers = {
13
           'Browser-Name': 'Chrome',
14
           '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)',
15
           'Host': 'mobileapi.flipkart.net'
16
        }
17
 
13754 kshitij.so 18
con = None
14257 kshitij.so 19
parser = optparse.OptionParser()
20
parser.add_option("-m", "--m", dest="mongoHost",
21
                      default="localhost",
22
                      type="string", help="The HOST where the mongo server is running",
23
                      metavar="mongo_host")
24
 
25
(options, args) = parser.parse_args()
26
 
13754 kshitij.so 27
bestSellers = []
13828 kshitij.so 28
now = datetime.now()
14379 kshitij.so 29
exceptionList = []
13754 kshitij.so 30
 
20357 kshitij.so 31
xstr = lambda s: s or ""
19645 kshitij.so 32
 
13754 kshitij.so 33
class __RankInfo:
34
 
19645 kshitij.so 35
    def __init__(self, identifier, rank, category, title, url):
13754 kshitij.so 36
        self.identifier = identifier
37
        self.rank  = rank
14379 kshitij.so 38
        self.category = category
19645 kshitij.so 39
        self.title = title
40
        self.url = url
13754 kshitij.so 41
 
18284 kshitij.so 42
 
13754 kshitij.so 43
def scrapeBestSellerMobiles():
44
    global bestSellers
19645 kshitij.so 45
    rank = 1
46
    for i in range(0,100,10):
47
        url = "http://mobileapi.flipkart.net/3/discover/getSearch?store=tyy/4io&start=%d&count=10"%(i)
48
        print url
49
        response_data = fetchResponseUsingProxy(url, headers, livePricing=None, proxy=False, flipkart=False)
50
        input_json = json.loads(response_data)
51
        for identifier, data in (input_json['RESPONSE']['product']).iteritems():
20357 kshitij.so 52
            t_title = data['value']['titles']['title']
53
            t_subtitle = data['value']['titles']['subtitle']
54
            title = (xstr(t_title) + " " + xstr(t_subtitle)).strip()
55
            print title
19645 kshitij.so 56
            url =  (data['value']['smartUrl']).replace("http://dl.flipkart.com/dl", "http://www.flipkart.com")
57
            r_info = __RankInfo(identifier, rank, 3, title, url)
13754 kshitij.so 58
            bestSellers.append(r_info)
19645 kshitij.so 59
            rank = rank + 1
60
 
13754 kshitij.so 61
 
62
def scrapeBestSellerTablets():
63
    global bestSellers
19645 kshitij.so 64
    rank = 1
65
    for i in range(0,100,10):
66
        url = "http://mobileapi.flipkart.net/3/discover/getSearch?store=tyy/hry&start=%d&count=10"%(i)
67
        print url
68
        response_data = fetchResponseUsingProxy(url, headers, livePricing=None, proxy=False, flipkart=False)
69
        input_json = json.loads(response_data)
70
        for identifier, data in (input_json['RESPONSE']['product']).iteritems():
20357 kshitij.so 71
            t_title = data['value']['titles']['title']
72
            t_subtitle = data['value']['titles']['subtitle']
73
            title = (xstr(t_title) + " " + xstr(t_subtitle)).strip()
74
            print title
19645 kshitij.so 75
            url =  (data['value']['smartUrl']).replace("http://dl.flipkart.com/dl", "http://www.flipkart.com")
20358 kshitij.so 76
            print url
77
            print identifier
78
            print "=================="
19645 kshitij.so 79
            r_info = __RankInfo(identifier, rank, 5, title, url)
13754 kshitij.so 80
            bestSellers.append(r_info)
19645 kshitij.so 81
            rank = rank + 1
13754 kshitij.so 82
 
19648 kshitij.so 83
def commitBestSellers():
14379 kshitij.so 84
    global exceptionList
13754 kshitij.so 85
    for x in bestSellers:
19648 kshitij.so 86
        col = get_mongo_connection(host=options.mongoHost).Catalog.MasterData.find({'identifier':x.identifier.strip()})
14379 kshitij.so 87
        if len(list(col)) == 0:
88
            exceptionList.append(x)
89
        else:
19648 kshitij.so 90
            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 91
 
92
def resetRanks(category):
19647 kshitij.so 93
    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 94
 
95
def sendEmail():
96
    message="""<html>
97
            <body>
98
            <h3>Flipkart Best Sellers not in master</h3>
99
            <table border="1" style="width:100%;">
100
            <thead>
101
            <tr><th>Identifier</th>
19645 kshitij.so 102
            <th>Title</th>
14379 kshitij.so 103
            <th>Category</th>
104
            <th>Rank</th>
19645 kshitij.so 105
            <th>URL</th>
14379 kshitij.so 106
            </tr></thead>
107
            <tbody>"""
108
    for item in exceptionList:
109
        message+="""<tr>
110
        <td style="text-align:center">"""+(item.identifier)+"""</td>
19645 kshitij.so 111
        <td style="text-align:center">"""+(item.title)+"""</td>
112
        <td style="text-align:center">"""+(CATEGORY_MAP.get(item.category))+"""</td>
14379 kshitij.so 113
        <td style="text-align:center">"""+str(item.rank)+"""</td>
19645 kshitij.so 114
        <td style="text-align:center">"""+(item.url)+"""</td>
14379 kshitij.so 115
        </tr>"""
116
    message+="""</tbody></table></body></html>"""
20358 kshitij.so 117
    message = message.decode('ascii', 'ignore')
19645 kshitij.so 118
    #recipients = ['kshitij.sood@saholic.com']
20172 aman.kumar 119
    recipients = ['rajneesh.arora@saholic.com','kshitij.sood@saholic.com','chaitnaya.vats@saholic.com','ritesh.chauhan@saholic.com','khushal.bhatia@saholic.com']
14379 kshitij.so 120
    msg = MIMEMultipart()
121
    msg['Subject'] = "Flipkart Best Sellers" + ' - ' + str(datetime.now())
122
    msg['From'] = ""
123
    msg['To'] = ",".join(recipients)
124
    msg.preamble = "Flipkart Best Sellers" + ' - ' + str(datetime.now())
125
    html_msg = MIMEText(message, 'html')
126
    msg.attach(html_msg)
127
 
128
    smtpServer = smtplib.SMTP('localhost')
129
    smtpServer.set_debuglevel(1)
130
    sender = 'dtr@shop2020.in'
131
    try:
132
        smtpServer.sendmail(sender, recipients, msg.as_string())
133
        print "Successfully sent email"
134
    except:
135
        print "Error: unable to send email."
13754 kshitij.so 136
 
137
def main():
138
    scrapeBestSellerMobiles()
19648 kshitij.so 139
    resetRanks(3)
13754 kshitij.so 140
    scrapeBestSellerTablets()
19648 kshitij.so 141
    resetRanks(5)
142
    commitBestSellers()
14379 kshitij.so 143
    sendEmail()
13754 kshitij.so 144
 
145
if __name__=='__main__':
146
    main()