def test_gta():
from jqdata import gta, query
df = gta.run_query(query(gta.STK_STOCKINFO))
assert len(df.index) == 3000
df = gta.run_query(gta.query(gta.STK_STOCKINFO).limit(10))
assert isinstance(df, pd.DataFrame)
assert len(df.index) == 10
pass
def dividend_rate(stock, end_date=None, paymentdate=2015, start_date=None, count=1, date=None, no_data_return=NaN, skip_paused=False):
def get_div(df, paymentdate):
df.PAYMENTDATE = pd.to_datetime(df.PAYMENTDATE)
try:
if len(df)>0:
div = df[[x.year == (int(paymentdate)+1) for x in df.PAYMENTDATE]]['DIVIDENTBT'].iloc[-1]
return float(div)
else:
return no_data_return
except:
return no_data_return
symbol = stock[:6]
df = gta.run_query(query(
gta.STK_MKT_DIVIDENT.SYMBOL,
gta.STK_MKT_DIVIDENT.PAYMENTDATE,
gta.STK_MKT_DIVIDENT.DIVIDENTBT,
).filter(gta.STK_MKT_DIVIDENT.SYMBOL.in_([symbol])
).order_by(gta.STK_MKT_DIVIDENT.PAYMENTDATE))
div = get_div(df, paymentdate)
from math import isnan
if isnan(div) == True:
div = get_div(df, paymentdate-1)
df = get_price(stock, start_date=start_date, count=count, end_date=end_date, fields=['close'], frequency='daily', fq=None)
df['dividend_rate'] = float(div)/df['close']*100
df['dividend_rate'] = df['dividend_rate'].round(4)
del df['close']
df.index = [stock]
return df
def fun_get_Divid_by_year(self, context, stocks):
year = context.current_dt.year - 1
#将当前股票池转换为国泰安的6位股票池
stocks_symbol = []
for s in stocks:
stocks_symbol.append(s[0:6])
df = gta.run_query(
query(
gta.STK_DIVIDEND.SYMBOL, # 股票代码
gta.STK_DIVIDEND.DECLAREDATE, # 分红消息的时间
).filter(
gta.STK_DIVIDEND.ISDIVIDEND == 'Y', #有分红的股票
gta.STK_DIVIDEND.DIVDENDYEAR == year,
gta.STK_DIVIDEND.TERMCODE == 'P2702', # 年度分红
gta.STK_DIVIDEND.SYMBOL.in_(stocks_symbol))).fillna(
value=0, method=None, axis=0)
# 转换时间格式
df['pubtime'] = map(
lambda x: int(x.split('-')[0] + x.split('-')[1] + x.split('-')[2]),
df['DECLAREDATE'])
# 取得当前时间
currenttime = int(
str(context.current_dt)[0:4] + str(context.current_dt)[5:7] +
str(context.current_dt)[8:10])
# 选择在当前时间能看到的记录
df = df[(df.pubtime < currenttime)]
# 得到目前看起来，有上一年度年度分红的股票
stocks_symbol_this_year = list(df['SYMBOL'])
# 得到目前看起来，上一年度没有年度分红的股票
stocks_symbol_past_year = list(
set(stocks_symbol) - set(stocks_symbol_this_year))
# 查有上一年度年度分红的
df1 = gta.run_query(
query(
gta.STK_DIVIDEND.SYMBOL, # 股票代码
gta.STK_DIVIDEND.DIVIDENTBT, # 股票分红
gta.STK_DIVIDEND.DECLAREDATE, # 分红消息的时间
gta.STK_DIVIDEND.DISTRIBUTIONBASESHARES # 分红时的股本基数
).filter(
gta.STK_DIVIDEND.ISDIVIDEND == 'Y', #有分红的股票
gta.STK_DIVIDEND.DIVDENDYEAR == year,
gta.STK_DIVIDEND.SYMBOL.in_(stocks_symbol_this_year))).fillna(
value=0, method=None, axis=0)
df1['pubtime'] = map(
lambda x: int(x.split('-')[0] + x.split('-')[1] + x.split('-')[2]),
df1['DECLAREDATE'])
currenttime = int(
str(context.current_dt)[0:4] + str(context.current_dt)[5:7] +
str(context.current_dt)[8:10])
df1 = df1[(df1.pubtime < currenttime)]
# 求上上年的年度分红
df2 = gta.run_query(
query(
gta.STK_DIVIDEND.SYMBOL, # 股票代码
gta.STK_DIVIDEND.DIVIDENTBT, # 股票分红
gta.STK_DIVIDEND.DECLAREDATE, # 分红消息的时间
gta.STK_DIVIDEND.DISTRIBUTIONBASESHARES # 分红时的股本基数
).filter(
gta.STK_DIVIDEND.ISDIVIDEND == 'Y', #有分红的股票
gta.STK_DIVIDEND.DIVDENDYEAR == (year - 1),
gta.STK_DIVIDEND.SYMBOL.in_(stocks_symbol_past_year))).fillna(
value=0, method=None, axis=0)
df2['pubtime'] = map(
lambda x: int(x.split('-')[0] + x.split('-')[1] + x.split('-')[2]),
df2['DECLAREDATE'])
currenttime = int(
str(context.current_dt)[0:4] + str(context.current_dt)[5:7] +
str(context.current_dt)[8:10])
df2 = df2[(df2.pubtime < currenttime)]
df = pd.concat((df2, df1))
df['SYMBOL'] = map(normalize_code, list(df['SYMBOL']))
df.index = list(df['SYMBOL'])
# 获取最新股本
q = query(valuation.code, valuation.capitalization).filter(
valuation.code.in_(list(df.index)))
df3 = get_fundamentals(q).fillna(value=0)
df3['SYMBOL'] = df3['code']
df3 = df3.drop(['code'], axis=1)
# 合并成一个 dataframe
df = df.merge(df3, on='SYMBOL')
df.index = list(df['SYMBOL'])
# 转换成 float
df['DISTRIBUTIONBASESHARES'] = map(float, df['DISTRIBUTIONBASESHARES'])
# 计算股份比值
df['CAP_RATIO'] = df['DISTRIBUTIONBASESHARES'] / (
df['capitalization'] * 10000)
df['DIVIDENTBT'] = map(float, df['DIVIDENTBT'])
# 计算相对于目前股份而言的分红额度
df['DIVIDENTBT'] = df['DIVIDENTBT'] * df['CAP_RATIO']
df = df.drop([
'SYMBOL', 'DECLAREDATE', 'DISTRIBUTIONBASESHARES',
'capitalization', 'CAP_RATIO'
],
axis=1)
#接下来这一步是考虑多次分红的股票，因此需要累加股票的多次分红
df = df.groupby(df.index).sum()
#得到当前股价
Price = history(1,
unit='1d',
field='close',
security_list=list(df.index),
df=True,
skip_paused=False,
fq='pre')
Price = Price.T
df['pre_close'] = Price
#计算股息率 = 股息/股票价格，* 10 是因为取到的是每 10 股分红
df['divpercent'] = df['DIVIDENTBT'] / (df['pre_close'] * 10)
df['code'] = np.array(df.index)
return df
def getDivid(context, stocks, year_watch=3):
year = context.current_dt.year - 1
#now = datetime.now()
#year = now.year-1
#将当前股票池转换为国泰安的6位股票池
stocks_symbol = []
for s in stocks:
stocks_symbol.append(s[0:6])
# 累加3年分红信息
# 如果知道前一年的分红，那么得到前一年的分红数据
df1 = gta.run_query(
query(
gta.STK_DIVIDEND.SYMBOL, #股票代码
gta.STK_DIVIDEND.DIVIDENTBT, #股票分红
gta.STK_DIVIDEND.TOTALDIVIDENDDISTRI, #派息数(实)
gta.STK_DIVIDEND.DECLAREDATE #分红消息的时间
).filter(
gta.STK_DIVIDEND.ISDIVIDEND == 'Y', #有分红的股票
gta.STK_DIVIDEND.DIVDENDYEAR == year,
#且分红信息在上一年度
gta.STK_DIVIDEND.SYMBOL.in_(stocks_symbol))).dropna(axis=0)
stocks_symbol_this_year = list(df1['SYMBOL'])
if year_watch == 3:
#知道前两年的分红数据
df2 = gta.run_query(
query(
gta.STK_DIVIDEND.SYMBOL, #股票代码
gta.STK_DIVIDEND.DIVIDENTBT, #股票分红
gta.STK_DIVIDEND.TOTALDIVIDENDDISTRI, #派息数(实)
gta.STK_DIVIDEND.DECLAREDATE #分红消息的时间
).filter(
gta.STK_DIVIDEND.ISDIVIDEND == 'Y', #有分红的股票
gta.STK_DIVIDEND.DIVDENDYEAR == year - 1,
#且分红信息在上一年度
gta.STK_DIVIDEND.SYMBOL.in_(stocks_symbol))).dropna(axis=0)
#知道前3年的分红数据
df3 = gta.run_query(
query(
gta.STK_DIVIDEND.SYMBOL, #股票代码
gta.STK_DIVIDEND.DIVIDENTBT, #股票分红
gta.STK_DIVIDEND.TOTALDIVIDENDDISTRI, #派息数(实)
gta.STK_DIVIDEND.DECLAREDATE #分红消息的时间
).filter(
gta.STK_DIVIDEND.ISDIVIDEND == 'Y', #有分红的股票
gta.STK_DIVIDEND.DIVDENDYEAR == year - 2,
#且分红信息在上一年度
gta.STK_DIVIDEND.SYMBOL.in_(stocks_symbol))).dropna(axis=0)
# 如果前一年的分红不知道，那么知道前4年的分红数据
df4 = gta.run_query(
query(
gta.STK_DIVIDEND.SYMBOL, #股票代码
gta.STK_DIVIDEND.DIVIDENTBT, #股票分红
gta.STK_DIVIDEND.TOTALDIVIDENDDISTRI, #派息数(实)
gta.STK_DIVIDEND.DECLAREDATE #分红消息的时间
).filter(
gta.STK_DIVIDEND.ISDIVIDEND == 'Y', #有分红的股票
gta.STK_DIVIDEND.DIVDENDYEAR == year - 3,
#且分红信息在上一年度
gta.STK_DIVIDEND.SYMBOL.in_(stocks_symbol),
gta.STK_DIVIDEND.SYMBOL.notin_(
stocks_symbol_this_year) #不知道今年信息的
)).dropna(axis=0)
df = pd.concat((df4, df3, df2, df1))
elif year_watch == 1:
# 如果前一年的分红不知道，那么知道前4年的分红数据
df2 = gta.run_query(
query(
gta.STK_DIVIDEND.SYMBOL, #股票代码
gta.STK_DIVIDEND.DIVIDENTBT, #股票分红
gta.STK_DIVIDEND.TOTALDIVIDENDDISTRI, #派息数(实)
gta.STK_DIVIDEND.DECLAREDATE #分红消息的时间
).filter(
gta.STK_DIVIDEND.ISDIVIDEND == 'Y', #有分红的股票
gta.STK_DIVIDEND.DIVDENDYEAR == year - 1,
#且分红信息在上一年度
gta.STK_DIVIDEND.SYMBOL.in_(stocks_symbol),
gta.STK_DIVIDEND.SYMBOL.notin_(
stocks_symbol_this_year) #不知道今年信息的
)).dropna(axis=0)
df = pd.concat((df2, df1))
else:
log.info('不支持1年和3年之外的参数！！！')
return
#print df[(df.SYMBOL == '601006')]
# 下面四行代码用于选择在当前时间内能已知去年股息信息的股票
df['pubtime'] = map(
lambda x: int(x.split('-')[0] + x.split('-')[1] + x.split('-')[2]),
df['DECLAREDATE'])
#print df['pubtime']
currenttime = int(
str(context.current_dt)[0:4] + str(context.current_dt)[5:7] +
str(context.current_dt)[8:10])
#currenttime = int(str(now.year)+'{:0>2}'.format(str(now.month))+'{:0>2}'.format(str(now.day)))
#print currenttime
# 筛选出pubtime小于当前时期的股票，然后剔除'DECLAREDATE','pubtime','SYMBOL'三列
# 并且将DIVIDENTBT 列转换为float
df = df[(df.pubtime < currenttime)]
df['SYMBOL'] = map(normalize_code, list(df['SYMBOL']))
df.index = list(df['SYMBOL'])
df = df.drop(['SYMBOL', 'pubtime', 'DECLAREDATE'], axis=1)
df['DIVIDENTBT'] = map(float, df['DIVIDENTBT'])
df['TOTALDIVIDENDDISTRI'] = map(float, df['TOTALDIVIDENDDISTRI'])
q_now = query(valuation.code, valuation.market_cap)
df_now = get_fundamentals(q_now)
df_now.index = list(df_now['code'])
#print df_now
#接下来这一步是考虑多次分红的股票，因此需要累加股票的多次分红
#按照股票代码分堆
df = df.groupby(df.index).sum()
df['market_cap'] = df_now['market_cap']
#得到当前股价
#Price=history(1, unit='1d', field='close', security_list=list(df.index), df=True, skip_paused=False, fq='pre')
#Price=get_price(list(df.index), count = 1, end_date=now , frequency='daily', fields='close')
#print Price['close']
#Price=Price['close'].T
#print Price
#Price=Price.T
#df['pre_close']=Price
#计算股息率 = 股息/股票价格
#df['divpercent']=df['DIVIDENTBT']/df['pre_close']
df['divpercent'] = df['TOTALDIVIDENDDISTRI'] / df[
'market_cap'] / 1000000 / year_watch
#print df
df['code'] = np.array(df.index)
#print df[(df.code == '601006.XSHG')]
df = df.sort(['divpercent'], ascending=[False])
df_name = get_all_securities(['stock'])
df['name'] = df_name['display_name']
return df
def test_gta2():
from jqdata import gta, query
print(gta.run_query(query(gta.STK_AF_FORECAST).limit(10)))
print(gta.run_query(query(gta.STK_AF_FORECAST.SYMBOL).limit(10)))
print(gta.run_query(gta.query(gta.STK_AF_FORECAST.SYMBOL).limit(10)))
pass