def __init__(self,name,now_time="2015-09-30"): # self.share=Share(name) self.yahoo_data=[] self.yahoo_feature=[] self.name=name self.now_time = now_time self.DIR_NAME="../data/" self.C = CsvHandler(self.name,self.now_time)
def setUp(self):
self.h1, self.h2, self.h3, self.h4 = XmlHandler(), TxtHandler(), JsonHandler(), CsvHandler()
self.h1.set_next(self.h2)
self.h2.set_next(self.h3)
self.h3.set_next(self.h4)
self.files_dict = os.path.dirname(os.path.dirname(os.path.abspath(__file__))) + '/tests/test_data/'
self.output_file = self.files_dict + 'result.res'
open(self.output_file, 'w').close()
class DataProducer:
def __init__(self,name,now_time="2015-09-30"):
# self.share=Share(name)
self.yahoo_data=[]
self.yahoo_feature=[]
self.name=name
self.now_time = now_time
self.DIR_NAME="../data/"
self.C = CsvHandler(self.name,self.now_time)
#csvをチェック
#CsvHandler.get_datai
#csvからaggregatorでデータをとってくる(まだ加工前)
# target
# x_high:最高値の変化量
# x_low:最高値の変化量
# x_vol:取引量
def make_datemap(self,term,k):
#何日おきにつくるか
print"making datemap csv..."
N=1
self.yahoo_data=self.C.get_data(self.name)
size=len(self.yahoo_data["aver"])
obj = self.C.read_csv(self.name)
g=open(self.DIR_NAME+self.name+"_datemap.csv","w")
fieldnames=("Column","Start","End","TargetDay")
headers = dict( (n,n) for n in fieldnames )
writer=csv.DictWriter(g,fieldnames)
writer.writerow(headers)
count = 1
day=k
while (day+term)<=size:
row=""
start = obj[day-k]["Date"]
end = obj[day]["Date"]
targetday = obj[day+term-1]["Date"]
writer.writerow({"Column":count,"Start":start,"End":end,"TargetDay":targetday})
count+=1
day+=N
g.close()
print 'Saving datemap: DONE'
def make_train_data(self,term,k,per):
#何日おきにつくるか
N=1
self.yahoo_data=self.C.get_data(self.name)
print"making trainning data..."
size=len(self.yahoo_data["aver"])
col=size-term-k+1
#初期化 最初の時点
self.x_yahoo_data=np.zeros((col,3*(k+1)))
self.y_yahoo_data=np.zeros(col)
#print len(self.yahoo_data["aver"]),len(self.yahoo_data["low"]),len(self.yahoo_data["high"])
day=k
index=0
#targetをつくる term内にper%あがったら1
while (day+term)<=size:
maxaver=np.max(self.yahoo_data["aver"][day:day+term])
x_high=list(self.yahoo_data["high"][day-k:day+1])
x_low=list(self.yahoo_data["low"][day-k:day+1])
x_vol=list(self.yahoo_data["vol"][day-k:day+1])
self.x_yahoo_data[index]=x_high+x_low+x_vol
self.y_yahoo_data[index]=maxaver/self.yahoo_data["aver"][day]-1
day+=N
index+=1
output={"data":self.x_yahoo_data,"target":self.y_yahoo_data}
return output
def make_pickle(self,Data):
filename="../data/"+self.name+".pkl"
with open(filename,'wb') as output:
six.moves.cPickle.dump(Data,output,-1)
print 'Saving pickle: DONE'
def plot(self,filename="graph"):
plt.plot(self.yahoo_data["aver"])
#plt.show()
plt.savefig(filename+".jpg")
def setUp(self):
self.CsvHandler = CsvHandler()