class RetrivalAndRankingChooseController(RankingChooseController):
'''Use the retrival ad category firstly, and ranking the ads.
'''
def __init__(self):
self.load_conf()
getters = self.build_vec_getter()
self.ranking = Ranking(self.builder, self.distance, getters[1], getters[2])
self.ad_c_cache, self.ad_cache = self.loader.load_ads()
self.number_content()
self.retrival = CategoryRetrivalStrategyCenter(self.builder, self.distance, getters[0], getters[2])
def get_content_ads(self, num):
'''Get the content whose id is num and the recommend ads
Args:
num : content id
Returns:
(content, [list of ad content])
'''
if num >= self.count:
return (('', ''), [])
print num, self.content_key[num]
content_obj = self.loader.load_content(self.content_key[num])
categorys = self.retrival.get_category(content_obj, self.ad_c_cache)
retrival_ads = []
print 'categorys:'
for category in categorys:
print category.name
retrival_ads += self.ad_cache.get_ads(category)
ads = self.ranking.ranking(content_obj, retrival_ads, topk = 2)
res_ad = []
for score, ad_key in ads:
res_ad.append(self.loader.get_ad_info(ad_key))
return (self.loader.get_content_info(self.content_key[num]), res_ad)
class RankingChooseController(ChooseController):
'''Just use the ranking process.
'''
def __init__(self):
self.load_conf()
getters = self.build_vec_getter()
self.ranking = Ranking(self.builder, self.distance, getters[1], getters[2])
self.ad_cache = self.loader.load_ads()[1]
self.number_content()
def load_conf(self):
'''Read config file and build builder, distance and loader'''
cf = ConfigParser.RawConfigParser(allow_no_value=True)
import os
paths = ['controllers', 'recommend', 'controller.cfg']
_file = os.getcwd()
for path in paths:
_file = os.path.join(_file, path)
with open(_file,'r') as configfile:
cf.readfp(configfile)
builder_class = cf.get("class", 'builder')
builder_args = eval(cf.get("args", 'builder'))
self.builder = getattr(builder, builder_class)(args = builder_args)
distance_class = cf.get("class", 'distance')
distance_args = eval(cf.get("args", 'distance'))
self.distance = getattr(distance, distance_class)(args = distance_args)
loader_class = cf.get("class", 'loader')
loader_args = eval(cf.get("args", 'loader'))
self.loader = getattr(loader, loader_class)(args = loader_args)
def build_vec_getter(self):
'''Build the ad and content vector getter
Returns:
tuple of ad vector getter and content vector getter
'''
ad_c_vec_getter = lambda vec : vec.get_vec()
ad_vec_getter = lambda vec : vec.get_vec()
content_vec_getter = lambda vec : vec.get_vec()
return (ad_c_vec_getter, ad_vec_getter, content_vec_getter)
def number_content(self):
'''Give all content number and save the key'''
self.count = 8
self.content_key = { 0 : {'category' : 'IT', 'file' : '10.txt'},
1 : {'category' : 'IT', 'file' : '11.txt'},
2 : {'category' : 'IT', 'file' : '18.txt'},
3 : {'category' : '体育', 'file' : '12.txt'},
4 : {'category' : '体育', 'file' : '13.txt'},
5 : {'category' : '体育', 'file' : '14.txt'},
6 : {'category' : '体育', 'file' : '16.txt'},
7 : {'category' : '体育', 'file' : '17.txt'}
}
def content_count(self):
return self.count
def get_content_ads(self, num):
'''Get the content whose id is num and the recommend ads
Args:
num : content id
Returns:
(content, [list of ad content])
'''
if num >= self.count:
return (('', ''), [])
content_obj = self.loader.load_content(self.content_key[num])
ads = self.ranking.ranking(content_obj, self.ad_cache, topk = 2)
res_ad = []
for score, ad_key in ads:
res_ad.append(self.loader.get_ad_info(ad_key))
return (self.loader.get_content_info(self.content_key[num]), res_ad)