from jqdata import finance
from sklearn import linear_model
from jqdata import *
import scipy.stats as stats
import numpy as np
import pandas as pd
import math
import time;

q= query(jy.LC_PerformanceForecast).order_by(jy.LC_PerformanceForecast.InfoPublDate.desc()).limit(3)
df = jy.run_query(q)

df.head(3)

def pct(start_date,stock,days): #计算股价涨跌幅
    start_date = datetime.datetime.strptime(start_date,'%Y-%m-%d %H:%M:%S')
    end_date = str(start_date + (datetime.timedelta(days)))
    p = get_price(stock,start_date = start_date,end_date = end_date)
    pct = p.close.pct_change().sum()
    return pct
#print pct_change('2019-01-25 00:00:00','000001.XSHE')

secu_category = [1] # 证券类别:1-A股，SecuMain表中的字段

incode = jy.run_query(query(
            jy.SecuMain.SecuCode,
            jy.SecuMain.CompanyCode,
        ).filter(
            jy.SecuMain.SecuCategory.in_(secu_category),
        ))
# print(incode)
q = query(
        jy.LC_PerformanceForecast
    ).filter(
        jy.LC_PerformanceForecast.CompanyCode.in_(incode['CompanyCode']), # 股票代码
        jy.LC_PerformanceForecast.ForcastType == 8, #8为公告为eps预减
    ).order_by(jy.LC_PerformanceForecast.InfoPublDate.desc() # 公告发布日期
    ).limit(3000)

df = jy.run_query(q).fillna(value=0, method=None, axis=0) 
df2 = pd.merge(incode, df, on='CompanyCode')
DF = df2[['SecuCode','InfoPublDate','ResultStatement']]
DF = DF.copy()
DF['date'] = DF.T.apply(lambda x:(str(x.InfoPublDate).split('+')[0]))
DF['code'] = DF.T.apply(lambda x:normalize_code(x.SecuCode))
DF['pct5'] = DF.T.apply(lambda x: pct(x.date,x.code,5)) #5天（非交易日）之后股价涨跌幅
DF['pct30'] = DF.T.apply(lambda x: pct(x.date,x.code,30)) #30天（非交易日）之后股价涨跌幅
DF['pct60'] = DF.T.apply(lambda x: pct(x.date,x.code,60)) #60天（非交易日）之后股价涨跌幅
DF = DF[(DF['pct30'] != 0)]
#print DF[['code','date','pct']]

print DF.pct5.describe() 
print DF.pct30.describe() 
print DF.pct60.describe()