def calculatebmi(listpr):
c = Context()
for p in listpr:
for timeattr in p.getalltimeattribute():
if timeattr.getattribute("HEIGHT") is not None and timeattr.getattribute("WEIGHT") is not None:
bmi = c.divide(timeattr.getattribute("HEIGHT"), c.power(timeattr.getattribute("WEIGHT"), Decimal("2")))
timeattr.setattribute("BMI", bmi)
def solve():
mod, cnt, num1, num2 = 10 ** 9, 2, 1, 1
x, y = (
Decimal(1) / Decimal(5).sqrt(),
(Decimal(1) + Decimal(5).sqrt()) / Decimal(2)
)
_context = Context()
while True:
num1, num2 = num2, (num1 + num2) % mod
_set = set(str(num1))
if '0' not in _set and len(_set) == 9:
_tmp = x.log10() + cnt * y.log10()
_num = int(_context.power(10, _tmp - int(_tmp) + 8))
_set = set(str(_num))
if '0' not in _set and len(_set) == 9:
return cnt
cnt += 1
def tahara_features(classify="dropout", is_classify="dropout", is_80=False):
u"""田原アルゴリズムに基づいて特徴語を抽出し、ユーザーベクトルを作成する
Args:
calassify:適用するユーザーのクラス
is_classify:どのユーザークラスらしさでベクトルを作成するか
is_80:上位80語制限
"""
import pymongo
from decimal import Decimal,Context
import math
from function_common import get_word_point
context=Context()
count=0
if classify == "dropout":
profiles=load_from_mongo(config.m_dropout.db,config.m_dropout.profile_coll)
elif classify=="normal":
profiles=load_from_mongo(config.m_normal.db,config.m_normal.profile_coll)
elif classify=="dropout_test":
profiles=load_from_mongo(config.m_dropout_test.db,config.m_dropout_test.profile_coll)
print len(profiles)
elif classify=="normal_test":
profiles=load_from_mongo(config.m_normal_test.db,config.m_normal_test.profile_coll)
print len(profiles)
else:
print "unknown classify ",classify
return
if is_classify =="dropout":
dropouts=load_from_mongo(config.m_dropout.db,config.m_dropout.tahara_coll,True)
names=[]
for d in dropouts.sort("rate",pymongo.DESCENDING):
count+=1
names.append(d["_id"])
if count >=1000:
break
if(is_80==True):
print "name,dropout,{0}".format(
",".join([str(n) for n in names if get_word_point(n) > 0]))
else:
print "name,dropout,{0}".format(",".join([str(n) for n in names]))
for profile in profiles:
if classify == "dropout" or classify == "normal":
wui=load_from_mongo("db","users_of_tehara",return_cursor=False,criteria={'_id':profile['_id']})
elif classify == "dropout_test" or classify == "normal_test":
wui=load_from_mongo("db","users_of_tehara_test",return_cursor=False,criteria={'_id':profile['_id']})
else:
print "unknown classify ",classify
return
if (wui is None) or (len(wui)==0):
continue
words= wui[0][u'words']
wui_wci=0
w2ui=0
w2ci=0
if is_classify =="dropout":
dropouts=load_from_mongo(config.m_dropout.db,config.m_dropout.tahara_coll,True)
else:
dropouts=load_from_mongo(config.m_normal.db,config.m_normal.tahara_coll,True)
if classify=="dropout":
tweets=load_from_mongo(config.m_dropout.db,config.m_dropout.tweets_coll,criteria={'$and':[{"user.screen_name":profile["_id"]},{'text':{'$regex':'^[^RT]'}},{'created_at':{'$gte':profile["date_of_start"],'$lte':profile["date_of_dropout"]}}]})
if len(tweets)==0:
tweets=load_from_mongo(config.m_dropout.db,"tweets_of_dropout?",criteria={'$and':[{"user.screen_name":profile["_id"]},{'text':{'$regex':'^[^RT]'}},{'created_at':{'$gte':profile["date_of_start"],'$lte':profile["date_of_dropout"]}}]})
tmp_sim_dropouts=[]
tmp_sim_dropouts.append(profile["_id"])
tmp_sim_dropouts.append(-1)
elif classify == "normal":
tweets=load_from_mongo(config.m_normal.db,config.m_normal.tweets_coll,criteria={'$and':[{"user.screen_name":profile["_id"]},{'text':{'$regex':'^[^RT]'}},{'created_at':{'$gte':profile["date_of_start"],'$lte':profile["date_of_dropout"]}}]})
tmp_sim_dropouts=[]
tmp_sim_dropouts.append(profile["_id"])
tmp_sim_dropouts.append(1)
elif classify == "dropout_test":
tweets=load_from_mongo(config.m_dropout_test.db,config.m_dropout_test.tweets_coll,criteria={'$and':[{"user.screen_name":profile["_id"]},{'text':{'$regex':'^[^RT]'}},{'created_at':{'$gte':profile["date_of_start"],'$lte':profile["date_of_dropout"]}}]})
if len(tweets)==0:
tweets=load_from_mongo(config.m_dropout_test.db,"tweets_of_dropout?",criteria={'$and':[{"user.screen_name":profile["_id"]},{'text':{'$regex':'^[^RT]'}},{'created_at':{'$gte':profile["date_of_start"],'$lte':profile["date_of_dropout"]}}]})
tmp_sim_dropouts=[]
tmp_sim_dropouts.append(profile["_id"])
tmp_sim_dropouts.append(-1)
elif classify == "normal_test":
tweets=load_from_mongo(config.m_normal_test.db,config.m_normal_test.tweets_coll,criteria={'$and':[{"user.screen_name":profile["_id"]},{'text':{'$regex':'^[^RT]'}},{'created_at':{'$gte':profile["date_of_start"],'$lte':profile["date_of_dropout"]}}]})
tmp_sim_dropouts=[]
tmp_sim_dropouts.append(profile["_id"])
tmp_sim_dropouts.append(-1)
# 実際につぶやきがあった日数をカウント
day_counter=set()
for tweet in tweets:
y=extract_time(tweet["created_at"],"Y")
m=extract_time(tweet["created_at"],"M")
d=extract_time(tweet["created_at"],"D")
days=y,":",m,":",d
day_counter.add(days)
total_days=len(day_counter)
if total_days == 0:
print "Error! ",profile["_id"],"has 0 tweets."
continue
tmp_sim_dropouts=[]
count=0
for d in dropouts.sort("rate",pymongo.DESCENDING):
count+=1
word=[w for w in words if w["name"] == d["_id"]]
if len(word)==0:
word=[{"count":0,"name":d["_id"],"days":0}]
else:
pass
cnt=Decimal(word[0]["count"])
rate=Decimal(d["rate"])
days=Decimal(word[0]["days"])
d_point=Decimal(get_word_point(d["_id"]))
if d_point == 0:
if is_classify=="dropout":
if count > 1000:
if is_80==True:
break
else:
pass
else:
if is_80==True:
continue
else:
pass
td=Decimal(total_days)
if is_classify=="dropout":
tmp_wui_wci=Decimal(days/td)
wui_wci+=tmp_wui_wci
tmp_w2ui=context.power(cnt,2)
w2ui+=tmp_w2ui
tmp_w2ci=context.power(rate,2)
w2ci+=tmp_w2ci
else:
tmp_wui_wci=Decimal(cnt*rate)
wui_wci+=(cnt*rate)
tmp_w2ui=Decimal(math.pow((Decimal(cnt)),2))
w2ui+=math.pow((cnt),2)
tmp_w2ci=Decimal(math.pow(Decimal(rate),2))
w2ci+=math.pow(rate,2)
if count >= 1000:
break
if tmp_wui_wci == 0 or tmp_w2ci == 0 or tmp_w2ui == 0 or cnt ==0:
tmp_sim_dropouts.append(0)
else:
tmp_sim_dropout=tmp_wui_wci
tmp_sim_dropouts.append(float(tmp_sim_dropout))
print profile["_id"],",True,{0}".format(
",".join([str(t) for t in tmp_sim_dropouts])
)
