def MACDcode(code,
label='Close',
fee=0.0003,
longd=26,
shortd=12,
difd=9,
position='Case4',
plot=True):
"""
code: 上海交易所和深圳交易所的股票代码
label: Close
fee: 手续费,0.0003
longd: 慢周期，26
shortd: 快周期，12
difd: DIF 的周期时间,9,
position: 判断买入的方式,Case1 以位置,Case2 为交叉,Case3 MACD柱行图
"""
try:
df = wt.get_k_data(code)
MACDdf(df, label, fee, longd, shortd, difd, position, plot)
except:
if len(code) == 6:
if code[0] == '0' or code[0] == '3':
code = 'YAHOO/SZ_' + code
elif code[0] == '6':
code = 'YAHOO/SS_' + code
elif len(code) == 4:
code = 'YAHOO/HK_' + code
df = wq.quandlyd(code)
MACDdf(df, label, fee, longd, shortd, difd, position, plot)
return
def KDJcode(code, label='Close', fee=0.0003, window=5, plot=True):
"""
"""
if len(code) == 6:
if code[0] in ['0', '3', '2', '6', '9']:
df = wt.get_k_data(code)
KDJdf(df, label, fee, window, plot)
else:
sys.exit()
elif len(code) == 4:
code = 'YAHOO/HK_' + code
df = wq.quandlyd(code)
KDJdf(df, label, fee, window, plot)
return
def Tutlecode(code,label='Close',N1=20,N2=10,fee=0.0005,path='',plot=True):
try:
df=wt.get_k_data(code)
d=Tutledf(df,label,N1,N2,fee,path,plot)
except:
if len(code)==6:
if code[0]=='0' or code[0]=='3':
code='YAHOO/SZ_'+code
elif code[0]=='6':
code='YAHOO/SS_'+code
elif len(code)==4:
code='YAHOO/HK_'+code
df=wq.quandlyd(code)
d=Tutledf(df,label,N1,N2,fee,path,plot)
return
def RSIcode(code, label='Close', fee=0.0003, longd=26, shortd=12, plot=True):
try:
df = wt.get_k_data(code)
RSIdf(df, label, fee, longd, shortd, plot)
except:
if len(code) == 6:
if code[0] == '0' or code[0] == '3':
code = 'YAHOO/SZ_' + code
elif code[0] == '6':
code = 'YAHOO/SS_' + code
elif len(code) == 4:
code = 'YAHOO/HK_' + code
df = wq.quandlyd(code)
RSIdf(df, label, fee, longd, shortd, plot)
return
def ChaikinADcode(code,
label='Close',
fee=0.0005,
M=5,
N=10,
path='',
plot=True):
try:
df = wt.get_k_data(code)
d = ChaikinADdf(df, label, fee, M, N, path, plot)
except:
if len(code) == 6:
if code[0] == '0' or code[0] == '3':
code = 'YAHOO/SZ_' + code
elif code[0] == '6':
code = 'YAHOO/SS_' + code
elif len(code) == 4:
code = 'YAHOO/HK_' + code
df = wq.quandlyd(code)
d = ChaikinADdf(df, label, fee, M, N, path, plot)
return
return df
def ChaikinADcode(code,
label='Close',
fee=0.0005,
M=5,
N=10,
path='',
plot=True):
try:
df = wt.get_k_data(code)
d = ChaikinADdf(df, label, fee, M, N, path, plot)
except:
if len(code) == 6:
if code[0] == '0' or code[0] == '3':
code = 'YAHOO/SZ_' + code
elif code[0] == '6':
code = 'YAHOO/SS_' + code
elif len(code) == 4:
code = 'YAHOO/HK_' + code
df = wq.quandlyd(code)
d = ChaikinADdf(df, label, fee, M, N, path, plot)
return
if __name__ == "__main__":
df = wt.get_k_data('000039')
d = ChaikinADdf(df)
def ChaikinADv2_code(code="",
label='Close',
fee=0.0005,
M=5,
N=10,
path='',
plot=True,
Wr=False):
if code == "":
sys.exit()
elif isinstance(code, str):
if len(code) == 6:
df = wt.get_k_data(code)
elif len(code) == 4:
df = wq.quandlyd(code)
df = df.iloc[-640:, ]
else:
sys.exit()
df = df.sort_index(ascending=True)
#print(df)
df = df.drop_duplicates()
df = df.reset_index()
df = df.rename(columns=lambda x: str(x).lower())
label = label.lower()
if 'p_change' not in df.columns:
df.loc[:, 'p_change'] = df[label] / df[label].shift(1) - 1
df.loc[:, 'volume_log'] = df['volume'].apply(np.log10)
df.loc[:, 'AD_m'] = pd.Series.rolling(
(2 * df[label] - df['high'] - df['low']) / (df['high'] - df['low']) *
df['volume_log'],
window=M).mean()
df.loc[:, 'AD_n'] = pd.Series.rolling(
(2 * df[label] - df['high'] - df['low']) / (df['high'] - df['low']) *
df['volume_log'],
window=N).mean()
df.loc[:, 'Chaikin'] = df['AD_m'] - df['AD_n']
buyi = df[(df['Chaikin'] > df['Chaikin'].shift(1)) |
((df['Chaikin'] > 0) & (df['Chaikin'].shift(1) < 0))].index
#buyi=df[(df['Chaikin'] > df['Chaikin'].shift(1))].index
df.loc[buyi, 'Sign'] = 1
selli = df[(df['Chaikin'] < df['Chaikin'].shift(1)) |
((df['Chaikin'] < 0) & (df['Chaikin'].shift(1) > 0))].index
#selli=df[(df['Chaikin']<df['Chaikin'].shift(1))].index
df.loc[selli, 'Sign'] = 0
df['Sign'].fillna(method='ffill', inplace=True)
df.loc[:, 'position'] = df['Sign'].shift(1)
df.loc[:,
'Cash_index'] = (1 +
(df['p_change'] - fee) * df['position']).cumprod()
_position_code(df, code, W=False)
df = df.set_index('date')
#print(df.tail())
if plot:
fig = plt.figure(figsize=(9, 5))
fig.subplots_adjust(hspace=0.15, wspace=0.05,\
left=0.06,right=0.99,\
top=0.97,bottom=0.09)
ax1 = fig.add_subplot(211)
df[['AD_m', 'AD_n']].plot(ax=ax1)
plt.grid(True)
ax2 = fig.add_subplot(212)
df['Cash_index'].plot(ax=ax2)
plt.grid(True)
plt.show()
#print(df.tail())
return
def MACDv2_code(code,
label='Close',
fee=0.0005,
ld=26,
sd=12,
md=9,
path='',
plot=True,
Wr=False):
"""
code:为输入的股票代码，大陆的为6个数字代码，香港的为4个数字代码。
like：string 600036,0005
label：为所采用的列，open,high,close,low 或是Adj Close, Close
ld,sd,md：为计算平均数的周期,以及dea 的计算周期
fee:为交易的费率
path:生成文件保存的地址及文件名，默认是不保存文件
"""
if code == None:
sys.exit()
elif isinstance(code, str):
if len(code) == 6:
df = wt.get_k_data(code)
elif len(code) == 4:
df = wq.quandlyd(code)
df = df.iloc[-640:, ]
#print(df)
else:
sys.exit()
#df=df.drop_duplicates()
#df.index=pd.to_datetime(df.index)
df = df.reset_index()
df = df.rename(columns=lambda x: str(x).lower())
label = label.lower()
if 'p_change' not in df.columns:
df.loc[:, 'p_change'] = df[label] / df[label].shift(1) - 1
dfi = df.shape[0]
df.loc[0, 'M_ld'] = df.loc[0, label]
df.loc[0, 'M_sd'] = df.loc[0, label]
for i in range(1, dfi):
df.loc[i:, 'M_ld'] = df.loc[i, label] * 2 / (
ld + 1) + df.loc[i - 1, 'M_ld'] * (ld - 1) / (ld + 1)
df.loc[i:, 'M_sd'] = df.loc[i, label] * 2 / (
sd + 1) + df.loc[i - 1, 'M_sd'] * (sd - 1) / (sd + 1)
df.loc[:, 'DIF'] = df['M_sd'] - df['M_ld']
df.loc[0, 'DEA'] = df.loc[0, 'DIF']
for i in range(1, dfi):
df.loc[i, 'DEA'] = df.loc[i - 1, 'DEA'] * (md - 1) / (
md + 1) + df.loc[i, 'DIF'] * 2 / (md + 1)
df.loc[:, 'BAR'] = 2 * (df['DIF'] - df['DEA'])
buyi1 = df[(df['DIF'] > df['DEA'])
& (df['DIF'] > df['DIF'].shift(-1))].index
df.loc[buyi1, 'Sign'] = 0
seli1 = df[(df['DIF'] < df['DEA'])
& (df['DIF'] < df['DIF'].shift(-1))].index
df.loc[seli1, 'Sign'] = 1
df['Sign'].fillna(method='ffill', inplace=True)
df.loc[:, 'position'] = df['Sign'].shift(1)
df.loc[:,
'Cash_index'] = (1 +
(df['p_change'] - fee) * df['position']).cumprod()
df.loc[:, 'vov'] = pd.Series.rolling(df['p_change'], window=9).std()
df = df.set_index('date')
_position_code(df, code)
if plot:
fig = plt.figure(figsize=(9, 5))
fig.subplots_adjust(hspace=0.15, wspace=0.05,\
left=0.06,right=0.99,\
top=0.97,bottom=0.09)
ax1 = fig.add_subplot(411)
df[label].plot(ax=ax1)
plt.grid(True)
ax2 = fig.add_subplot(412)
df['Cash_index'].plot(ax=ax2)
plt.grid(True)
ax3 = fig.add_subplot(413)
df[['p_change', 'vov']].plot(ax=ax3)
plt.grid(True)
ax4 = fig.add_subplot(414)
df['BAR'].plot(kind='bar', ax=ax4)
ax5 = ax4.twinx()
df[['DIF', 'DEA']].plot(kind='line', ax=ax4)
plt.grid(True)
monthsLoc = mpl.dates.MonthLocator()
weeksLoc = mpl.dates.WeekdayLocator()
ax4.xaxis.set_major_locator(monthsLoc)
ax4.xaxis.set_minor_locator(weeksLoc)
monthsFmt = mpl.dates.DateFormatter('%b')
ax4.xaxis.set_major_formatter(monthsFmt)
plt.show()
if path != '':
df.to_csv(path)
return
def Tutlev2_code(code=None,
label='Close',
N1=20,
N2=10,
fee=0.0005,
path='',
plot=True,
Wr=False):
"""
code:为输入的股票代码，大陆的为6个数字代码，香港的为4个数字代码。
like：string 600036,0005
df:为输入的DataFrame数据
label：为所采用的列，open,high,close,low 或是Adj Close
N1,N2：为计算平均数的周期
fee:为交易的费率
path:生成文件保存的地址及文件名，默认是不保存文件
"""
if code == None:
sys.exit()
elif isinstance(code, str):
if len(code) == 6:
df = wt.get_k_data(code)
elif len(code) == 4:
df = wq.quandlyd(code)
df = df.iloc[-640:, ]
#print(df)
else:
sys.exit()
if isinstance(df, pd.DataFrame):
df = df.sort_index(ascending=True)
df = df.drop_duplicates()
df = df.reset_index()
df = df.rename(columns=lambda x: str(x).lower())
label = label.lower()
index_data = df
if 'p_change' not in df.columns:
index_data['p_change'] = (
index_data[label] / index_data[label].shift(1) - 1) * 100
# 保留这几个需要的字段：'date', 'high', 'low', 'close', 'change'
index_data = index_data[[
'date', 'open', 'high', 'low', label, 'p_change'
]]
index_data.loc[:, 'High_Close_Price_N1_Day'] = pd.Series.rolling(
index_data['high'], window=N1, center=False).max()
index_data['High_Close_Price_N1_Day'].fillna(
value=index_data['high'].expanding(min_periods=1).max())
index_data.loc[:, 'Low_Close_Price_N2_Day'] = pd.Series.rolling(
index_data['low'], window=N2, center=False).min()
index_data['Low_Close_Price_N2_Day'].fillna(
value=index_data['low'].expanding(min_periods=1).min())
#index_data['Low_Close_Price_N2_Day']=index_data['Low_Close_Price_N2_Day'].shift()
# 当当天的【close】> 昨天的【最近N1个交易日的最高点】时，将【收盘发出的信号】设定为1
buy_index = index_data[
index_data['close'] > index_data['High_Close_Price_N1_Day'].shift(
1)].index
index_data.loc[buy_index, 'Sign'] = 1
# 当当天的【close】< 昨天的【最近N2个交易日的最低点】时，将【收盘发出的信号】设定为0
sell_index = index_data[
index_data['close'] < index_data['Low_Close_Price_N2_Day'].shift(
1)].index
index_data.loc[sell_index, 'Sign'] = 0
# 计算每天的仓位，当天持有上证指数时，仓位为1，当天不持有上证指数时，仓位为0
index_data['Sign'].fillna(method='ffill', inplace=True)
index_data['position'] = index_data['Sign'].shift(1)
#index_data['position'].fillna(method='ffill', inplace=True)
#当仓位为1时，已当天的开盘价买入股票，当仓位为0时，以收盘价卖出该股份。计算从数据期内的收益
index_data['Cash_index'] = (
(index_data['p_change'] / 100 - fee) * index_data['position'] +
1.0).cumprod()
#initial_idx = index_data.iloc[0]['close'] / (1 + (index_data.iloc[0]['p_change']/100))
initial_idx = 1
index_data['Cash_index'] *= initial_idx
_position_code(index_data, code)
#print ('The turtle methon Signal:')
if path != '':
index_data.to_csv(path)
# ==========计算每年指数的收益以及海龟交易法则的收益
index_data['p_change_turtle'] = (
index_data['p_change']) * index_data['position']
index_data = index_data.reset_index(drop=True)
index_data = index_data.set_index('date')
if plot:
index_data[[label, 'p_change', 'Cash_index']].plot(subplots=True,
figsize=(9, 5),
grid=True)
plt.grid(True)
plt.show()
return
def KDJv2_code(code=None,
label='close',
window=9,
fee=0.0005,
path='',
plot=True,
Wr=False):
"""
code:为输入的股份代码，6位的为境内的代码，4位的为香港的代码
label：为所采用的列，open,high,close,low 或是Adj Close
fee:为交易的费率
path:生成文件保存的地址及文件名，默认是不保存文件
"""
if code == None:
sys.exit()
elif isinstance(code, str):
if len(code) == 6:
df = wt.get_k_data(code)
elif len(code) == 4:
df = wq.quandlyd(code)
df = df.iloc[-640:, ]
#print(df)
else:
sys.exit()
if isinstance(df, pd.DataFrame):
df = df.sort_index(ascending=True)
df = df.drop_duplicates()
df = df.reset_index()
df = df.rename(columns=lambda x: str(x).lower())
label = label.lower()
stock_dataT = df
if 'p_change' not in df.columns:
stock_dataT['p_change'] = (
stock_dataT[label] / stock_dataT[label].shift(1) - 1) * 100
stock_data = stock_dataT.loc[:, ('date', 'high', 'low', label,
'p_change')].copy()
# 计算KDJ指标
stock_data['low_list'] = pd.Series.rolling(stock_data['low'],
window=window,
center=False).min()
stock_data['low_list'].fillna(value=stock_data['low'].expanding(
min_periods=1).min())
stock_data['high_list'] = pd.Series.rolling(stock_data['high'],
window=window,
center=False).max()
stock_data['high_list'].fillna(value=stock_data['high'].expanding(
min_periods=1).max())
stock_data['rsv'] = (stock_data[label] - stock_data['low_list']) / (
stock_data['high_list'] - stock_data['low_list']) * 100
stock_data['KDJ_K'] = pd.Series.ewm(stock_data['rsv'],
ignore_na=False,
min_periods=0,
adjust=True,
com=2).mean()
stock_data['KDJ_D'] = pd.Series.ewm(stock_data['KDJ_K'],
ignore_na=False,
min_periods=0,
adjust=True,
com=2).mean()
stock_data['KDJ_J'] = 3 * stock_data['KDJ_K'] - 2 * stock_data['KDJ_D']
# 计算KDJ指标金叉、死叉情况
###通常就敏感性而言，J值最强，K值次之，D值最慢，而就安全性而言，J值最差，K值次之，D值最稳
##金叉用1表示，死叉用0表示
buyi = stock_data[(stock_data['KDJ_K'] > stock_data['KDJ_D'])
& (stock_data['KDJ_K'].shift(-1) <
stock_data['KDJ_D'].shift(-1))].index
stock_data.loc[buyi, 'Sign'] = 0
selli = stock_data[(stock_data['KDJ_K'] < stock_data['KDJ_D'])
& (stock_data['KDJ_K'].shift(-1) >
stock_data['KDJ_D'].shift(-1))].index
stock_data.loc[selli, 'Sign'] = 1
stock_data['Sign'].fillna(method='ffill', inplace=True)
##测试反向操作的效果，即买进信号则卖出，卖出信号则买进。
stock_data['position'] = stock_data['Sign'].shift(1)
#stock_data['position'].fillna(method='ffill', inplace=True)
#当仓位为1时，已当天的开盘价买入股票，当仓位为0时，以收盘价卖出该股份。计算从数据期内的收益
stock_data['Cash_index'] = ((stock_data['p_change']/100 -fee) * \
stock_data['position']+1).cumprod()
#initial_idx = stock_data.iloc[0]['close'] / (1 + (stock_data.iloc[0]['p_change']/100))
initial_idx = 1
stock_data['Cash_index'] *= initial_idx
stock_data = stock_data.set_index('date')
_position_code(stock_data, code, W=Wr)
"""
if (stock_data.iloc[-1]['position']==1)&(stock_data.iloc[-2]['position']==1):
print ('持有这只股票Holding code')
elif (stock_data.iloc[-1]['position']==1)&(stock_data.iloc[-2]['position']==0):
print ('开始买进这只股票Buy code')
elif (stock_data.iloc[-1]['position']==0)&(stock_data.iloc[-2]['position']==1):
print ('开始卖出这只股票Sell code')
else:
print ('保持空仓Keeping Short Position code')
"""
if path != '':
stock_data.to_csv(path)
stock_data['vov'] = pd.Series.rolling(stock_data['p_change'],
window=window,
center=False).std()
if plot:
stock_data[[label, 'p_change', 'vov',
'Cash_index']].plot(subplots=True,
figsize=(9, 5),
grid=True)
plt.grid(True)
plt.show()
return
df.loc[selli, 'Sign'] = 0
elif position == 'Case4':
#交叉情况
buyi=df[(df['DIF'].shift(-1)<df['DEA'].shift(-1)) & \
(df['DIF']>df['DEA'])].index
selli=df[((df['DIF'].shift(-1)>df['DEA'].shift(-1)) & \
(df['DIF']<df['DEA']))].index
df.loc[buyi, 'Sign'] = 0
df.loc[selli, 'Sign'] = 1
df['Sign'].fillna(method='ffill', inplace=True)
df['position'] = df['Sign'].shift(1)
df['Cash_ind'] = (1 + (df['p_change'] - fee) * df['position']).cumprod()
_position(df)
df = df.set_index('date')
if plot:
plott(df, label)
return
if __name__ == "__main__":
df = wt.get_k_data('601166')
MACDdf(df, position='Case4')
#MACDcode('601166')
