| 13829 |
kshitij.so |
1 |
import urllib2
|
|
|
2 |
import simplejson as json
|
|
|
3 |
import pymongo
|
|
|
4 |
from dtr.utils.utils import to_java_date
|
| 13918 |
kshitij.so |
5 |
from datetime import datetime, timedelta
|
| 13829 |
kshitij.so |
6 |
from operator import itemgetter
|
|
|
7 |
from dtr.utils.AmazonPriceOnlyScraper import AmazonScraper
|
| 14121 |
kshitij.so |
8 |
from dtr.utils import FlipkartScraper,NewFlipkartScraper
|
| 13829 |
kshitij.so |
9 |
|
|
|
10 |
con = None
|
|
|
11 |
SOURCE_MAP = {'AMAZON':1,'FLIPKART':2,'SNAPDEAL':3}
|
|
|
12 |
scraperFk = FlipkartScraper.FlipkartScraper()
|
|
|
13 |
scraperAmazon = AmazonScraper()
|
| 14121 |
kshitij.so |
14 |
scraperProductPage = NewFlipkartScraper.FlipkartProductPageScraper()
|
| 13829 |
kshitij.so |
15 |
|
|
|
16 |
headers = {
|
|
|
17 |
'User-agent':'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/535.11 (KHTML, like Gecko) Chrome/17.0.963.56 Safari/535.11',
|
|
|
18 |
'Accept' : 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
|
|
|
19 |
'Accept-Language' : 'en-US,en;q=0.8',
|
|
|
20 |
'Accept-Charset' : 'ISO-8859-1,utf-8;q=0.7,*;q=0.3'
|
|
|
21 |
}
|
|
|
22 |
|
|
|
23 |
def get_mongo_connection(host='localhost', port=27017):
|
|
|
24 |
global con
|
|
|
25 |
if con is None:
|
|
|
26 |
print "Establishing connection %s host and port %d" %(host,port)
|
|
|
27 |
try:
|
|
|
28 |
con = pymongo.MongoClient(host, port)
|
|
|
29 |
except Exception, e:
|
|
|
30 |
print e
|
|
|
31 |
return None
|
|
|
32 |
return con
|
|
|
33 |
|
|
|
34 |
def returnLatestPrice(data, source_id):
|
| 13918 |
kshitij.so |
35 |
now = datetime.now()
|
| 13829 |
kshitij.so |
36 |
if source_id == 1:
|
|
|
37 |
try:
|
| 13918 |
kshitij.so |
38 |
if data['identifier'] is None or len(data['identifier'].strip())==0:
|
| 13829 |
kshitij.so |
39 |
return {}
|
| 13918 |
kshitij.so |
40 |
|
|
|
41 |
try:
|
| 13975 |
kshitij.so |
42 |
if data['priceUpdatedOn'] > to_java_date(now - timedelta(minutes=5)):
|
| 13918 |
kshitij.so |
43 |
print "sku id is already updated",data['_id']
|
| 13920 |
kshitij.so |
44 |
return {'_id':data['_id'],'available_price':data['available_price'],'in_stock':data['in_stock'],'source_id':1,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
| 13918 |
kshitij.so |
45 |
except:
|
|
|
46 |
pass
|
|
|
47 |
|
|
|
48 |
|
| 13829 |
kshitij.so |
49 |
url = "http://www.amazon.in/gp/offer-listing/%s/ref=olp_sort_ps"%(data['identifier'])
|
|
|
50 |
lowestPrice = 0.0
|
|
|
51 |
lowestPrice = scraperAmazon.read(url)
|
| 13929 |
kshitij.so |
52 |
print "LowestPrice ",lowestPrice
|
| 13829 |
kshitij.so |
53 |
inStock = 0
|
|
|
54 |
if lowestPrice > 0:
|
|
|
55 |
inStock = 1
|
|
|
56 |
if lowestPrice > 0:
|
| 13971 |
kshitij.so |
57 |
get_mongo_connection().Catalog.MasterData.update({'_id':data['_id']}, {'$set' : {'available_price':lowestPrice,'updatedOn':to_java_date(now),'priceUpdatedOn':to_java_date(now),'in_stock':inStock}}, multi=True)
|
| 13918 |
kshitij.so |
58 |
get_mongo_connection().Catalog.Deals.update({'_id':data['_id']}, {'$set' : {'available_price':lowestPrice , 'in_stock':inStock}}, multi=True)
|
| 13829 |
kshitij.so |
59 |
else:
|
| 13929 |
kshitij.so |
60 |
lowestPrice = data['available_price']
|
| 13977 |
kshitij.so |
61 |
get_mongo_connection().Catalog.MasterData.update({'_id':data['_id']}, {'$set' : {'updatedOn':to_java_date(now),'in_stock':0,'priceUpdatedOn':to_java_date(now)}}, multi=True)
|
| 13918 |
kshitij.so |
62 |
get_mongo_connection().Catalog.Deals.update({'_id':data['_id']}, {'$set' : {'in_stock':0}}, multi=True)
|
|
|
63 |
|
|
|
64 |
try:
|
|
|
65 |
recomputeDeal(data['skuBundleId'])
|
|
|
66 |
except:
|
|
|
67 |
print "Unable to compute deal for ",data['skuBundleId']
|
|
|
68 |
|
| 13829 |
kshitij.so |
69 |
return {'_id':data['_id'],'available_price':lowestPrice,'in_stock':inStock,'source_id':1,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
|
|
70 |
except:
|
| 13920 |
kshitij.so |
71 |
return {'_id':data['_id'],'available_price':data['available_price'],'in_stock':data['in_stock'],'source_id':1,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
| 13829 |
kshitij.so |
72 |
|
|
|
73 |
elif source_id ==3:
|
|
|
74 |
try:
|
| 13918 |
kshitij.so |
75 |
if data['identifier'] is None or len(data['identifier'].strip())==0:
|
| 13829 |
kshitij.so |
76 |
return {}
|
| 13918 |
kshitij.so |
77 |
|
|
|
78 |
try:
|
| 13975 |
kshitij.so |
79 |
if data['priceUpdatedOn'] > to_java_date(now - timedelta(minutes=5)):
|
| 13920 |
kshitij.so |
80 |
print "sku id is already updated",data['_id']
|
|
|
81 |
return {'_id':data['_id'],'available_price':data['available_price'],'in_stock':data['in_stock'],'source_id':3,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
| 13919 |
kshitij.so |
82 |
|
|
|
83 |
except Exception as e:
|
| 13920 |
kshitij.so |
84 |
print "Exception snapdeal"
|
| 13919 |
kshitij.so |
85 |
print e
|
| 13918 |
kshitij.so |
86 |
pass
|
|
|
87 |
|
|
|
88 |
|
| 13829 |
kshitij.so |
89 |
url="http://www.snapdeal.com/acors/json/gvbps?supc=%s&catId=175&sort=sellingPrice"%(data['identifier'])
|
|
|
90 |
req = urllib2.Request(url,headers=headers)
|
|
|
91 |
response = urllib2.urlopen(req)
|
|
|
92 |
json_input = response.read()
|
|
|
93 |
vendorInfo = json.loads(json_input)
|
|
|
94 |
lowestOfferPrice = 0
|
|
|
95 |
inStock = 0
|
|
|
96 |
for vendor in vendorInfo:
|
|
|
97 |
lowestOfferPrice = float(vendor['sellingPrice'])
|
|
|
98 |
stock = vendor['buyableInventory']
|
|
|
99 |
if stock > 0 and lowestOfferPrice > 0:
|
|
|
100 |
inStock = 1
|
|
|
101 |
break
|
|
|
102 |
|
|
|
103 |
print lowestOfferPrice
|
|
|
104 |
print inStock
|
|
|
105 |
print "*************"
|
|
|
106 |
if inStock == 1:
|
| 13971 |
kshitij.so |
107 |
get_mongo_connection().Catalog.MasterData.update({'_id':data['_id']}, {'$set' : {'available_price':lowestOfferPrice,'updatedOn':to_java_date(now),'priceUpdatedOn':to_java_date(now),'in_stock':inStock}}, multi=True)
|
| 13918 |
kshitij.so |
108 |
get_mongo_connection().Catalog.Deals.update({'_id':data['_id']}, {'$set' : {'available_price':lowestOfferPrice , 'in_stock':inStock}}, multi=True)
|
| 13829 |
kshitij.so |
109 |
else:
|
| 13929 |
kshitij.so |
110 |
lowestOfferPrice = data['available_price']
|
| 13977 |
kshitij.so |
111 |
get_mongo_connection().Catalog.MasterData.update({'_id':data['_id']}, {'$set' : {'updatedOn':to_java_date(now),'in_stock':inStock,'priceUpdatedOn':to_java_date(now)}}, multi=True)
|
| 13918 |
kshitij.so |
112 |
get_mongo_connection().Catalog.Deals.update({'_id':data['_id']}, {'$set' : {'in_stock':inStock}}, multi=True)
|
|
|
113 |
|
|
|
114 |
try:
|
|
|
115 |
recomputeDeal(data['skuBundleId'])
|
|
|
116 |
except:
|
|
|
117 |
print "Unable to compute deal for ",data['skuBundleId']
|
|
|
118 |
|
| 13829 |
kshitij.so |
119 |
return {'_id':data['_id'],'available_price':lowestOfferPrice,'in_stock':inStock,'source_id':3,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
|
|
120 |
except:
|
| 13920 |
kshitij.so |
121 |
return {'_id':data['_id'],'available_price':data['available_price'],'in_stock':data['in_stock'],'source_id':3,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
| 13829 |
kshitij.so |
122 |
|
|
|
123 |
elif source_id == 2:
|
|
|
124 |
try:
|
| 13918 |
kshitij.so |
125 |
if data['identifier'] is None or len(data['identifier'].strip())==0:
|
| 13829 |
kshitij.so |
126 |
return {}
|
| 13918 |
kshitij.so |
127 |
|
|
|
128 |
try:
|
| 13975 |
kshitij.so |
129 |
if data['priceUpdatedOn'] > to_java_date(now - timedelta(minutes=5)):
|
| 13918 |
kshitij.so |
130 |
print "sku id is already updated",data['_id']
|
| 13920 |
kshitij.so |
131 |
return {'_id':data['_id'],'available_price':data['available_price'],'in_stock':data['in_stock'],'source_id':2,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
| 13918 |
kshitij.so |
132 |
except:
|
|
|
133 |
pass
|
|
|
134 |
|
|
|
135 |
lowestSp = 0
|
|
|
136 |
inStock = 0
|
| 14121 |
kshitij.so |
137 |
try:
|
|
|
138 |
if data['marketPlaceUrl']!="" or data['marketPlaceUrl'] !="http://www.flipkart.com/ps/%s"%(data['identifier']):
|
|
|
139 |
result = scraperProductPage.read(data['marketPlaceUrl'])
|
|
|
140 |
if result.get('lowestPrice')!=0:
|
|
|
141 |
lowestSp = result.get('lowestPrice')
|
|
|
142 |
inStock = result.get('inStock')
|
|
|
143 |
except:
|
|
|
144 |
print "Unable to scrape product page ",data['identifier']
|
|
|
145 |
if lowestSp ==0:
|
|
|
146 |
url = "http://www.flipkart.com/ps/%s"%(data['identifier'])
|
|
|
147 |
vendorsData = scraperFk.read(url)
|
|
|
148 |
sortedVendorsData = []
|
|
|
149 |
sortedVendorsData = sorted(vendorsData, key=itemgetter('sellingPrice'))
|
|
|
150 |
print "data",sortedVendorsData
|
|
|
151 |
for vData in sortedVendorsData:
|
|
|
152 |
lowestSp = vData['sellingPrice']
|
|
|
153 |
break
|
|
|
154 |
if lowestSp > 0:
|
|
|
155 |
inStock = 1
|
| 13829 |
kshitij.so |
156 |
print lowestSp
|
|
|
157 |
print inStock
|
|
|
158 |
if lowestSp > 0:
|
| 13971 |
kshitij.so |
159 |
get_mongo_connection().Catalog.MasterData.update({'_id':data['_id']}, {'$set' : {'available_price':lowestSp,'updatedOn':to_java_date(now),'priceUpdatedOn':to_java_date(now),'in_stock':inStock}}, multi=True)
|
| 13918 |
kshitij.so |
160 |
get_mongo_connection().Catalog.Deals.update({'_id':data['_id']}, {'$set' : {'available_price':lowestSp , 'in_stock':inStock}}, multi=True)
|
| 13829 |
kshitij.so |
161 |
else:
|
| 13929 |
kshitij.so |
162 |
lowestSp = data['available_price']
|
| 13977 |
kshitij.so |
163 |
get_mongo_connection().Catalog.MasterData.update({'_id':data['_id']}, {'$set' : {'updatedOn':to_java_date(now),'in_stock':inStock,'priceUpdatedOn':to_java_date(now)}}, multi=True)
|
| 13918 |
kshitij.so |
164 |
get_mongo_connection().Catalog.Deals.update({'_id':data['_id']}, {'$set' : {'in_stock':inStock}}, multi=True)
|
|
|
165 |
|
|
|
166 |
|
|
|
167 |
try:
|
|
|
168 |
recomputeDeal(data['skuBundleId'])
|
|
|
169 |
except:
|
|
|
170 |
print "Unable to compute deal for ",data['skuBundleId']
|
|
|
171 |
|
| 13829 |
kshitij.so |
172 |
return {'_id':data['_id'],'available_price':lowestSp,'in_stock':inStock,'source_id':2,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
|
|
173 |
except:
|
| 13920 |
kshitij.so |
174 |
return {'_id':data['_id'],'available_price':data['available_price'],'in_stock':data['in_stock'],'source_id':2,'source_product_name':data['source_product_name'],'marketPlaceUrl':data['marketPlaceUrl'],'thumbnail':data['thumbnail']}
|
| 13829 |
kshitij.so |
175 |
else:
|
|
|
176 |
return {}
|
|
|
177 |
|
| 13918 |
kshitij.so |
178 |
|
|
|
179 |
def recomputeDeal(skuBundleId):
|
|
|
180 |
"""Lets recompute deal for this bundle"""
|
|
|
181 |
print "Recomputing for bundleId",skuBundleId
|
| 13829 |
kshitij.so |
182 |
|
| 13918 |
kshitij.so |
183 |
similarItems = list(get_mongo_connection().Catalog.Deals.find({'skuBundleId':skuBundleId}).sort([('available_price',pymongo.ASCENDING)]))
|
|
|
184 |
bestPrice = float("inf")
|
|
|
185 |
bestOne = None
|
|
|
186 |
bestSellerPoints = 0
|
|
|
187 |
toUpdate = []
|
|
|
188 |
for similarItem in similarItems:
|
| 13971 |
kshitij.so |
189 |
if similarItem['in_stock'] == 0 or similarItem['maxprice'] is None or similarItem['maxprice'] < similarItem['available_price']:
|
| 13918 |
kshitij.so |
190 |
get_mongo_connection().Catalog.Deals.update({ '_id' : similarItem['_id'] }, {'$set':{'showDeal':0 }})
|
|
|
191 |
continue
|
|
|
192 |
if similarItem['available_price'] < bestPrice:
|
|
|
193 |
bestOne = similarItem
|
|
|
194 |
bestPrice = similarItem['available_price']
|
|
|
195 |
bestSellerPoints = similarItem['bestSellerPoints']
|
|
|
196 |
elif similarItem['available_price'] == bestPrice and bestSellerPoints < similarItem['bestSellerPoints']:
|
|
|
197 |
bestOne = similarItem
|
|
|
198 |
bestPrice = similarItem['available_price']
|
|
|
199 |
bestSellerPoints = similarItem['bestSellerPoints']
|
|
|
200 |
else:
|
|
|
201 |
pass
|
|
|
202 |
if bestOne is not None:
|
|
|
203 |
for similarItem in similarItems:
|
|
|
204 |
toUpdate.append(similarItem['_id'])
|
|
|
205 |
toUpdate.remove(bestOne['_id'])
|
|
|
206 |
get_mongo_connection().Catalog.Deals.update({ '_id' : bestOne['_id'] }, {'$set':{'showDeal':1 }})
|
|
|
207 |
if len(toUpdate) > 0:
|
|
|
208 |
get_mongo_connection().Catalog.Deals.update({ '_id' : { "$in": toUpdate } }, {'$set':{'showDeal':0 }},upsert=False, multi=True)
|
| 13829 |
kshitij.so |
209 |
|
|
|
210 |
def getLatestPrice(skuBundleId, source_id):
|
|
|
211 |
temp = []
|
|
|
212 |
itemIds = list(get_mongo_connection().Catalog.MasterData.find({'skuBundleId':skuBundleId,'source_id' : source_id}))
|
|
|
213 |
for item in itemIds:
|
|
|
214 |
temp.append(returnLatestPrice(item, source_id))
|
|
|
215 |
return temp
|
|
|
216 |
|
| 13864 |
kshitij.so |
217 |
def getLatestPriceById(id):
|
|
|
218 |
item = list(get_mongo_connection().Catalog.MasterData.find({'_id':id}))
|
| 13866 |
kshitij.so |
219 |
return returnLatestPrice(item[0], item[0]['source_id'])
|
| 13829 |
kshitij.so |
220 |
|
|
|
221 |
|
|
|
222 |
def main():
|
| 13929 |
kshitij.so |
223 |
print getLatestPriceById(8850)
|
| 13829 |
kshitij.so |
224 |
|
|
|
225 |
if __name__=='__main__':
|
| 13971 |
kshitij.so |
226 |
main()
|
|
|
227 |
|
|
|
228 |
"""21.06$"""
|