def app():
st.write("""
# S&P 500 Stock Analyzer
Shown below are the **Fundamentals**, **News Sentiment**, **Bollinger Bands** and **Comprehensive Report (Compared with SPY as a whole as benchmark)** of your selected stock!
""")
st.markdown("***")
st.sidebar.header('User Input Parameters')
symbol, start, end = user_input_features()
start = pd.to_datetime(start)
end = pd.to_datetime(end)
#symbol1 = get_symbol(symbol.upper())
#st.subheader(symbol1)
stock = finvizfinance(symbol.lower())
stock_chart = stock.ticker_charts()
st.image(stock_chart)
# Read data
data = yf.download(symbol, start, end, threads=False)
# ## SMA and EMA
#Simple Moving Average
data['SMA'] = talib.SMA(data['Adj Close'], timeperiod=20)
# Exponential Moving Average
data['EMA'] = talib.EMA(data['Adj Close'], timeperiod=20)
# Plot
st.subheader(f"""
Simple Moving Average vs. Exponential Moving Average\n {symbol}
""")
st.line_chart(data[['Adj Close', 'SMA', 'EMA']])
# Bollinger Bands
data['upper_band'], data['middle_band'], data['lower_band'] = talib.BBANDS(
data['Adj Close'], timeperiod=20)
# Plot
st.subheader(f"""
Bollinger Bands\n {symbol}
""")
st.line_chart(
data[['Adj Close', 'upper_band', 'middle_band', 'lower_band']])
# ## MACD (Moving Average Convergence Divergence)
# MACD
data['macd'], data['macdsignal'], data['macdhist'] = talib.MACD(
data['Adj Close'], fastperiod=12, slowperiod=26, signalperiod=9)
# Plot
st.subheader(f"""
Moving Average Convergence Divergence\n {symbol}
""")
st.line_chart(data[['macd', 'macdsignal']])
# ## RSI (Relative Strength Index)
# RSI
data['RSI'] = talib.RSI(data['Adj Close'], timeperiod=14)
# Plot
st.subheader(f"""
Relative Strength Index\n {symbol}
""")
st.line_chart(data['RSI'])
st.markdown("***")
st.subheader("Fundamentals: ")
st.dataframe(get_fundamentals(symbol))
st.markdown("***")
# ## Latest News
st.subheader("Latest News: ")
st.table(get_news(symbol).head(5))
st.markdown("***")
# ## Recent Insider Trades
st.subheader("Recent Insider Trades: ")
st.table(get_insider(symbol).head(5))
st.markdown("***")
st.write("Generating comprehensive stock report...")
st.write("**please wait for some time... **")
st.write(
"This section will compare the historical performance of your selected stock with SPDR S&P 500 Trust ETF (Ticker: SPY) as benchmark."
)
stock_report(symbol)
# ## Stock report
st.subheader(f"""**{symbol} Stock Report**""")
#st.header(symbol + " Stock Report")
HtmlFile = open("report.html", 'r', encoding='utf-8')
source_code = HtmlFile.read()
#print(source_code)
components.html(source_code, height=9000)
st.write(
"Disclaimer: The data are collected from Google, Yahoo Finance and Finviz"
)
def ZIGZAG(self,
df,
minbars=12,
bb_period=20,
bb_dev=2.0,
nan_value=0.0,
dropna=True,
level=logging.WARN):
"""Builds a ZIGZAG indicator based on Bollinger Bands overbought and oversell signals
Keyword arguments:
df -- Datafeed to apply indicator
minbars -- Min. number of bars per flip to avoid discarding (default 12)
bb_period -- Bollinger bands period (default 20)
bb_dev -- Bollinger bands deviation (default 2.0)
nan_value -- Values for zigzag indicator during search phase (default 0.0)
dropna -- Flag to delete NaN values, else fill them with nan_value
level -- logging level (default WARN)
"""
class ActionCtrl():
class ActionType(Enum):
NoActions = 0
SearchingHigh = 1
SearchingLow = 2
def __init__(self, high, low, idx, delta, level=logging.WARN):
self.__logger = logging.getLogger(__name__)
self.__logger.setLevel(level)
self.curr = ActionCtrl.ActionType.NoActions
self.last_high = high
self.last_high_idx = idx
self.beforelast_high = high
self.beforelast_high_idx = idx
self.last_swing_high_idx = idx
self.last_low = low
self.last_low_idx = idx
self.beforelast_low = low
self.beforelast_low_idx = idx
self.last_swing_low_idx = idx
self.delta = delta
self.events = ZIGZAG_Events()
self.__logger.debug(
'New action at idx={}: last_high={}, last_low={}, min-delta={}'
.format(idx, self.last_high, self.last_low, self.delta))
#---end action::__init__
def __result(self):
if self.curr == ActionCtrl.ActionType.SearchingHigh:
return 'high'
elif self.curr == ActionCtrl.ActionType.SearchingLow:
return 'low'
return 'no-action'
#---end action::__result
# this function updates MAX|MIN values with last recorded depending on the current action
def zigzag(self, x, df):
log = 'Procesing [{}]:'.format(x.name)
self.events.clear()
# check if HIGH must be updated
max_value = x.HIGH
if self.curr == ActionCtrl.ActionType.SearchingHigh and max_value > self.last_high:
#self.beforelast_high = self.last_high
#self.beforelast_high_idx = self.last_high_idx
self.last_high = max_value
self.last_high_idx = x.name
log += ' new HIGH={}'.format(max_value)
self.__logger.debug(log)
return self.__result()
# check if LOW must be updated
min_value = x.LOW
if self.curr == ActionCtrl.ActionType.SearchingLow and min_value < self.last_low:
#self.beforelast_low = self.last_low
#self.beforelast_low_idx = self.last_low_idx
self.last_low = min_value
self.last_low_idx = x.name
log += ' new LOW={}'.format(min_value)
self.__logger.debug(log)
return self.__result()
# check if search HIGH starts
if self.curr != ActionCtrl.ActionType.SearchingHigh and max_value > x.BOLLINGER_HI:
_prev_action = self.curr
self.events.ZIGZAG_StartMaxSearch = True
self.curr = ActionCtrl.ActionType.SearchingHigh
# check delta condition
curr_delta = (x.name - self.last_high_idx)
if _prev_action == ActionCtrl.ActionType.NoActions:
# in first swing doesnt apply
curr_delta = self.delta + 1
if curr_delta < self.delta:
log += ' ERR_DELTA \/ ={}'.format(curr_delta)
df.at[self.last_high_idx, 'ZIGZAG'] = nan_value
df.at[self.last_high_idx, 'ACTION'] = 'high'
if max_value > self.last_high:
log += ' replace_HIGH @[{}]=>{}'.format(
self.last_high_idx, max_value)
self.last_high = max_value
self.last_high_idx = x.name
else:
log += ' keep_HIGH @[{}]=>{}'.format(
self.last_high_idx, self.last_high)
df.at[self.last_low_idx, 'ZIGZAG'] = nan_value
df.at[self.last_low_idx, 'ACTION'] = 'high'
log += ' remove LOW @[{}]'.format(self.last_low_idx)
self.last_low = self.beforelast_low
self.last_low_idx = self.beforelast_low_idx
self.__logger.info(log)
else:
# save last low
df.at[self.last_low_idx, 'ZIGZAG'] = self.last_low
self.beforelast_low = self.last_low
self.beforelast_low_idx = self.last_low_idx
# starts high recording
#self.beforelast_high = self.last_high
#self.beforelast_high_idx = self.last_high_idx
self.last_high = max_value
self.last_high_idx = x.name
log += ' save LOW @[{}]={}, new HIGH=>{}'.format(
self.last_low_idx, self.last_low, max_value)
self.__logger.debug(log)
return self.__result()
# check if search LOW starts
if self.curr != ActionCtrl.ActionType.SearchingLow and min_value < x.BOLLINGER_LO:
_prev_action = self.curr
self.events.ZIGZAG_StartMinSearch = True
self.curr = ActionCtrl.ActionType.SearchingLow
# check delta condition
curr_delta = (x.name - self.last_low_idx)
if _prev_action == ActionCtrl.ActionType.NoActions:
# in first swing doesnt apply
curr_delta = self.delta + 1
if curr_delta < self.delta:
log += ' ERR_DELTA /\ ={}'.format(curr_delta)
df.at[self.last_low_idx, 'ZIGZAG'] = nan_value
df.at[self.last_low_idx, 'ACTION'] = 'low'
if min_value < self.last_low:
log += ' replace_LOW @[{}]=>{}'.format(
self.last_low_idx, min_value)
self.last_low = min_value
self.last_low_idx = x.name
else:
log += ' keep_LOW @[{}]=>{}'.format(
self.last_low_idx, self.last_low)
df.at[self.last_high_idx, 'ZIGZAG'] = nan_value
df.at[self.last_high_idx, 'ACTION'] = 'low'
log += ' remove HIGH @[{}]'.format(self.last_high_idx)
self.last_high = self.beforelast_high
self.last_high_idx = self.beforelast_high_idx
self.__logger.info(log)
else:
# save last high
df.at[self.last_high_idx, 'ZIGZAG'] = self.last_high
self.beforelast_high = self.last_high
self.beforelast_high_idx = self.last_high_idx
# starts low recording
#self.beforelast_low = self.last_low
#self.beforelast_low_idx = self.last_low_idx
self.last_low = min_value
self.last_low_idx = x.name
log += ' save HIGH @[{}]={}, new LOW=>{}'.format(
self.last_high_idx, self.last_high, min_value)
self.__logger.debug(log)
return self.__result()
if self.curr == ActionCtrl.ActionType.SearchingLow:
log += ' curr LOW @[{}]=>{}'.format(
self.last_low_idx, self.last_low)
elif self.curr == ActionCtrl.ActionType.SearchingHigh:
log += ' curr HIGH @[{}]=>{}'.format(
self.last_high_idx, self.last_high)
self.__logger.debug(log)
return self.__result()
#---end action::zigzag
# registers previous zzpoints and their indexes at current row.Px and row.Px_idx
def points(self, row, df, nan_value):
log = 'row [{}] zigzag={}: '.format(row.name, row.ZIGZAG)
# get last 5 zigzag points
zzpoints = df.ZIGZAG[(df.index < row.name)
& (df.ZIGZAG != nan_value) &
(df.ACTION.str.contains('in-progress')
== False)]
# at least requires 6 points, else discard
if zzpoints.shape[0] < 6:
log += 'exit-zzpoints-count={} '.format(zzpoints.shape[0])
self.__logger.debug(log)
return
# update rows
for i in range(1, 7):
df.at[row.name, 'P{}'.format(i)] = zzpoints.iloc[-i]
df.at[row.name, 'P{}_idx'.format(i)] = zzpoints.index[-i]
log += 'Updated!! '
self.__logger.debug(log)
#---end action::points
# clear events
self.__events.clear()
# copy dataframe and calculate bollinger bands if not yet present
_df = df.copy()
_df['BOLLINGER_HI'], _df['BOLLINGER_MA'], _df[
'BOLLINGER_LO'] = talib.BBANDS(_df.CLOSE,
timeperiod=bb_period,
nbdevup=bb_dev,
nbdevdn=bb_dev,
matype=0)
_df['BOLLINGER_WIDTH'] = _df['BOLLINGER_HI'] - _df['BOLLINGER_LO']
boll_b = (_df.CLOSE - _df['BOLLINGER_LO']) / _df['BOLLINGER_WIDTH']
boll_b[np.isnan(boll_b)] = 0.5
boll_b[np.isinf(boll_b)] = 0.5
_df['BOLLINGER_b'] = boll_b
if dropna:
_df.dropna(inplace=True)
_df.reset_index(drop=True, inplace=True)
else:
_df.fillna(value=nan_value, inplace=True)
# Initially no actions are in progress, record first high and low values creating an ActionCtrl object
action = ActionCtrl(
high=_df['HIGH'][0], #max(_df['OPEN'][0], _df['CLOSE'][0]),
low=_df['LOW'][0], #min(_df['OPEN'][0], _df['CLOSE'][0]),
idx=_df.iloc[0].name,
delta=minbars,
level=level)
_df['ZIGZAG'] = nan_value
_df['ACTION'] = 'no-action'
_df['ACTION'] = _df.apply(lambda x: action.zigzag(x, _df), axis=1)
# fills last element as pending
if _df.ZIGZAG.iloc[-1] == nan_value:
if _df.ACTION.iloc[-1] == 'high':
_df.at[action.last_high_idx, 'ZIGZAG'] = action.last_high
_df.at[action.last_high_idx, 'ACTION'] = 'high-in-progress'
else:
_df.at[action.last_low_idx, 'ZIGZAG'] = action.last_low
_df.at[action.last_low_idx, 'ACTION'] = 'low-in-progress'
# now adds point p1 to p6 backtrace of values and indexes:
for i in range(1, 7):
_df['P{}'.format(i)] = nan_value
_df['P{}_idx'.format(i)] = 0
_df.apply(lambda x: action.points(x, _df, nan_value), axis=1)
self.__df = _df
self.__action = action
return self.__df, self.__action.events
def bbands(source, timeperiod=5, nbdevup=2, nbdevdn=2, matype=0):
return talib.BBANDS(source, timeperiod, nbdevup, nbdevdn, matype)
# ## SMA and EMA
#Simple Moving Average
data['SMA'] = ta.SMA(data['Adj Close'], timeperiod=20)
# Exponential Moving Average
data['EMA'] = ta.EMA(data['Adj Close'], timeperiod=20)
# Plot
data[['Adj Close', 'SMA', 'EMA']].plot(figsize=(10, 5))
plt.subplots()
plt.show()
# ## Bollinger Bands
# Bollinger Bands
data['upper_band'], data['middle_band'], data['lower_band'] = ta.BBANDS(
data['Adj Close'], timeperiod=20)
# Plot
data[['Adj Close', 'upper_band', 'middle_band',
'lower_band']].plot(figsize=(10, 5))
plt.show()
# ## MACD (Moving Average Convergence Divergence)
# MACD
data['macd'], data['macdsignal'], data['macdhist'] = ta.MACD(data['Adj Close'],
fastperiod=12,
slowperiod=26,
signalperiod=9)
data[['macd', 'macdsignal']].plot(figsize=(10, 5))
plt.show()
def process_indextools(self):
"""======================更新辅助指标=================="""
###更新closes
close = self.newBar['close']
self.closes.append(close)
if len(self.closes) < self.macd_len: return #如果收盘价列表不够用于计算，则继续添加
if len(self.fenxingTuples) <= 1: return #如果两个分型都没有，则继续等待分型
now_fenxing = self.fenxingTuples[-1]
###判断分型是否更新
###未更新
if self.last_fenxing_datetime == now_fenxing[0]:
self.macd_area_buffer.append(
talib.MACD(np.array(self.closes))[-1][-1]) #buffer添加一个
###更新了
else:
###替代前分型
if self.last_fenxing_type == now_fenxing[2]:
#macd处理
self.macd_area_buffer.append(
talib.MACD(np.array(self.closes))[-1][-1]) #buffer添加一个
self.macd_area_backup.extend(
self.macd_area_buffer) #backup加入buffer
macd_area = sum(self.macd_area_backup) #计算
#self.macd_areas[-1] = macd_area#替换最后一个红蓝柱面积
self.macd_areas[-1] = (now_fenxing[0], macd_area) #替换最后一个红蓝柱面积
self.macd_area_buffer = [] #清空buffer
#bolling处理
ceil = talib.BBANDS(np.array(self.closes))[0][-2]
floor = talib.BBANDS(np.array(self.closes))[-1][-2]
self.powerful_area = ((now_fenxing[1] > ceil) |
(now_fenxing[1] < floor))
#self.bolling_fenxing[-1] = self.powerful_area
self.bolling_fenxing[-1] = (now_fenxing[0], self.powerful_area)
#判断变量
self.last_fenxing_datetime = now_fenxing[0]
###增加新的分型
else:
#macd处理
self.macd_area_backup = [] #清空backup
self.macd_area_buffer.append(
talib.MACD(np.array(self.closes))[-1][-1]) #buffer添加一个
self.macd_area_backup.extend(
self.macd_area_buffer) #backup加入buffer
macd_area = sum(self.macd_area_backup)
#self.macd_areas.append(sum(self.macd_area_backup))#增加一个红蓝柱面积
self.macd_areas.append((now_fenxing[0], macd_area)) #增加一个红蓝柱面积
self.macd_area_buffer = [] #清空buffer
#bolling处理
ceil = talib.BBANDS(np.array(self.closes))[0][-2]
floor = talib.BBANDS(np.array(self.closes))[-1][-2]
self.powerful_area = ((now_fenxing[1] > ceil) |
(now_fenxing[1] < floor))
#self.bolling_fenxing.append(self.powerful_area)
self.bolling_fenxing.append(
(now_fenxing[0], self.powerful_area))
#判断变量
self.last_fenxing_datetime = now_fenxing[0]
self.last_fenxing_type = now_fenxing[2]
def test_case(self):
close, open, ma2 = [], [], []
pre_ma2 = []
true_test = []
boll = {
'upper': [],
'middler': [],
'lower': []
}
boll3 = {
'upper': [],
'middler': [],
'lower': []
}
class DemoStrategy(Strategy):
def on_init(self, ctx):
"""初始化数据"""
ctx.ma2 = MA(ctx.close, 2)
ctx.boll = BOLL(ctx.close, 2)
def on_bar(self, ctx):
if ctx.curbar>=2:
## @todo + - * /
true_test.append(ctx.ma2-0 == ctx.ma2[0])
pre_ma2.append(ctx.ma2[3])
ma2.append(ctx.ma2[0])
close.append(ctx.close[0])
open.append(ctx.open[0])
boll['upper'].append(ctx.boll['upper'][0])
boll['middler'].append(ctx.boll['middler'][0])
boll['lower'].append(ctx.boll['lower'][0])
boll3['upper'].append(ctx.boll['upper'][3])
boll3['middler'].append(ctx.boll['middler'][3])
boll3['lower'].append(ctx.boll['lower'][3])
def on_final(self, ctx):
return
def on_exit(self, ctx):
return
## @todo 滚动的时候无法通过测试
set_pcontracts([pcontract('BB.SHFE', '1.Minute')], 0)
add_strategy([DemoStrategy('A1')])
run()
self.assertTrue(all(true_test), "指标转化错误!")
logger.info('----------------- 指标转化测试成功 -------------------')
# 分别测试单值和多值指标函数。
source_ma2 = talib.SMA(np.asarray(close), 2)
true_test, false_test = [], []
for i in xrange(0, len(close)):
if i >= 1:
true_test.append(ma2[i] == source_ma2[i])
else:
false_test.append(ma2[i] == ma2[i])
self.assertFalse(any(false_test), "单值指标正例测试失败!")
self.assertTrue(all(true_test), "单值指标正例测试失败!")
source_ma2 = talib.SMA(np.asarray(open), 2)
true_test, false_test = [], []
for i in xrange(0, len(open)):
if i >= 1:
true_test.append(ma2[i] == source_ma2[i])
else:
false_test.append(ma2[i] == ma2[i])
self.assertFalse(any(false_test), "单值指标反例测试失败!")
self.assertFalse(all(true_test), "单值指标反例测试失败!")
logger.info('----------------- 单值指标测试成功 -------------------')
true_test, false_test = [], []
source_ma2 = talib.SMA(np.asarray(close), 2)
for i in xrange(0, len(close)):
if i >= 4:
true_test.append(pre_ma2[i] == source_ma2[i-3])
else:
# 指标序列变量的默认值为nan
false_test.append(pre_ma2[i] == pre_ma2[i])
self.assertTrue(all(true_test), "单值指标回溯正例测试成功")
self.assertFalse(any(false_test), "单值指标回溯正例测试成功")
true_test, false_test = [], []
source_ma2 = talib.SMA(np.asarray(open), 2)
for i in xrange(0, len(close)):
if i >= 4:
true_test.append(pre_ma2[i] == source_ma2[i-3])
else:
# 指标序列变量的默认值为nan
false_test.append(pre_ma2[i] == pre_ma2[i])
self.assertFalse(all(true_test), "单值指标回溯反例测试成功")
self.assertFalse(any(false_test), "单值指标回溯反例测试成功")
logger.info('----------------- 单值指标回溯测试成功 -------------------')
#
u, m, l = talib.BBANDS(np.asarray(close), 2, 2, 2)
true_test, false_test = [], []
for i in xrange(0, len(close)):
if i >= 1:
true_test.append(boll['upper'][i] == u[i])
true_test.append(boll['middler'][i] == m[i])
true_test.append(boll['lower'][i] == l[i])
else:
false_test.append(boll['upper'][i] == boll['upper'][i])
false_test.append(boll['middler'][i] == boll['middler'][i])
false_test.append(boll['lower'][i] == boll['lower'][i])
self.assertFalse(any(false_test), "多值指标正例测试失败!")
self.assertTrue(all(true_test), "多值指标正例测试失败!")
u, m, l = talib.BBANDS(np.asarray(open), 2, 2, 2)
true_test, false_test = [], []
for i in xrange(0, len(close)):
if i >= 1:
true_test.append(boll['upper'][i] == u[i])
true_test.append(boll['middler'][i] == m[i])
true_test.append(boll['lower'][i] == l[i])
else:
false_test.append(boll['upper'][i] == boll['upper'][i])
false_test.append(boll['middler'][i] == boll['middler'][i])
false_test.append(boll['lower'][i] == boll['lower'][i])
self.assertFalse(any(false_test), "多值指标反例测试失败!")
self.assertFalse(all(true_test), "多值指标反例测试失败!")
logger.info('----------------- 多值指标测试成功 -------------------')
true_test, false_test = [], []
u, m, l = talib.BBANDS(np.asarray(close), 2, 2, 2)
for i in xrange(0, len(close)):
if i >= 4:
true_test.append(boll3['upper'][i] == u[i-3])
else:
false_test.append(boll3['upper'][i] == boll3['upper'][i])
self.assertTrue(all(true_test), "多值指标回溯正例测试成功")
self.assertFalse(any(false_test), "多值指标回溯正例测试成功")
true_test, false_test = [], []
u, m, l = talib.BBANDS(np.asarray(open), 2, 2, 2)
for i in xrange(0, len(close)):
if i >= 4:
true_test.append(boll3['upper'][i] == u[i-3])
else:
false_test.append(boll3['upper'][i] == boll3['upper'][i])
self.assertFalse(all(true_test), "多值指标回溯反例测试成功")
self.assertFalse(any(false_test), "多值指标回溯反例测试成功")
logger.info('----------------- 多值指标回溯测试成功 -------------------')
def math_transform_process(event):
print(event.widget.get())
math_transform = event.widget.get()
upperband, middleband, lowerband = ta.BBANDS(close, timeperiod=5, nbdevup=2, nbdevdn=2, matype=0)
fig, axes = plt.subplots(2, 1, sharex=True)
ax1, ax2 = axes[0], axes[1]
axes[0].plot(close, 'rd-', markersize=3)
axes[0].plot(upperband, 'y-')
axes[0].plot(middleband, 'b-')
axes[0].plot(lowerband, 'y-')
axes[0].set_title(math_transform, fontproperties="SimHei")
if math_transform == '反余弦':
real = ta.ACOS(close)
axes[1].plot(real, 'r-')
elif math_transform == '反正弦':
real = ta.ASIN(close)
axes[1].plot(real, 'r-')
elif math_transform == '反正切':
real = ta.ATAN(close)
axes[1].plot(real, 'r-')
elif math_transform == '向上取整':
real = ta.CEIL(close)
axes[1].plot(real, 'r-')
elif math_transform == '余弦':
real = ta.COS(close)
axes[1].plot(real, 'r-')
elif math_transform == '双曲余弦':
real = ta.COSH(close)
axes[1].plot(real, 'r-')
elif math_transform == '指数':
real = ta.EXP(close)
axes[1].plot(real, 'r-')
elif math_transform == '向下取整':
real = ta.FLOOR(close)
axes[1].plot(real, 'r-')
elif math_transform == '自然对数':
real = ta.LN(close)
axes[1].plot(real, 'r-')
elif math_transform == '常用对数':
real = ta.LOG10(close)
axes[1].plot(real, 'r-')
elif math_transform == '正弦':
real = ta.SIN(close)
axes[1].plot(real, 'r-')
elif math_transform == '双曲正弦':
real = ta.SINH(close)
axes[1].plot(real, 'r-')
elif math_transform == '平方根':
real = ta.SQRT(close)
axes[1].plot(real, 'r-')
elif math_transform == '正切':
real = ta.TAN(close)
axes[1].plot(real, 'r-')
elif math_transform == '双曲正切':
real = ta.TANH(close)
axes[1].plot(real, 'r-')
plt.show()
def shaping_ohlc(ohlc, columns, drop=False, bbtp=9, over_rate=0.00035):
#columns dependency reverse resolution
cdrr = {
'O/O': ['O/O', 'ATR/O'],
'H/O': ['H/O', 'H-L/O', 'ATR/O'],
'L/O': ['L/O', 'H-L/O', 'ATR/O'],
'C/O': ['C/O', 'MACDS/O', 'MACDM/O', 'MACDL/O'],
'O/C': ['O/C'],
'H/C': ['H/C', 'H-L/C'],
'L/C': ['L/C', 'H-L/C'],
'C/C': ['C/C'],
'H-L/O': ['H-L/O'],
'H-L/C': ['H-L/C'],
'O-1/C': ['O-1/C'],
'H-1/C': ['H-1/C'],
'L-1/C': ['L-1/C'],
'C-1/C': ['C-1/C'],
'PositiveLine':
['PositiveLine', 'SignLine', '3SignLine', '6SignLine', '12SignLine'],
'NegativeLine':
['NegativeLine', 'SignLine', '3SignLine', '6SignLine', '12SignLine'],
'SignLine': ['SignLine'],
'3SignLine': ['3SignLine'],
'6SignLine': ['6SignLine'],
'12SignLine': ['12SignLine'],
'BB': ['BBUpper', 'BBMiddle', 'BBLower', 'BBOver', 'BBUnder'],
'BBOver': ['BBOver'],
'BBUnder': ['BBUnder'],
'MACDS': ['MACDS'],
'MACDM': ['MACDM'],
'MACDL': ['MACDL'],
'MACDS/O': ['MACDS/O'],
'MACDM/O': ['MACDM/O'],
'MACDL/O': ['MACDL/O'],
'ATR': ['ATR'],
'ATR/O': ['ATR/O'],
'RSI': ['RSI'],
'Time': ['Time'],
'Over': ['Over', 'Target_Over'],
'1moku-base': [
'1moku-base', '1moku-precedent-span1', '1moku-signal1',
'1moku-signal2', '1moku-signal3', '1moku-signal4', '1moku-signal5',
'1moku-signal6'
],
'1moku-conversion': [
'1moku-conversion', '1moku-precedent-span1', '1moku-signal1',
'1moku-signal2', '1moku-signal3', '1moku-signal4', '1moku-signal5',
'1moku-signal6'
],
'1moku-precedent-span1': [
'1moku-precedent-span1', '1moku-signal3', '1moku-signal4',
'1moku-signal5', '1moku-signal6'
],
'1moku-precedent-span2': [
'1moku-precedent-span2', '1moku-signal3', '1moku-signal4',
'1moku-signal5', '1moku-signal6'
],
'1moku-late-span': ['1moku-late-span'],
'1moku-signal1': ['1moku-signal1'],
'1moku-signal2': ['1moku-signal2'],
'1moku-signal3': ['1moku-signal3'],
'1moku-signal4': ['1moku-signal4'],
'1moku-signal5': ['1moku-signal5'],
'1moku-signal6': ['1moku-signal6'],
'Target_Open': [
'Target_Open', 'Target_O/O', 'Target_H/O', 'Target_L/O',
'Target_C/O', 'Target_O/C'
],
'Target_High': ['Target_High', 'Target_H/O', 'Target_H/C'],
'Target_Low': ['Target_Low', 'Target_L/O', 'Target_L/C'],
'Target_Close': [
'Target_Close', 'Target_C/O', 'Target_O/C', 'Target_H/C',
'Target_L/C', 'Target_C/C'
],
'Target_O/O': ['Target_O/O'],
'Target_H/O': ['Target_H/O'],
'Target_L/O': ['Target_L/O'],
'Target_C/O': ['Target_C/O'],
'Target_O/C': ['Target_O/C'],
'Target_H/C': ['Target_H/C'],
'Target_L/C': ['Target_L/C'],
'Target_C/C': ['Target_C/C'],
'Target_Over': ['Target_Over'],
}
if check_cdrr('O/O', cdrr, columns):
ohlc['O/O'] = ohlc['Open'] / ohlc['Open']
if check_cdrr('H/O', cdrr, columns):
ohlc['H/O'] = ohlc['High'] / ohlc['Open']
if check_cdrr('L/O', cdrr, columns):
ohlc['L/O'] = ohlc['Low'] / ohlc['Open']
if check_cdrr('C/O', cdrr, columns):
ohlc['C/O'] = ohlc['Close'] / ohlc['Open']
if check_cdrr('O/C', cdrr, columns):
ohlc['O/C'] = ohlc['Open'] / ohlc['Close']
if check_cdrr('H/C', cdrr, columns):
ohlc['H/C'] = ohlc['High'] / ohlc['Close']
if check_cdrr('L/C', cdrr, columns):
ohlc['L/C'] = ohlc['Low'] / ohlc['Close']
if check_cdrr('C/C', cdrr, columns):
ohlc['C/C'] = ohlc['Close'] / ohlc['Close']
if check_cdrr('H-L/O', cdrr, columns):
ohlc['H-L/O'] = ohlc['H/O'] - ohlc['L/O']
if check_cdrr('H-L/C', cdrr, columns):
ohlc['H-L/C'] = ohlc['H/C'] - ohlc['L/C']
if check_cdrr('O-1/C', cdrr, columns):
ohlc['O-1/C'] = ohlc['Open'].shift(1) / ohlc['Close']
if check_cdrr('H-1/C', cdrr, columns):
ohlc['H-1/C'] = ohlc['High'].shift(1) / ohlc['Close']
if check_cdrr('L-1/C', cdrr, columns):
ohlc['L-1/C'] = ohlc['Low'].shift(1) / ohlc['Close']
if check_cdrr('C-1/C', cdrr, columns):
ohlc['C-1/C'] = ohlc['Close'].shift(1) / ohlc['Close']
if check_cdrr('PositiveLine', cdrr, columns):
ohlc['PositiveLine'] = (ohlc['Open'] <= ohlc['Close']) * 1
if check_cdrr('NegativeLine', cdrr, columns):
ohlc['NegativeLine'] = (ohlc['Open'] > ohlc['Close']) * (-1)
if check_cdrr('SignLine', cdrr, columns):
ohlc['SignLine'] = ohlc['PositiveLine'] + ohlc['NegativeLine']
if check_cdrr('3SignLine', cdrr, columns):
ohlc['3SignLine'] = ohlc['SignLine'].rolling(3).sum()
if check_cdrr('6SignLine', cdrr, columns):
ohlc['6SignLine'] = ohlc['SignLine'].rolling(6).sum()
if check_cdrr('12SignLine', cdrr, columns):
ohlc['12SignLine'] = ohlc['SignLine'].rolling(12).sum()
if check_cdrr('BB', cdrr, columns):
upper, middle, lower = talib.BBANDS(ohlc['Close'],
timeperiod=bbtp,
nbdevup=2,
nbdevdn=2,
matype=0)
ohlc['BBUpper'] = upper
ohlc['BBMiddle'] = middle
ohlc['BBLower'] = lower
if check_cdrr('BBOver', cdrr, columns):
ohlc['BBOver'] = ohlc['BBUpper'] < ohlc['Close']
if check_cdrr('BBUnder', cdrr, columns):
ohlc['BBUnder'] = ohlc['Close'] < ohlc['BBLower']
if check_cdrr('MACDS', cdrr, columns):
ohlc['MACDS'] = talib.MACD(ohlc['Close'],
fastperiod=6,
slowperiod=19,
signalperiod=9)[0]
if check_cdrr('MACDM', cdrr, columns):
ohlc['MACDM'] = talib.MACD(ohlc['Close'],
fastperiod=12,
slowperiod=26,
signalperiod=9)[0]
if check_cdrr('MACDL', cdrr, columns):
ohlc['MACDL'] = talib.MACD(ohlc['Close'],
fastperiod=19,
slowperiod=39,
signalperiod=9)[0]
if check_cdrr('MACDS/O', cdrr, columns):
ohlc['MACDS/O'] = talib.MACD(ohlc['C/O'],
fastperiod=6,
slowperiod=19,
signalperiod=9)[0]
if check_cdrr('MACDM/O', cdrr, columns):
ohlc['MACDM/O'] = talib.MACD(ohlc['C/O'],
fastperiod=12,
slowperiod=26,
signalperiod=9)[0]
if check_cdrr('MACDL/O', cdrr, columns):
ohlc['MACDL/O'] = talib.MACD(ohlc['C/O'],
fastperiod=19,
slowperiod=39,
signalperiod=9)[0]
if check_cdrr('ATR', cdrr, columns):
ohlc['ATR'] = talib.ATR(ohlc['High'],
ohlc['Low'],
ohlc['Close'],
timeperiod=14)
if check_cdrr('ATR/O', cdrr, columns):
ohlc['ATR/O'] = talib.ATR(ohlc['H/O'],
ohlc['L/O'],
ohlc['O/O'],
timeperiod=14)
if check_cdrr('RSI', cdrr, columns):
ohlc['RSI'] = talib.RSI(ohlc['Close'], timeperiod=14)
if check_cdrr('Time', cdrr, columns):
#time relation
tr = {
'60': [60, 60],
'300': [300, 12],
'3600': [3600, 24],
}
period = str(int(ohlc['timestamp'].diff().mode()))
ohlc['Time'] = ohlc['timestamp'] // tr[period][0] % tr[period][1]
if check_cdrr('Over', cdrr, columns):
ohlc['Over'] = ((ohlc['Open'] * (1 + over_rate) <= ohlc['High']) *
(ohlc['Open'] *
(1 - over_rate) >= ohlc['Low'])).astype(bool)
if check_cdrr('1moku-base', cdrr, columns):
ohlc['1moku-base'] = (ohlc['High'].rolling(26).max() +
ohlc['Low'].rolling(26).min()) / 2
if check_cdrr('1moku-conversion', cdrr, columns):
ohlc['1moku-conversion'] = (ohlc['High'].rolling(9).max() +
ohlc['Low'].rolling(9).min()) / 2
if check_cdrr('1moku-precedent-span1', cdrr, columns):
ohlc['1moku-precedent-span1'] = (
(ohlc['1moku-base'] + ohlc['1moku-conversion']) / 2).shift(26)
if check_cdrr('1moku-precedent-span2', cdrr, columns):
ohlc['1moku-precedent-span2'] = (
(ohlc['High'].rolling(52).max() + ohlc['Low'].rolling(52).min()) /
2).shift(26)
if check_cdrr('1moku-late-span', cdrr, columns):
ohlc['1moku-late-span'] = ohlc['Close'].shift(-26)
if check_cdrr('1moku-signal1', cdrr, columns):
ohlc['1moku-signal1'] = (
(ohlc['1moku-conversion'] - ohlc['1moku-base'] > 0) * 1).diff(1)
if check_cdrr('1moku-signal2', cdrr, columns):
ohlc['1moku-signal2'] = (ohlc['1moku-conversion'] - ohlc['1moku-base']
> 0) * 1
if check_cdrr('1moku-signal3', cdrr, columns):
ohlc['1moku-signal3'] = (
((ohlc['Close'] > ohlc['1moku-precedent-span1']) * 1) *
((ohlc['Close'] > ohlc['1moku-precedent-span2']) * 1)).diff(1)
if check_cdrr('1moku-signal4', cdrr, columns):
ohlc['1moku-signal4'] = (
(ohlc['Close'] > ohlc['1moku-precedent-span1']) * 1) * (
(ohlc['Close'] > ohlc['1moku-precedent-span2']) * 1)
if check_cdrr('1moku-signal5', cdrr, columns):
ohlc['1moku-signal5'] = (
((ohlc['Close'] < ohlc['1moku-precedent-span1']) * 1) *
((ohlc['Close'] < ohlc['1moku-precedent-span2']) * 1)).diff(1)
if check_cdrr('1moku-signal6', cdrr, columns):
ohlc['1moku-signal6'] = (
(ohlc['Close'] < ohlc['1moku-precedent-span1']) * 1) * (
(ohlc['Close'] < ohlc['1moku-precedent-span2']) * 1)
if check_cdrr('Target_Open', cdrr, columns):
ohlc['Target_Open'] = ohlc['Open'].shift(-1)
if check_cdrr('Target_High', cdrr, columns):
ohlc['Target_High'] = ohlc['High'].shift(-1)
if check_cdrr('Target_Low', cdrr, columns):
ohlc['Target_Low'] = ohlc['Low'].shift(-1)
if check_cdrr('Target_Close', cdrr, columns):
ohlc['Target_Close'] = ohlc['Close'].shift(-1)
if check_cdrr('Target_O/O', cdrr, columns):
ohlc['Target_O/O'] = ohlc['Target_Open'] / ohlc['Open']
if check_cdrr('Target_H/O', cdrr, columns):
ohlc['Target_H/O'] = ohlc['Target_High'] / ohlc['Open']
if check_cdrr('Target_L/O', cdrr, columns):
ohlc['Target_L/O'] = ohlc['Target_Low'] / ohlc['Open']
if check_cdrr('Target_C/O', cdrr, columns):
ohlc['Target_C/O'] = ohlc['Target_Close'] / ohlc['Open']
if check_cdrr('Target_O/C', cdrr, columns):
ohlc['Target_O/C'] = ohlc['Target_Open'] / ohlc['Close']
if check_cdrr('Target_H/C', cdrr, columns):
ohlc['Target_H/C'] = ohlc['Target_High'] / ohlc['Close']
if check_cdrr('Target_L/C', cdrr, columns):
ohlc['Target_L/C'] = ohlc['Target_Low'] / ohlc['Close']
if check_cdrr('Target_C/C', cdrr, columns):
ohlc['Target_C/C'] = ohlc['Target_Close'] / ohlc['Close']
if check_cdrr('Target_Over', cdrr, columns):
ohlc['Target_Over'] = ohlc['Over'].shift(-1).astype(bool)
ohlc = ohlc[columns]
if drop:
ohlc = ohlc.dropna()
return ohlc
def strategy_bandwidth(series, is_backtest=True):
'''策略：布林通道宽度'''
# 实时线
close = np.cumsum(series).astype(float)
# 布林线
upperband, middleband, lowerband = ta.BBANDS(close,
timeperiod=5,
nbdevup=2,
nbdevdn=2,
matype=0)
bandwidth = upperband - lowerband
bandwidth_mean = np.nan_to_num(bandwidth).mean()
current_signal = 0 # 当前信号：无
# 计算趋势
mean = ta.MA(close, timeperiod=5, matype=0)
#mean = np.nan_to_num(mean) #Nan转换为0
diff_mean = np.diff(mean) # 注意：长度减1
#trend = np.where(diff_mean > 0, 1, -1)
trend = []
for i in range(len(diff_mean)):
if diff_mean[i] == 0:
trend.append(trend[-1])
else:
trend.append([-1, 1][diff_mean[i] > 0])
# 如果是回测
if is_backtest:
error_periods = 0 # 哨兵：之前错误期数
earn = [] # 收益记录
order = [] # 投注资金记录
for i in range(2, len(series)):
# 判断是否有信号，什么信号
#if bandwidth[i-1] < bandwidth[i]: # 增：趋势区间
if bandwidth[i - 1] > bandwidth_mean: # 宽：趋势区间
#current_signal = series[-1] # 与前一期相同
current_signal = trend[i - 2] # 与前一期趋势相同
else: # 窄：震荡区间
current_signal = -series[i - 1] # 与前一期相反（注意不一样！！）
# 若无信号，跳过，不做任何操作
#if current_signal == 0: continue
if current_signal == 0:
order.append(0)
earn.append(0)
continue
# 看料
buy = current_signal
# 下单
order.append(times[error_periods])
# 判断对错，计算收益
if buy == series[i]:
error_periods = 0 # 错误期数置零
earn.append(odds * order[-1]) # 赔率odds=1.96
else:
error_periods += 1 # 错误期数加一
earn.append(0)
print('投注记录:\n', order[-show_period:], '\n合计：',
sum(order[-show_period:]))
print('收益记录:\n', earn[-show_period:], '\n合计：',
sum(earn[-show_period:]))
print('收益率：', (sum(earn[-show_period:]) - sum(order[-show_period:])) /
sum(order[-show_period:]))
return order, earn, bandwidth, bandwidth_mean
# 否则只显示最后一期
else:
# 应肖的需求，显示下期结果
if bandwidth[-1] > mean: # 宽：趋势区间
current_signal = trend[-1] # 与前一期（趋势）相同
else: # 窄：震荡区间
current_signal = -series[-1] # 与前一期相反
print('=' * 30)
print('通道宽度： ', end='')
print_red_color_text(current_signal)
print('=' * 30)
return current_signal
def get_factors(index,
Open,
Close,
High,
Low,
Volume,
rolling = 26,
drop=False,
normalization=True):
tmp = pd.DataFrame()
tmp['tradeTime'] = index
#累积/派发线（Accumulation / Distribution Line，该指标将每日的成交量通过价格加权累计，
#用以计算成交量的动量。属于趋势型因子
tmp['AD'] = talib.AD(High, Low, Close, Volume)
# 佳庆指标（Chaikin Oscillator），该指标基于AD曲线的指数移动均线而计算得到。属于趋势型因子
tmp['ADOSC'] = talib.ADOSC(High, Low, Close, Volume, fastperiod=3, slowperiod=10)
# 平均动向指数，DMI因子的构成部分。属于趋势型因子
tmp['ADX'] = talib.ADX(High, Low, Close,timeperiod=14)
# 相对平均动向指数，DMI因子的构成部分。属于趋势型因子
tmp['ADXR'] = talib.ADXR(High, Low, Close,timeperiod=14)
# 绝对价格振荡指数
tmp['APO'] = talib.APO(Close, fastperiod=12, slowperiod=26)
# Aroon通过计算自价格达到近期最高值和最低值以来所经过的期间数，帮助投资者预测证券价格从趋势到区域区域或反转的变化，
#Aroon指标分为Aroon、AroonUp和AroonDown3个具体指标。属于趋势型因子
tmp['AROONDown'], tmp['AROONUp'] = talib.AROON(High, Low,timeperiod=14)
tmp['AROONOSC'] = talib.AROONOSC(High, Low,timeperiod=14)
# 均幅指标（Average TRUE Ranger），取一定时间周期内的股价波动幅度的移动平均值，
#是显示市场变化率的指标，主要用于研判买卖时机。属于超买超卖型因子。
tmp['ATR14']= talib.ATR(High, Low, Close, timeperiod=14)
tmp['ATR6']= talib.ATR(High, Low, Close, timeperiod=6)
# 布林带
tmp['Boll_Up'],tmp['Boll_Mid'],tmp['Boll_Down']= talib.BBANDS(Close, timeperiod=20, nbdevup=2, nbdevdn=2, matype=0)
# 均势指标
tmp['BOP'] = talib.BOP(Open, High, Low, Close)
#5日顺势指标（Commodity Channel Index），专门测量股价是否已超出常态分布范围。属于超买超卖型因子。
tmp['CCI5'] = talib.CCI(High, Low, Close, timeperiod=5)
tmp['CCI10'] = talib.CCI(High, Low, Close, timeperiod=10)
tmp['CCI20'] = talib.CCI(High, Low, Close, timeperiod=20)
tmp['CCI88'] = talib.CCI(High, Low, Close, timeperiod=88)
# 钱德动量摆动指标（Chande Momentum Osciliator），与其他动量指标摆动指标如相对强弱指标（RSI）和随机指标（KDJ）不同，
# 钱德动量指标在计算公式的分子中采用上涨日和下跌日的数据。属于超买超卖型因子
tmp['CMO_Close'] = talib.CMO(Close,timeperiod=14)
tmp['CMO_Open'] = talib.CMO(Close,timeperiod=14)
# DEMA双指数移动平均线
tmp['DEMA6'] = talib.DEMA(Close, timeperiod=6)
tmp['DEMA12'] = talib.DEMA(Close, timeperiod=12)
tmp['DEMA26'] = talib.DEMA(Close, timeperiod=26)
# DX 动向指数
tmp['DX'] = talib.DX(High, Low, Close,timeperiod=14)
# EMA 指数移动平均线
tmp['EMA6'] = talib.EMA(Close, timeperiod=6)
tmp['EMA12'] = talib.EMA(Close, timeperiod=12)
tmp['EMA26'] = talib.EMA(Close, timeperiod=26)
# KAMA 适应性移动平均线
tmp['KAMA'] = talib.KAMA(Close, timeperiod=30)
# MACD
tmp['MACD_DIF'],tmp['MACD_DEA'],tmp['MACD_bar'] = talib.MACD(Close, fastperiod=12, slowperiod=24, signalperiod=9)
# 中位数价格 不知道是什么意思
tmp['MEDPRICE'] = talib.MEDPRICE(High, Low)
# 负向指标 负向运动
tmp['MiNUS_DI'] = talib.MINUS_DI(High, Low, Close,timeperiod=14)
tmp['MiNUS_DM'] = talib.MINUS_DM(High, Low,timeperiod=14)
# 动量指标（Momentom Index），动量指数以分析股价波动的速度为目的，研究股价在波动过程中各种加速，
#减速，惯性作用以及股价由静到动或由动转静的现象。属于趋势型因子
tmp['MOM'] = talib.MOM(Close, timeperiod=10)
# 归一化平均值范围
tmp['NATR'] = talib.NATR(High, Low, Close,timeperiod=14)
# OBV 能量潮指标（On Balance Volume，OBV），以股市的成交量变化来衡量股市的推动力，
#从而研判股价的走势。属于成交量型因子
tmp['OBV'] = talib.OBV(Close, Volume)
# PLUS_DI 更向指示器
tmp['PLUS_DI'] = talib.PLUS_DI(High, Low, Close,timeperiod=14)
tmp['PLUS_DM'] = talib.PLUS_DM(High, Low, timeperiod=14)
# PPO 价格振荡百分比
tmp['PPO'] = talib.PPO(Close, fastperiod=6, slowperiod= 26, matype=0)
# ROC 6日变动速率（Price Rate of Change），以当日的收盘价和N天前的收盘价比较，
#通过计算股价某一段时间内收盘价变动的比例，应用价格的移动比较来测量价位动量。属于超买超卖型因子。
tmp['ROC6'] = talib.ROC(Close, timeperiod=6)
tmp['ROC20'] = talib.ROC(Close, timeperiod=20)
#12日量变动速率指标（Volume Rate of Change），以今天的成交量和N天前的成交量比较，
#通过计算某一段时间内成交量变动的幅度，应用成交量的移动比较来测量成交量运动趋向，
#达到事先探测成交量供需的强弱，进而分析成交量的发展趋势及其将来是否有转势的意愿，
#属于成交量的反趋向指标。属于成交量型因子
tmp['VROC6'] = talib.ROC(Volume, timeperiod=6)
tmp['VROC20'] = talib.ROC(Volume, timeperiod=20)
# ROC 6日变动速率（Price Rate of Change），以当日的收盘价和N天前的收盘价比较，
#通过计算股价某一段时间内收盘价变动的比例，应用价格的移动比较来测量价位动量。属于超买超卖型因子。
tmp['ROCP6'] = talib.ROCP(Close, timeperiod=6)
tmp['ROCP20'] = talib.ROCP(Close, timeperiod=20)
#12日量变动速率指标（Volume Rate of Change），以今天的成交量和N天前的成交量比较，
#通过计算某一段时间内成交量变动的幅度，应用成交量的移动比较来测量成交量运动趋向，
#达到事先探测成交量供需的强弱，进而分析成交量的发展趋势及其将来是否有转势的意愿，
#属于成交量的反趋向指标。属于成交量型因子
tmp['VROCP6'] = talib.ROCP(Volume, timeperiod=6)
tmp['VROCP20'] = talib.ROCP(Volume, timeperiod=20)
# RSI
tmp['RSI'] = talib.RSI(Close, timeperiod=14)
# SAR 抛物线转向
tmp['SAR'] = talib.SAR(High, Low, acceleration=0.02, maximum=0.2)
# TEMA
tmp['TEMA6'] = talib.TEMA(Close, timeperiod=6)
tmp['TEMA12'] = talib.TEMA(Close, timeperiod=12)
tmp['TEMA26'] = talib.TEMA(Close, timeperiod=26)
# TRANGE 真实范围
tmp['TRANGE'] = talib.TRANGE(High, Low, Close)
# TYPPRICE 典型价格
tmp['TYPPRICE'] = talib.TYPPRICE(High, Low, Close)
# TSF 时间序列预测
tmp['TSF'] = talib.TSF(Close, timeperiod=14)
# ULTOSC 极限振子
tmp['ULTOSC'] = talib.ULTOSC(High, Low, Close, timeperiod1=7, timeperiod2=14, timeperiod3=28)
# 威廉指标
tmp['WILLR'] = talib.WILLR(High, Low, Close, timeperiod=14)
# 标准化
if normalization:
factors_list = tmp.columns.tolist()[1:]
if rolling >= 26:
for i in factors_list:
tmp[i] = (tmp[i] - tmp[i].rolling(window=rolling, center=False).mean()) /tmp[i].rolling(window=rolling, center=False).std()
elif rolling < 26 & rolling > 0:
print ('Recommended rolling range greater than 26')
elif rolling <=0:
for i in factors_list:
tmp[i] = (tmp[i] - tmp[i].mean())/tmp[i].std()
if drop:
tmp.dropna(inplace=True)
tmp.set_index('tradeTime', inplace=True)
return tmp
def ALL():
def candles(window, period):
period = str(period)
window = WINDOW[0]
url = 'https://api.hitbtc.com/api/2/public/candles/BTCUSD?period=' + window + '&limit=' + period
r = requests.get(url)
retorno = {}
resposta = r.json()
MEAN, TIMESTAMP = [], []
for i in range(len(np.array(resposta))):
MEAN.append((float(np.array(resposta)[i]['max']) +
float(np.array(resposta)[i]['min'])) / 2)
TIMESTAMP.append(np.array(resposta)[i]['timestamp'])
return list(MEAN), TIMESTAMP
def vol(window, period):
period = str(period)
window = WINDOW[0]
url = 'https://api.hitbtc.com/api/2/public/candles/BTCUSD?period=' + window + '&limit=' + period
r = requests.get(url)
retorno = {}
resposta = r.json()
VOL, TIMESTAMP = [], []
for i in range(len(np.array(resposta))):
VOL.append((float(np.array(resposta)[i]['volume'])))
TIMESTAMP.append(np.array(resposta)[i]['timestamp'])
return list(VOL), TIMESTAMP
from scipy.signal import savgol_filter
WINDOW = ['M1', 'M3', 'M5', 'M15', 'M30', 'H1', 'H4', 'D1', 'W1']
window = 1000
lista, timestamp = candles(WINDOW[0], window)
w = 13
#x = moving_average(lista, w)
x = savgol_filter(lista, 63, 3)
#x = savgol_filter(lista, 123, 3)
d_window = 10
dydx = x[1:] - x[:-1]
val = []
val.append(0)
index = []
l = 130
j = 190
for i in range(0, len(lista) - l, l):
a = min(lista[i:i + j])
if a != val[-1]:
index.append(int(np.argwhere(np.array(lista) == a)[0]))
val.append(a)
val.pop(0)
a = min(lista[-80:])
val.append(a)
index.append(int(np.argwhere(np.array(lista) == a)))
y_s = val
x_s = index
val = []
val.append(0)
index = []
for i in range(0, len(lista) - l, l):
a = max(lista[i:i + j])
if a != val[-1]:
index.append(int(np.argwhere(np.array(lista) == a)[0]))
val.append(a)
val.pop(0)
a = max(lista[-80:])
val.append(a)
index.append(int(np.argwhere(np.array(lista) == a)))
'''
indexMAX = []
indexMIN = []
for i in range(len(x)-1):
if dydx[i]<0 and sum(dydx[i+1:i+50]>0)==49:
indexMAX.append(i)
if dydx[i]>0 and sum(dydx[i+1:i+50]<0)==49:
indexMIN.append(i)
#plt.scatter( np.array(indexMIN) , np.array(lista)[np.array(indexMIN)] )
#plt.plot(x)
#plt.plot(lista[w:])
y_s = np.array(lista)[list(np.array(indexMIN)+int(w/2))]
x_s = np.array(indexMIN)+w/2
#A = (y_s[0]-y_s[1])/(x_s[0]-x_s[1])
#B = y_s[0]-A*x_s[0]
#X_s=np.array(range(-10,200))
#Y_s = A*X+B
y_r = np.array(lista)[list(np.array(indexMAX)+int(w/2))]
x_r = np.array(indexMAX)+w/2
A = (y_r[0]-y_r[1])/(x_r[0]-x_r[1])
B = y_r[0]-A*x_r[0]
X_r = np.array(range(-10,200))
Y_r = A*X+B
'''
y_r = val
x_r = index
plt.figure(figsize=(10, 5))
plt.plot(lista)
c_gap = 1
a_r = []
a_s = []
for i in range(0, len(y_r) - c_gap, c_gap):
A = (y_r[i] - y_r[i + c_gap]) / (x_r[i] - x_r[i + c_gap] + 0.001)
B = y_r[i] - A * x_r[i]
a_r.append(A)
X_r1 = np.array(range(int(x_r[i]), int(x_r[i + c_gap])))
Y_r1 = A * X_r1 + B
#Y_r1 = A*X+B
plt.plot(X_r1, Y_r1, color='green')
if i == list(range(0, len(y_r) - c_gap, c_gap))[-1]:
X_r1 = np.array(range(int(x_r[i]), int(len(lista) + 60)))
Y_r1 = A * X_r1 + B
plt.plot(X_r1, Y_r1, color='green')
c_gap = 1
for i in range(0, len(y_r) - c_gap, c_gap):
try:
A = (y_s[i] - y_s[i + c_gap]) / (x_s[i] - x_s[i + c_gap])
a_s.append(A)
B = y_s[i] - A * x_s[i]
X_s = np.array(range(int(x_s[i]), int(x_s[i + c_gap])))
Y_s = A * X_s + B
plt.plot(X_s, Y_s, color='red')
if i == list(range(0, len(y_r) - c_gap, c_gap))[-1]:
X_s = np.array(range(int(x_s[i]), int(len(lista) + 60)))
Y_s = A * X_s + B
plt.plot(X_s, Y_s, color='red')
except:
i = len(x_s) - 1 - c_gap
A = (y_s[i] - y_s[i + c_gap]) / (x_s[i] - x_s[i + c_gap])
a_s.append(A)
B = y_s[i] - A * x_s[i]
X_s = np.array(range(int(x_s[i]), int(len(lista) + 60)))
Y_s = A * X_s + B
plt.plot(X_s, Y_s, color='red')
plt.plot(x)
plt.grid()
plt.savefig('dydx/Sup_Res' + str(time.time()).split('.')[0] + '.png')
plt.title('Support/Resistence')
plt.xlim(0, len(lista) + 60)
plt.savefig('../INDICATORS/' + 'support_res' + str(time.time()) + '.png')
plt.cla()
plt.clf()
#plt.plot(X_r,Y_r)
#plt.plot(X_s,Y_s)
#plt.plot(x)
#plt.plot(lista)
#plt.xlim(20,115)
#plt.ylim(min(lista)-20,max(lista)+20)
#plt.legend(['MA 10 periodos','preço'])
#plt.figure(1)
#plt.plot(a_s)
#plt.plot(a_r)
price = np.array(lista)
price = (price - price.mean())
price = price / max(price)
dydx = savgol_filter(dydx, 63, 3)
price = savgol_filter(price, 63, 3)
plt.figure(66)
plt.figure(figsize=(10, 4))
#plt.figure(3,figsize=(6,2))
plt.plot(dydx / max(abs(dydx)))
plt.plot(price)
plt.plot(np.zeros(len(dydx)))
plt.legend(['dydx', 'price'], loc='best')
plt.savefig('dydx/Derivate' + str(time.time()).split('.')[0] + '.png')
plt.title('Derivate')
plt.savefig('../INDICATORS/' + 'phase_in_t' + str(time.time()) + '.png')
plt.cla()
plt.clf()
r = int(-1)
s = int(-1)
plt.figure(50)
plt.figure(figsize=(10, 5))
cem = y_r[r] - y_s[s]
fibonacci = np.ones((400, 5)) * np.array(
[0.236, 0.382, 0.50, 0.618, 0.786]) * cem + y_s[-1]
x_f = list(range(x_s[s], x_s[s] + len(fibonacci), 1))
#plt.plot(x)
plt.plot(lista)
plt.plot(x_f, fibonacci)
plt.ylim(y_r[r] - 200, y_s[s] + 200)
plt.legend(['price', '0.236', '0.382', '0.50', '0.618', '0.786'])
plt.grid()
plt.title('Fibonacci')
plt.savefig('dydx/Fibonacci' + str(time.time()).split('.')[0] + '.png')
plt.savefig('../INDICATORS/' + 'Fibonacci_' + str(time.time()) + '.png')
import numpy
import talib
output = talib.SMA(np.array(lista))
from talib import MA_Type
upper, middle, lower = talib.BBANDS(np.array(lista), 44)
#talib.MOM(lista, timeperiod=5)
plt.figure(60)
plt.figure(figsize=(10, 5))
plt.plot(upper)
plt.plot(lower)
plt.plot(middle)
plt.plot(np.array(lista))
plt.grid()
volume, timestamp = vol(WINDOW[0], window)
plt.legend(['upper', 'lower', 'mid', 'price'])
plt.title('Bollinger')
plt.savefig('../INDICATORS/' + 'dydxBollinger' +
str(time.time()).split('.')[0] + '.png')
plt.cla()
plt.clf()
price = np.array(lista)
std = np.ones(len(volume)) * np.array(volume).std()
mean = np.ones(len(volume)) * np.array(volume).mean()
plt.figure(70)
plt.figure(figsize=(10, 5))
plt.plot(volume)
plt.plot(mean + 2 * std)
plt.plot((price - (min(price))))
plt.grid()
plt.legend(['volume', 'mean+2sigma', 'price (normalized)'])
plt.title('Anomaly')
plt.savefig('dydx/anomaly' + str(time.time()).split('.')[0] + '.png')
plt.savefig('../INDICATORS/' + 'dydx_' + str(time.time()) + '.png')
plt.cla()
plt.clf()
plt.figure(777)
#plt.figure(figsize= (5,5))
dyd2x = dydx[1:] - dydx[:-1]
state1 = (dydx / max(abs(dydx)))[1:]
state2 = dyd2x
lenght = 200
state2 = state2[-lenght:]
state1 = state1[-lenght:]
plt.grid()
t = np.linspace(0, 1, num=len(state1))
dotcolors = [(0.0, 0.0, 0.0, a) for a in t]
plt.scatter(state1, state2, c=dotcolors, s=30, edgecolors='None')
plt.plot(state1, state2, linewidth=0.4, c='black')
[min(state2), max(state2)]
[min(state1), max(state2)]
[0, 0]
plt.plot([min(state1), max(state1)], [0, 0], color='black')
plt.plot([0, 0], [min(state2), max(state2)], color='black')
plt.title('Phase Diagram')
plt.savefig('../INDICATORS/' + 'phase_mean' + str(time.time()) + '.png')
plt.xlabel("y'(x)")
plt.ylabel("y''(x)")
plt.cla()
plt.clf()
plt.figure(999)
price = np.array(lista)
price = (price - price.mean())
price = price / max(price)
price = savgol_filter(price, 63, 3)
#plt.figure(figsize= (5,5))
dyd2x = dydx[1:] - dydx[:-1]
state1 = (dydx / max(abs(dydx)))[1:]
state2 = price
lenght = 300
state2 = state2[-lenght:]
state1 = state1[-lenght:]
plt.grid()
t = np.linspace(0, 1, num=len(state1))
dotcolors = [(0.0, 0.0, 0.0, a) for a in t]
plt.scatter(state2, state1, c=dotcolors, s=30, edgecolors='None')
plt.plot(state2, state1, linewidth=0.4, c='black')
plt.title("Phase Diagram price y'(x)")
plt.xlabel("x")
plt.ylabel("y'(x)")
plt.savefig('../INDICATORS/' + 'phase' + str(time.time()) + '.png')
[min(state2), max(state2)]
[min(state1), max(state2)]
[0, 0]
plt.plot([0, 0], [min(state1), max(state1)], color='black')
plt.plot([min(state2), max(state2)], [0, 0], color='black')
plt.cla()
plt.clf()
def _extract_feature(candle, params):
o = candle.open
h = candle.high
l = candle.low
c = candle.close
v = candle.volume
# OHLCV
features = pd.DataFrame()
features['open'] = o
features['high'] = h
features['low'] = l
features['close'] = c
features['volume'] = v
# RSI
features['rsi_s'] = ta.RSI(c, timeperiod=params['rsi_timeperiod_s'])
features['rsi_m'] = ta.RSI(c, timeperiod=params['rsi_timeperiod_m'])
features['rsi_l'] = ta.RSI(c, timeperiod=params['rsi_timeperiod_l'])
# STOCHF
fastk, fastd = ta.STOCHF(h,
l,
c,
fastk_period=params['stockf_fastk_period_s'],
fastd_period=params['stockf_fastd_period_s'],
fastd_matype=ta.MA_Type.T3)
change_stockf = calc_change(fastk, fastd)
change_stockf.index = fastk.index
features['fastk_s'] = fastk
features['fastd_s'] = fastd
features['fast_change_s'] = change_stockf
fastk, fastd = ta.STOCHF(h,
l,
c,
fastk_period=params['stockf_fastk_period_m'],
fastd_period=params['stockf_fastd_period_m'],
fastd_matype=ta.MA_Type.T3)
change_stockf = calc_change(fastk, fastd)
change_stockf.index = fastk.index
features['fastk_m'] = fastk
features['fastd_m'] = fastd
features['fast_change_m'] = change_stockf
fastk, fastd = ta.STOCHF(h,
l,
c,
fastk_period=params['stockf_fastk_period_l'],
fastd_period=params['stockf_fastd_period_l'],
fastd_matype=ta.MA_Type.T3)
change_stockf = calc_change(fastk, fastd)
change_stockf.index = fastk.index
features['fastk_l'] = fastk
features['fastd_l'] = fastd
features['fast_change_l'] = change_stockf
# WILLR
features['willr_s'] = ta.WILLR(h,
l,
c,
timeperiod=params['willr_timeperiod_s'])
features['willr_m'] = ta.WILLR(h,
l,
c,
timeperiod=params['willr_timeperiod_m'])
features['willr_l'] = ta.WILLR(h,
l,
c,
timeperiod=params['willr_timeperiod_l'])
# MACD
macd, macdsignal, macdhist = ta.MACDEXT(
c,
fastperiod=params['macd_fastperiod_s'],
slowperiod=params['macd_slowperiod_s'],
signalperiod=params['macd_signalperiod_s'],
fastmatype=ta.MA_Type.T3,
slowmatype=ta.MA_Type.T3,
signalmatype=ta.MA_Type.T3)
change_macd = calc_change(macd, macdsignal)
change_macd.index = macd.index
features['macd_s'] = macd
features['macdsignal_s'] = macdsignal
features['macdhist_s'] = macdhist
features['change_macd_s'] = change_macd
macd, macdsignal, macdhist = ta.MACDEXT(
c,
fastperiod=params['macd_fastperiod_m'],
slowperiod=params['macd_slowperiod_m'],
signalperiod=params['macd_signalperiod_m'],
fastmatype=ta.MA_Type.T3,
slowmatype=ta.MA_Type.T3,
signalmatype=ta.MA_Type.T3)
change_macd = calc_change(macd, macdsignal)
change_macd.index = macd.index
features['macd_m'] = macd
features['macdsignal_m'] = macdsignal
features['macdhist_m'] = macdhist
features['change_macd_m'] = change_macd
# ROCP
rocp = ta.ROCP(c, timeperiod=params['rocp_timeperiod_s'])
change_rocp = calc_change(
rocp, pd.Series(np.zeros(len(candle)), index=candle.index))
change_rocp.index = rocp.index
features['rocp_s'] = rocp
features['change_rocp_s'] = change_rocp
rocp = ta.ROCP(c, timeperiod=params['rocp_timeperiod_m'])
change_rocp = calc_change(
rocp, pd.Series(np.zeros(len(candle)), index=candle.index))
change_rocp.index = rocp.index
features['rocp_m'] = rocp
features['change_rocp_m'] = change_rocp
rocp = ta.ROCP(c, timeperiod=params['rocp_timeperiod_l'])
change_rocp = calc_change(
rocp, pd.Series(np.zeros(len(candle)), index=candle.index))
change_rocp.index = rocp.index
features['rocp_l'] = rocp
features['change_rocp_l'] = change_rocp
# ボリンジャーバンド
bb_upper, bb_middle, bb_lower = ta.BBANDS(
c,
timeperiod=params['bbands_timeperiod_s'],
nbdevup=params['bbands_nbdevup_s'],
nbdevdn=params['bbands_nbdevdn_s'],
matype=ta.MA_Type.T3)
bb_trend1 = pd.Series(np.zeros(len(candle)), index=candle.index)
bb_trend1[c > bb_upper] = 1
bb_trend1[c < bb_lower] = -1
bb_trend2 = pd.Series(np.zeros(len(candle)), index=candle.index)
bb_trend2[c > bb_middle] = 1
bb_trend2[c < bb_middle] = -1
features['bb_upper_s'] = bb_upper
features['bb_middle_s'] = bb_middle
features['bb_lower_s'] = bb_lower
features['bb_trend1_s'] = bb_trend1
features['bb_trend2_s'] = bb_trend2
bb_upper, bb_middle, bb_lower = ta.BBANDS(
c,
timeperiod=params['bbands_timeperiod_m'],
nbdevup=params['bbands_nbdevup_m'],
nbdevdn=params['bbands_nbdevdn_m'],
matype=ta.MA_Type.T3)
bb_trend1 = pd.Series(np.zeros(len(candle)), index=candle.index)
bb_trend1[c > bb_upper] = 1
bb_trend1[c < bb_lower] = -1
bb_trend2 = pd.Series(np.zeros(len(candle)), index=candle.index)
bb_trend2[c > bb_middle] = 1
bb_trend2[c < bb_middle] = -1
features['bb_upper_m'] = bb_upper
features['bb_middle_m'] = bb_middle
features['bb_lower_m'] = bb_lower
features['bb_trend1_m'] = bb_trend1
features['bb_trend2_m'] = bb_trend2
bb_upper, bb_middle, bb_lower = ta.BBANDS(
c,
timeperiod=params['bbands_timeperiod_l'],
nbdevup=params['bbands_nbdevup_l'],
nbdevdn=params['bbands_nbdevdn_l'],
matype=ta.MA_Type.T3)
bb_trend1 = pd.Series(np.zeros(len(candle)), index=candle.index)
bb_trend1[c > bb_upper] = 1
bb_trend1[c < bb_lower] = -1
bb_trend2 = pd.Series(np.zeros(len(candle)), index=candle.index)
bb_trend2[c > bb_middle] = 1
bb_trend2[c < bb_middle] = -1
features['bb_upper_l'] = bb_upper
features['bb_middle_l'] = bb_middle
features['bb_lower_l'] = bb_lower
features['bb_trend1_l'] = bb_trend1
features['bb_trend2_l'] = bb_trend2
# 出来高
features['obv'] = ta.OBV(c, v)
features['ad'] = ta.AD(h, l, c, v)
features['adosc_s'] = ta.ADOSC(h,
l,
c,
v,
fastperiod=params['fastperiod_adosc_s'],
slowperiod=params['slowperiod_adosc_s'])
# ローソク足
features['CDL2CROWS'] = ta.CDL2CROWS(o, h, l, c)
features['CDL3BLACKCROWS'] = ta.CDL3BLACKCROWS(o, h, l, c)
features['CDL3INSIDE'] = ta.CDL3INSIDE(o, h, l, c)
features['CDL3LINESTRIKE'] = ta.CDL3LINESTRIKE(o, h, l, c)
features['CDL3OUTSIDE'] = ta.CDL3OUTSIDE(o, h, l, c)
features['CDL3STARSINSOUTH'] = ta.CDL3STARSINSOUTH(o, h, l, c)
features['CDL3WHITESOLDIERS'] = ta.CDL3WHITESOLDIERS(o, h, l, c)
features['CDLABANDONEDBABY'] = ta.CDLABANDONEDBABY(o,
h,
l,
c,
penetration=0)
features['CDLADVANCEBLOCK'] = ta.CDLADVANCEBLOCK(o, h, l, c)
features['CDLBELTHOLD'] = ta.CDLBELTHOLD(o, h, l, c)
features['CDLBREAKAWAY'] = ta.CDLBREAKAWAY(o, h, l, c)
features['CDLCLOSINGMARUBOZU'] = ta.CDLCLOSINGMARUBOZU(o, h, l, c)
features['CDLCONCEALBABYSWALL'] = ta.CDLCONCEALBABYSWALL(o, h, l, c)
features['CDLCOUNTERATTACK'] = ta.CDLCOUNTERATTACK(o, h, l, c)
features['CDLDARKCLOUDCOVER'] = ta.CDLDARKCLOUDCOVER(o,
h,
l,
c,
penetration=0)
features['CDLDOJI'] = ta.CDLDOJI(o, h, l, c)
features['CDLDOJISTAR'] = ta.CDLDOJISTAR(o, h, l, c)
features['CDLDRAGONFLYDOJI'] = ta.CDLDRAGONFLYDOJI(o, h, l, c)
features['CDLENGULFING'] = ta.CDLENGULFING(o, h, l, c)
features['CDLEVENINGDOJISTAR'] = ta.CDLEVENINGDOJISTAR(o,
h,
l,
c,
penetration=0)
features['CDLEVENINGSTAR'] = ta.CDLEVENINGSTAR(o, h, l, c, penetration=0)
features['CDLGAPSIDESIDEWHITE'] = ta.CDLGAPSIDESIDEWHITE(o, h, l, c)
features['CDLGRAVESTONEDOJI'] = ta.CDLGRAVESTONEDOJI(o, h, l, c)
features['CDLHAMMER'] = ta.CDLHAMMER(o, h, l, c)
features['CDLHANGINGMAN'] = ta.CDLHANGINGMAN(o, h, l, c)
features['CDLHARAMI'] = ta.CDLHARAMI(o, h, l, c)
features['CDLHARAMICROSS'] = ta.CDLHARAMICROSS(o, h, l, c)
features['CDLHIGHWAVE'] = ta.CDLHIGHWAVE(o, h, l, c)
features['CDLHIKKAKE'] = ta.CDLHIKKAKE(o, h, l, c)
features['CDLHIKKAKEMOD'] = ta.CDLHIKKAKEMOD(o, h, l, c)
features['CDLHOMINGPIGEON'] = ta.CDLHOMINGPIGEON(o, h, l, c)
features['CDLIDENTICAL3CROWS'] = ta.CDLIDENTICAL3CROWS(o, h, l, c)
features['CDLINNECK'] = ta.CDLINNECK(o, h, l, c)
features['CDLINVERTEDHAMMER'] = ta.CDLINVERTEDHAMMER(o, h, l, c)
features['CDLKICKING'] = ta.CDLKICKING(o, h, l, c)
features['CDLKICKINGBYLENGTH'] = ta.CDLKICKINGBYLENGTH(o, h, l, c)
features['CDLLADDERBOTTOM'] = ta.CDLLADDERBOTTOM(o, h, l, c)
features['CDLLONGLEGGEDDOJI'] = ta.CDLLONGLEGGEDDOJI(o, h, l, c)
features['CDLMARUBOZU'] = ta.CDLMARUBOZU(o, h, l, c)
features['CDLMATCHINGLOW'] = ta.CDLMATCHINGLOW(o, h, l, c)
features['CDLMATHOLD'] = ta.CDLMATHOLD(o, h, l, c, penetration=0)
features['CDLMORNINGDOJISTAR'] = ta.CDLMORNINGDOJISTAR(o,
h,
l,
c,
penetration=0)
features['CDLMORNINGSTAR'] = ta.CDLMORNINGSTAR(o, h, l, c, penetration=0)
features['CDLONNECK'] = ta.CDLONNECK(o, h, l, c)
features['CDLPIERCING'] = ta.CDLPIERCING(o, h, l, c)
features['CDLRICKSHAWMAN'] = ta.CDLRICKSHAWMAN(o, h, l, c)
features['CDLRISEFALL3METHODS'] = ta.CDLRISEFALL3METHODS(o, h, l, c)
features['CDLSEPARATINGLINES'] = ta.CDLSEPARATINGLINES(o, h, l, c)
features['CDLSHOOTINGSTAR'] = ta.CDLSHOOTINGSTAR(o, h, l, c)
features['CDLSHORTLINE'] = ta.CDLSHORTLINE(o, h, l, c)
features['CDLSPINNINGTOP'] = ta.CDLSPINNINGTOP(o, h, l, c)
features['CDLSTALLEDPATTERN'] = ta.CDLSTALLEDPATTERN(o, h, l, c)
features['CDLSTICKSANDWICH'] = ta.CDLSTICKSANDWICH(o, h, l, c)
features['CDLTAKURI'] = ta.CDLTAKURI(o, h, l, c)
features['CDLTASUKIGAP'] = ta.CDLTASUKIGAP(o, h, l, c)
features['CDLTHRUSTING'] = ta.CDLTHRUSTING(o, h, l, c)
features['CDLTRISTAR'] = ta.CDLTRISTAR(o, h, l, c)
features['CDLUNIQUE3RIVER'] = ta.CDLUNIQUE3RIVER(o, h, l, c)
features['CDLUPSIDEGAP2CROWS'] = ta.CDLUPSIDEGAP2CROWS(o, h, l, c)
features['CDLXSIDEGAP3METHODS'] = ta.CDLXSIDEGAP3METHODS(o, h, l, c)
window = 5
features_ext = features
for w in range(window):
tmp = features.shift(periods=60 * (w + 1), freq='S')
tmp.columns = [c + '_' + str(w + 1) + 'w' for c in features.columns]
features_ext = pd.concat([features_ext, tmp], axis=1)
return features_ext
def trades():
print(datetime.datetime.now())
cancel_order()
#获取持仓
holding_result = futureAPI.get_specific_position(future_id)
# print(holding_result)
#获取多仓可平仓数量，空仓可平仓数量
long_amount = holding_result['holding'][0]['long_avail_qty']
short_amount = holding_result['holding'][0]['short_avail_qty']
#获取多仓数量、空仓数量
long_qty = float(holding_result['holding'][0]['long_qty'])
short_qty = float(holding_result['holding'][0]['short_qty'])
#获取多仓收益率、空仓收益率
long_ratio = float(holding_result['holding'][0]['long_pnl_ratio'])
short_ratio = float(holding_result['holding'][0]['short_pnl_ratio'])
#获取多仓收益、空仓收益
long_pnl = float(holding_result['holding'][0]['long_pnl'])
short_pnl = float(holding_result['holding'][0]['short_pnl'])
print('多仓数量：', long_qty, '可平：', long_amount, '空仓数量：', short_qty, '可平：',
short_amount)
print('多仓收益：', long_pnl, '多仓收益率：', long_ratio)
print('空仓收益：', short_pnl, '空仓收益率：', short_ratio)
k_result = futureAPI.get_kline('BTC-USDT-200515', '')
df = pd.DataFrame(
k_result,
columns=['time', 'open', 'high', 'low', 'close', 'volume', 'vc'])
upper_band, middle_band, lower_band = talib.BBANDS(
df['close'].apply(float).values, timeperiod=20, nbdevup=2, nbdevdn=2)
pre_close_price = float(df['close'].values[-1])
print(pre_close_price, upper_band[-1], lower_band[-1])
#止盈止损部分
#多仓平多止盈
if float(long_amount) > 0 and long_ratio > 0.06:
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
sell_price = price_depth['asks'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='3',
order_type='0',
price=sell_price,
size=long_amount,
match_price='0')
print('多仓平多止盈')
#空仓平空止盈
if float(short_amount) > 0 and short_ratio > 0.06:
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
buy_price = price_depth['bids'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='4',
order_type='0',
price=buy_price,
size=short_amount,
match_price='0')
print('空仓平空止盈')
#多仓平多止损
if float(long_amount) > 0 and long_ratio < -0.03:
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
sell_price = price_depth['asks'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='3',
order_type='0',
price=sell_price,
size=long_amount,
match_price='0')
print('多仓平多止损')
#空仓平空止损
if float(short_amount) > 0 and short_ratio < -0.03:
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
buy_price = price_depth['bids'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='4',
order_type='0',
price=buy_price,
size=short_amount,
match_price='0')
print('空仓平空止损')
#交易部分
#交易数量
amount = 15
#获取仓位信息
holding_result = futureAPI.get_specific_position(future_id)
#获取多仓可平仓数量，空仓可平仓数量
long_amount = holding_result['holding'][0]['long_avail_qty']
short_amount = holding_result['holding'][0]['short_avail_qty']
if pre_close_price > upper_band[-1]:
print('超出上轨')
if float(long_amount) > 0 and float(short_amount) > 0:
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
buy_price = price_depth['bids'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='4',
order_type='0',
price=buy_price,
size=short_amount,
match_price='0')
print('有空有多，平空持多，持有多仓：', long_amount)
if float(short_amount) > 0 and float(long_amount) == 0:
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
#平空
buy_price = price_depth['bids'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='4',
order_type='0',
price=buy_price,
size=short_amount,
match_price='0')
#开多
order_result = futureAPI.take_order(instrument_id=future_id,
type='1',
order_type='0',
price=buy_price,
size=str(amount),
match_price='0')
print('有空无多，平空开多')
if float(short_amount) == 0 and float(long_amount) > 0:
print('无空有多，持有多仓：', long_amount)
if float(short_amount) == 0 and float(long_amount) == 0:
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
buy_price = price_depth['bids'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='1',
order_type='0',
price=buy_price,
size=str(amount),
match_price='0')
print('无空无多，买入开多')
if pre_close_price < lower_band[-1]:
print('低于下轨')
if float(long_amount) > 0 and float(short_amount) > 0:
#平多
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
sell_price = price_depth['asks'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='3',
order_type='0',
price=sell_price,
size=long_amount,
match_price='0')
print('有多有空，平多持空，持有空仓：', short_amount)
if float(long_amount) > 0 and float(short_amount) == 0:
#平多
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
sell_price = price_depth['asks'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='3',
order_type='0',
price=sell_price,
size=long_amount,
match_price='0')
#开空
order_result = futureAPI.take_order(instrument_id=future_id,
type='2',
order_type='0',
price=sell_price,
size=str(amount),
match_price='0')
print('有多无空，平多开空')
if float(long_amount) == 0 and float(short_amount) > 0:
print('无多有空，持有空仓：', short_amount)
if float(long_amount) == 0 and float(short_amount) == 0:
price_depth = futureAPI.get_depth('BTC-USDT-200515', '2', '')
sell_price = price_depth['asks'][1][0]
order_result = futureAPI.take_order(instrument_id=future_id,
type='2',
order_type='0',
price=sell_price,
size=str(amount),
match_price='0')
print('无多无空，卖出开空')
else:
print('区间内运行')
time.sleep(60)
def procBar(bar1, bar2, pos, trade):
global pairSeries
global tsPairratio
global tsZscore
global tsDates
global indSmaZscore
global intDatapoints
global sentEntryOrder
global sentExitOrder
global intSMALength
global dblUpperThreshold
global dblLowerThreshold
global instPair1Factor
global instPair2Factor
global dblQty
global dblQty2
global crossAbove
global crossBelow
if bar1['Close'] > 0 and bar2['Close'] > 0:
xd = bar1['Close'] * instPair1Factor
yd = bar2['Close'] * instPair2Factor
sym1 = bar1['Symbol']
sym2 = bar2['Symbol']
logging.info("s106:procBar:" + sym1 + "," + sym2)
if not pairSeries.has_key(sym1):
pairSeries[sym1] = list()
if not pairSeries.has_key(sym2):
pairSeries[sym2] = list()
if not tsPairratio.has_key(sym1 + sym2):
tsPairratio[sym1 + sym2] = list()
if not tsPairratio.has_key(sym2 + sym1):
tsPairratio[sym2 + sym1] = list()
if not tsZscore.has_key(sym1 + sym2):
tsZscore[sym1 + sym2] = list()
if not tsZscore.has_key(sym2 + sym1):
tsZscore[sym2 + sym1] = list()
if not tsDates.has_key(sym1):
tsDates[sym1] = list()
if not tsDates.has_key(sym2):
tsDates[sym2] = list()
if not tsDates.has_key(sym1 + sym2):
tsDates[sym1 + sym2] = list()
if not tsDates.has_key(sym2 + sym1):
tsDates[sym2 + sym1] = list()
if not crossAbove.has_key(sym1 + sym2):
crossAbove[sym1 + sym2] = False
if not crossBelow.has_key(sym1 + sym2):
crossBelow[sym1 + sym2] = False
if not sentEntryOrder.has_key(sym1 + sym2):
sentEntryOrder[sym1 + sym2] = False
if not sentExitOrder.has_key(sym1 + sym2):
sentExitOrder[sym1 + sym2] = False
if bar1['Date'] not in tsDates[sym1]:
pairSeries[sym1].append(bar1['Close'])
tsDates[sym1].append(bar1['Date'])
if bar2['Date'] not in tsDates[sym2]:
pairSeries[sym2].append(bar2['Close'])
tsDates[sym2].append(bar2['Date'])
dblRatioData = bar1['Close'] / bar2['Close']
tsPairratio[sym1 + sym2].append(dblRatioData)
dblRatioData2 = bar2['Close'] / bar1['Close']
tsPairratio[sym2 + sym1].append(dblRatioData2)
if len(tsPairratio[sym1 + sym2]) < intDatapoints or len(
tsPairratio[sym2 + sym1]) < intDatapoints:
return []
iStart = len(tsPairratio[sym1 + sym2]) - intDatapoints
iEnd = len(tsPairratio[sym1 + sym2]) - 1
dblAverage = np.mean(tsPairratio[sym1 + sym2][iStart:iEnd])
dblRatioStdDev = np.std(tsPairratio[sym1 + sym2][iStart:iEnd])
dblResidualsData = (dblRatioData - dblAverage)
dblZscoreData = (dblRatioData - dblAverage) / dblRatioStdDev
tsZscore[sym1 + sym2].append(dblZscoreData)
tsDates[sym1 + sym2].append(bar1['Date'])
iStart2 = len(tsPairratio[sym2 + sym1]) - intDatapoints
iEnd2 = len(tsPairratio[sym2 + sym1]) - 1
dblAverage2 = np.mean(tsPairratio[sym2 + sym1][iStart2:iEnd2])
dblRatioStdDev2 = np.std(tsPairratio[sym2 + sym1][iStart2:iEnd2])
dblResidualsData2 = (dblRatioData2 - dblAverage2)
dblZscoreData2 = (dblRatioData2 - dblAverage2) / dblRatioStdDev2
tsZscore[sym2 + sym1].append(dblZscoreData2)
tsDates[sym2 + sym1].append(bar2['Date'])
signals = pd.DataFrame()
signals['Date'] = tsDates[sym1 + sym2]
signals['tsZscore'] = tsZscore[sym1 + sym2]
signals['tsZscore2'] = tsZscore[sym2 + sym1]
#signals['indSmaZscore']=pd.rolling_mean(signals['tsZscore'], intSMALength, min_periods=1)
#signals['indSmaZscore2']=pd.rolling_mean(signals['tsZscore2'], intSMALength, min_periods=1)
(signals['indBbu'], signals['indBbm'], signals['indBbl']) = ta.BBANDS(
np.array(signals['tsZscore']),
timeperiod=intSMALength,
# number of non-biased standard deviations from the mean
nbdevup=dblBBUOrder,
nbdevdn=dblBBLOrder,
# Moving average type: simple moving average here
matype=0)
(signals['indBbu2'], signals['indBbm2'],
signals['indBbl2']) = ta.BBANDS(
np.array(signals['tsZscore2']),
timeperiod=intSMALength,
# number of non-biased standard deviations from the mean
nbdevup=dblBBUOrder,
nbdevdn=dblBBLOrder,
# Moving average type: simple moving average here
matype=0)
signals['indSmaZscore'] = ta.SMA(np.array(signals['tsZscore']),
intSMALength)
signals['indSmaZscore2'] = ta.SMA(np.array(signals['tsZscore2']),
intSMALength)
signals = signals.set_index('Date')
# tsPricePair1.Add(bar1['Date']Time, bar1['Close']);
# tsPricePair2.Add(bar2['Date']Time, bar2['Close']);
# double dblBeta = tsPricePair1.GetCovariance( tsPricePair2 ) /tsPricePair1.GetVariance(Bars.Ago(intDatapoints).DateTime, Bars.Ago(1).DateTime);
#print 'ZScore:' + str(dblZscoreData) + ' Upper: ' + str(dblUpperThreshold) + + ' Lower: ' + str(dblLowerThreshold)
if len(signals['tsZscore']) < 5 or len(
signals['indSmaZscore']) < 5 or len(
signals['indSmaZscore2']) < 5:
return []
#updateCointData();
if trade:
#print strOrderComment
(crossBelow[sym1 + sym2],
crossAbove[sym1 + sym2]) = crossCheck(signals, sym1 + sym2,
'tsZscore', 'indSmaZscore')
(z1CBBbu, z1CABbu) = crossCheck(signals, 'bb' + sym1 + sym2,
'tsZscore', 'indBbu')
(z1CBBbl, z1CABbl) = crossCheck(signals, 'bb2' + sym1 + sym2,
'tsZscore2', 'indBbl')
#print ' crossAbove: ' + str(crossAbove) + ' crossBelow: ' + str(crossBelow)
#crossBelow = signals['tsZscore'].iloc[-1] >= signals['indSmaZscore'].iloc[-1] and crossBelow.any()
#crossAbove = signals['tsZscore'].iloc[-1] <= signals['indSmaZscore'].iloc[-1] and crossAbove.any()
#print ' crossAbove: ' + str(crossAbove) + ' crossBelow: ' + str(crossBelow)
if not sentEntryOrder[sym1 + sym2] and not pos.has_key(
bar1['Symbol']) and not pos.has_key(bar2['Symbol']):
#dblQty = Math.Ceiling(1/dblBeta);
#dblQty2 = Math.Ceiling(dblBeta);
# Create the set of short and long simple moving averages over the
# respective periods
# Create a 'signal' (invested or not invested) when the short moving average crosses the long
# moving average, but only for the period greater than the shortest moving average window
if dblZscoreData >= dblUpperThreshold and \
z1CABbu:
#crossAbove[sym1+sym2] == 1 and \
#Sell(instPair1, dblQty, strOrderComment);
#Buy(instPair2, dblQty2, strOrderComment2);
#if (myParam.hasOpt)
#{
# Sell(myParam.getSym1OptInst(PutCall.Put, bar1['Close'], this), myParam.sym1OptQty, strOrderComment);
# Sell(myParam.getSym2OptInst(PutCall.Call, bar2['Close'], this), myParam.sym2OptQty, strOrderComment2);
#}
sentEntryOrder[sym1 + sym2] = True
sentExitOrder[sym1 + sym2] = False
strOrderComment = '{"Entry": 1, "Exit": 0, "symPair": "' + sym1 + sym2 + '", "zScore": ' + str(
round(dblZscoreData, 2)) + ', "zSMA": ' + str(
round(signals.iloc[-1]['indSmaZscore'], 2)) + '}'
strOrderComment2 = '{"Entry": 1, "Exit": 0, "symPair": "' + sym1 + sym2 + '", "zScore": ' + str(
round(dblZscoreData2, 2)) + ', "zSMA": ' + str(
round(signals.iloc[-1]['indSmaZscore2'], 2)) + '}'
return ([bar1['Symbol'], -abs(dblQty), strOrderComment],
[bar2['Symbol'],
abs(dblQty2), strOrderComment2])
elif dblZscoreData <= -1 * dblUpperThreshold and \
z1CABbu:
#crossBelow[sym1+sym2] and \
#Buy(instPair1, dblQty, strOrderComment);
#Sell(instPair2, dblQty2, strOrderComment2);
#if (myParam.hasOpt)
#{
# Sell(myParam.getSym1OptInst(PutCall.Call, bar1['Close'], this), myParam.sym1OptQty, strOrderComment);
# Sell(myParam.getSym2OptInst(PutCall.Put, bar2['Close'], this), myParam.sym2OptQty, strOrderComment2);
#}
sentEntryOrder[sym1 + sym2] = True
sentExitOrder[sym1 + sym2] = False
strOrderComment = '{"Entry": 1, "Exit": 0, "symPair": "' + sym1 + sym2 + '", "zScore": ' + str(
round(dblZscoreData, 2)) + ', "zSMA": ' + str(
round(signals.iloc[-1]['indSmaZscore'], 2)) + '}'
strOrderComment2 = '{"Entry": 1, "Exit": 0, "symPair": "' + sym1 + sym2 + '", "zScore": ' + str(
round(dblZscoreData2, 2)) + ', "zSMA": ' + str(
round(signals.iloc[-1]['indSmaZscore2'], 2)) + '}'
return ([bar1['Symbol'],
abs(dblQty), strOrderComment],
[bar2['Symbol'], -abs(dblQty2), strOrderComment2])
elif not sentExitOrder[sym1 + sym2] and pos.has_key(
bar1['Symbol']) and pos.has_key(bar2['Symbol']):
if pos[bar1['Symbol']] < 0 and pos[bar2['Symbol']] > 0 and \
dblZscoreData <= dblLowerThreshold and crossAbove[sym1+sym2]==1 and z1CBBbl:
#Buy(instPair1, dblQty, strOrderComment);
#Sell(instPair2, dblQty2, strOrderComment2);
#if (myParam.hasOpt)
#{
# Buy(myParam.instPair1opt, myParam.sym1OptQty, strOrderComment);
# Buy(myParam.instPair2opt, myParam.sym2OptQty, strOrderComment);
#}
sentEntryOrder[sym1 + sym2] = False
sentExitOrder[sym1 + sym2] = True
strOrderComment = '{"Entry": 0, "Exit": 1, "symPair": "' + sym1 + sym2 + '", "zScore": ' + str(
round(dblZscoreData, 2)) + ', "zSMA": ' + str(
round(signals.iloc[-1]['indSmaZscore'], 2)) + '}'
strOrderComment2 = '{"Entry": 0, "Exit": 1, "symPair": "' + sym1 + sym2 + '", "zScore": ' + str(
round(dblZscoreData2, 2)) + ', "zSMA": ' + str(
round(signals.iloc[-1]['indSmaZscore2'], 2)) + '}'
return ([bar1['Symbol'],
abs(dblQty), strOrderComment],
[bar2['Symbol'], -abs(dblQty2), strOrderComment2])
elif pos[bar1['Symbol']] > 0 and pos[bar2['Symbol']] < 0 and \
dblZscoreData >= -1 * dblLowerThreshold and crossBelow[sym1+sym2] == 1 and z1CBBbl:
#Sell(instPair1, dblQty, strOrderComment);
#Buy(instPair2, dblQty2, strOrderComment2);
#if (myParam.hasOpt)
#{
# Buy(myParam.instPair1opt, myParam.sym1OptQty, strOrderComment);
# Buy(myParam.instPair2opt, myParam.sym2OptQty, strOrderComment);
#}
sentEntryOrder[sym1 + sym2] = False
sentExitOrder[sym1 + sym2] = True
strOrderComment = '{"Entry": 0, "Exit": 1, "symPair": "' + sym1 + sym2 + '", "zScore": ' + str(
round(dblZscoreData, 2)) + ', "zSMA": ' + str(
round(signals.iloc[-1]['indSmaZscore'], 2)) + '}'
strOrderComment2 = '{"Entry": 0, "Exit": 1, "symPair": "' + sym1 + sym2 + '", "zScore": ' + str(
round(dblZscoreData2, 2)) + ', "zSMA": ' + str(
round(signals.iloc[-1]['indSmaZscore2'], 2)) + '}'
return ([bar1['Symbol'], -abs(dblQty), strOrderComment],
[bar2['Symbol'],
abs(dblQty2), strOrderComment2])
def get_feature_from_vector(vector, dtype=None):
feature_dict = {}
feature_dict['ema_10'] = tl.EMA(vector, timeperiod=10)
feature_dict['ema_16'] = tl.EMA(vector, timeperiod=16)
feature_dict['ema_22'] = tl.EMA(vector, timeperiod=22)
feature_dict['sma_10'] = tl.SMA(vector, timeperiod=10)
feature_dict['sma_16'] = tl.SMA(vector, timeperiod=16)
feature_dict['sma_22'] = tl.SMA(vector, timeperiod=22)
if dtype == 'adj_close':
upper, middle, lower = tl.BBANDS(vector,
timeperiod=20,
nbdevup=2,
nbdevdn=2)
feature_dict['bbands_20_upper'] = upper
feature_dict['bbands_20_lower'] = lower
upper, middle, lower = tl.BBANDS(vector,
timeperiod=26,
nbdevup=2,
nbdevdn=2)
feature_dict['bbands_26_upper'] = upper
feature_dict['bbands_26_lower'] = lower
upper, middle, lower = tl.BBANDS(vector,
timeperiod=32,
nbdevup=2,
nbdevdn=2)
feature_dict['bbands_32_upper'] = upper
feature_dict['bbands_32_lower'] = lower
feature_dict['momentum_12'] = tl.MOM(vector, timeperiod=12)
feature_dict['momentum_18'] = tl.MOM(vector, timeperiod=18)
feature_dict['momentum_24'] = tl.MOM(vector, timeperiod=24)
ema = tl.EMA(feature_dict['momentum_12'], timeperiod=2)
feature_dict['momentum_to_ema_12'] = np.divide(
feature_dict['momentum_12'], ema)
feature_dict['ema_to_mom_12'] = ema
ema = tl.EMA(feature_dict['momentum_18'], timeperiod=2)
feature_dict['momentum_to_ema_18'] = np.divide(
feature_dict['momentum_18'], ema)
feature_dict['ema_to_mom_18'] = ema
ema = tl.EMA(feature_dict['momentum_24'], timeperiod=2)
feature_dict['momentum_to_ema_24'] = np.divide(
feature_dict['momentum_24'], ema)
feature_dict['ema_to_mom_24'] = ema
feature_dict['rate_of_change_10'] = tl.ROCP(vector, timeperiod=10)
feature_dict['rate_of_change_16'] = tl.ROCP(vector, timeperiod=16)
feature_dict['rate_of_change_22'] = tl.ROCP(vector, timeperiod=22)
macd = tl.MACD(vector, fastperiod=12, slowperiod=18)[0]
feature_dict['macd_0_18'] = macd
ema = tl.EMA(feature_dict['macd_0_18'], timeperiod=9)
feature_dict['macds_0_18'] = ema
feature_dict['macdsr_0_18'] = np.divide(feature_dict['macd_0_18'], ema)
macd = tl.MACD(vector, fastperiod=12, slowperiod=18)[1]
feature_dict['macd_1_18'] = macd
ema = tl.EMA(macd, timeperiod=9)
feature_dict['macds_1_18'] = ema
feature_dict['macdsr_1_18'] = np.divide(feature_dict['macd_1_18'], ema)
macd = tl.MACD(vector, fastperiod=12, slowperiod=18)[2]
feature_dict['macd_2_18'] = macd
ema = tl.EMA(macd, timeperiod=9)
feature_dict['macds_2_18'] = ema
feature_dict['macdsr_2_18'] = np.divide(feature_dict['macd_2_18'], ema)
macd = tl.MACD(vector, fastperiod=12, slowperiod=24)[0]
feature_dict['macd_0_24'] = macd
ema = tl.EMA(feature_dict['macd_0_24'], timeperiod=9)
feature_dict['macds_0_24'] = ema
feature_dict['macdsr_0_24'] = np.divide(feature_dict['macd_0_24'], ema)
macd = tl.MACD(vector, fastperiod=12, slowperiod=24)[1]
feature_dict['macd_1_24'] = macd
ema = tl.EMA(feature_dict['macd_1_24'], timeperiod=9)
feature_dict['macds_1_24'] = ema
feature_dict['macdsr_1_24'] = np.divide(feature_dict['macd_1_24'], ema)
macd = tl.MACD(vector, fastperiod=12, slowperiod=24)[2]
feature_dict['macd_2_24'] = macd
ema = tl.EMA(feature_dict['macd_2_24'], timeperiod=9)
feature_dict['macds_2_24'] = ema
feature_dict['macdsr_2_24'] = np.divide(feature_dict['macd_2_24'], ema)
macd = tl.MACD(vector, fastperiod=12, slowperiod=30)[0]
feature_dict['macd_0_30'] = macd
ema = tl.EMA(feature_dict['macd_0_30'], timeperiod=9)
feature_dict['macds_0_30'] = ema
feature_dict['macdsr_0_30'] = np.divide(feature_dict['macd_0_30'], ema)
macd = tl.MACD(vector, fastperiod=12, slowperiod=30)[1]
feature_dict['macd_1_30'] = macd
ema = tl.EMA(feature_dict['macd_1_30'], timeperiod=9)
feature_dict['macds_1_30'] = ema
feature_dict['macdsr_1_30'] = np.divide(feature_dict['macd_1_30'], ema)
macd = tl.MACD(vector, fastperiod=12, slowperiod=30)[2]
feature_dict['macd_2_30'] = macd
ema = tl.EMA(feature_dict['macd_2_30'], timeperiod=9)
feature_dict['macds_2_30'] = ema
feature_dict['macdsr_2_30'] = np.divide(feature_dict['macd_2_30'], ema)
feature_dict['rsi_8'] = tl.RSI(vector, timeperiod=8)
feature_dict['rsi_14'] = tl.RSI(vector, timeperiod=14)
feature_dict['rsi_20'] = tl.RSI(vector, timeperiod=20)
return feature_dict
def strategy_five(series, is_backtest=True):
'''策略：布林通道宽度、金叉死叉、顺势而为、KDJ零百转换、随机漫步'''
# 实时线
close = np.cumsum(series).astype(float)
# 布林通道宽度
upperband, middleband, lowerband = ta.BBANDS(close,
timeperiod=5,
nbdevup=2,
nbdevdn=2,
matype=0)
bandwidth = upperband - lowerband
mean = np.nan_to_num(bandwidth).mean()
# 金叉死叉
ma_fast = ta.MA(close, timeperiod=8, matype=0)
ma_slow = ta.MA(close, timeperiod=20, matype=0)
# 顺势而为
ma = ta.MA(close, timeperiod=10, matype=0)
ma = np.nan_to_num(ma) #Nan转换为0
# KDJ极限
fastk, fastd = ta.STOCHRSI(close,
timeperiod=14,
fastk_period=10,
fastd_period=5,
fastd_matype=0)
# 随机漫步
rnd = []
signals = []
current_signal = [0] * 5 # 当前信号：无
# 如果是回测
if is_backtest:
pre_is_right = True # 哨兵：上期投注正确
earn = [] # 收益记录
order = [] # 投注资金记录
for i in range(2, len(series)):
# 布林通道宽度
#if bandwidth[i-1] < bandwidth[i]: # 增：趋势区间
if bandwidth[i - 1] > mean: # 宽：趋势区间
current_signal[0] = series[i - 1] # 与前一期相同 //与前一期（趋势）相同
else: # 窄：震荡区间
current_signal[0] = -series[i - 1] # 与前一期相反
# 金叉死叉
y1, y2 = ma_fast[i - 2], ma_fast[i - 1]
y3, y4 = ma_slow[i - 2], ma_slow[i - 1]
if y1 < y3 and y2 > y4: # 金叉
current_signal[1] = 1
elif y1 > y3 and y2 < y4: # 死叉
current_signal[1] = -1
# 顺势而为
if ma[i - 2] < ma[i - 1]: # 上升趋势
current_signal[2] = 1
else: # 下降趋势
current_signal[2] = -1
# KDJ极限
if fastk[i - 1] == 0: # 上升趋势
current_signal[3] = 1
elif fastk[i] == 100: # 下降趋势
current_signal[3] = -1
# 随机漫步
rnd.append(np.random.choice([-1, 1]))
current_signal[4] = rnd[-1]
# 若无信号，跳过，不做任何操作
if sum(current_signal) == 0: continue
# 看料
buy = [-1, 1][sum(current_signal) > 0]
if i < len(series) - 1: # 若不是最后一期，则
signals.append(buy)
# 下单
order.append(times[error_periods])
# 判断对错，计算收益
if buy == series[i]:
error_periods = 0 # 错误期数置零
earn.append(odds * order[-1]) # 赔率odds=1.96
else:
error_periods += 1 # 错误期数加一
earn.append(0)
else:
print(buy)
print('投注记录:\n', order[-show_period:], '\n合计：',
sum(order[-show_period:]))
print('收益记录:\n', earn[-show_period:], '\n合计：',
sum(earn[-show_period:]))
print('收益率：', (sum(earn[-show_period:]) - sum(order[-show_period:])) /
sum(order[-show_period:]))
return order, earn, signals
else:
# 布林通道宽度
if bandwidth[-1] > mean: # 宽：趋势区间
current_signal[0] = series[-1] # 与前一期相同 //与前一期（趋势）相同
else: # 窄：震荡区间
current_signal[0] = -series[-1] # 与前一期相反
# 金叉死叉
y1, y2 = ma_fast[-2], ma_fast[-1]
y3, y4 = ma_slow[-2], ma_slow[-1]
if y1 < y3 and y2 > y4: # 金叉
current_signal[1] = 1
elif y1 > y3 and y2 < y4: # 死叉
current_signal[1] = -1
# 顺势而为
if ma[-2] < ma[-1]: # 上升趋势
current_signal[2] = 1
else: # 下降趋势
current_signal[2] = -1
# KDJ极限
if fastk[-1] == 0: # 上升趋势
current_signal[3] = 1
elif fastk[-1] == 100: # 下降趋势
current_signal[3] = -1
# 随机漫步
current_signal[4] = np.random.choice([-1, 1])
print('=' * 30)
print('五种综合： ', end='')
print_red_color_text([-1, 1][sum(current_signal) > 0])
print('=' * 30)
return [-1, 1][sum(current_signal) > 0]
df.loc[idx, 'total_money'] = record_info['money_remain'] + record_info[
'amount_remain'] * df.loc[idx, 'close']
return df.tail(1)['total_money'].values[0]
if __name__ == '__main__':
""" ----------------------- 准备训练数据 ------------------------------ """
stk_code = '300508'
sh_index = ts.get_k_data(code=stk_code)
closed = sh_index['close'].values
sh_index['upper'], sh_index['middle'], sh_index['lower'] = talib.BBANDS(
closed,
timeperiod=10,
# number of non-biased standard deviations from the mean
nbdevup=2,
nbdevdn=2,
# Moving average type: simple moving average here
matype=0)
sh_index = sh_index.dropna(how='any').tail(100)
""" --------------------- 定义 ------------------------ """
r_bit_len = 5 # 网格二进制长度
r_amount = 3 # 网格数量
DNA_SIZE = r_bit_len * r_amount # DNA 长度
POP_SIZE = 100 # 种群大小
CROSS_RATE = 0.8 # mating probability (DNA crossover)
MUTATION_RATE = 0.03 # mutation probability
N_GENERATIONS = 200 # 种群代数
X_BOUND = [0, 20] # 基因的上下限（此处应该设置为“相对price”的上下限）
print("Current Time =", current_time)
RSI_OVERBOUGHT = 75
RSI_OVERSOLD = 30
cols = ['T', 'O', 'H', 'L', 'C', 'V']
data = pd.read_csv('data/live.csv')
df = pd.DataFrame(data)
df.drop(df.columns[6:], axis=1, inplace=True)
df.columns = cols
df['EMA30'] = EMA(df.C, 90)
df['EMA20'] = EMA(df.C, 10)
df['EMA10'] = EMA(df.C, 5)
df['EMA200'] = EMA(df.C, 90)
df.loc[(df['EMA10'] >= df['EMA200']), 'emaUptrend'] = True
df.loc[(df['EMA10'] <= df['EMA200']), 'emaDowntrend'] = True
df['RSI'] = talib.RSI(df.C)
df['upper'], df['middle'], df['lower'] = talib.BBANDS(df.C, matype=MA_Type.T3)
df['BBsellSig'] = df.C > df.upper
df['BBbuySig'] = df.C < df.lower
df['BBsellSig'] = df.C > df.upper
df['rsiOversold'] = df.RSI < RSI_OVERSOLD
df['rsiOverbought'] = df.RSI > RSI_OVERBOUGHT
df['macd'], df["sig"], df['diff'] = MACDFIX(df.C)
df.loc[(df['diff'] >= 0), 'macdBuy'] = True
df.loc[(df['EMA20'] <= df['EMA30']) & (df['EMA10'] <= df['EMA20']) &
(df['C'] <= df['EMA10']), 'EMAsellSig'] = True
df.loc[(df['EMA20'] >= df['EMA30']) & (df['EMA10'] >= df['EMA20']) &
(df['C'] >= df['EMA10']), 'EMAbuySig'] = True
###BUY
df.loc[(df.emaUptrend & df.macdBuy), 'BUY'] = 1
##SELL
def attribute_gen(data):
'''
generate attributes from cleansed data of local database
:param data: is tick level np.array with 28 columns
:return: X, y, Xpred are dataframe with datetime index
'''
#######################################################################################
# indexed by datetime
data_indexed = pd.DataFrame(data)
data_indexed = data_indexed.set_index(
2) # the original 28 column is removed automatically
data_indexed.drop([0, 1], axis=1, inplace=True) # 8503
# type(data_indexed[7][0]) # it is int, we need float instead
data_indexed[7] = data_indexed[7].astype(float)
#######################################################################################
'''
# five minutes first & last & mean
# 1
data_indexed_5m_first = data_indexed.resample('5T').first()
data_indexed_5m_first = data_indexed_5m_first.between_time('9:30', '15:00')
inde = data_indexed_5m_first[data_indexed_5m_first[3].isna()].index
data_indexed_5m_first.drop(inde, inplace=True)
# data_indexed_5m_first = data_indexed_5m_first.fillna(method='backfill')
data_indexed_5m_first.drop(data_indexed_5m_first.tail(1).index, inplace=True)
data_indexed_5m_first.drop(data_indexed_5m_first.tail(1).index, inplace=True)
# 2
data_indexed_5m_last = data_indexed.resample('5T').last() # 354
data_indexed_5m_last = data_indexed_5m_last.between_time('9:30', '15:00') # 133
inde = data_indexed_5m_last[data_indexed_5m_last[3].isna()].index
data_indexed_5m_last.drop(inde, inplace=True) # 96
# data_indexed_5m_last = data_indexed_5m_last.fillna(method='backfill')
# cut the last item
data_indexed_5m_last.drop(data_indexed_5m_last.tail(1).index, inplace=True)
# 3
data_indexed_5m_last = data_indexed_5m_last.between_time('9:30', '15:00') # 133
inde = data_indexed_5m_last[data_indexed_5m_last[3].isna()].index
data_indexed_5m_last.drop(inde, inplace=True) # 96
# data_indexed_5m_last = data_indexed_5m_last.fillna(method='backfill')
# cut the last item
data_indexed_5m_last.drop(data_indexed_5m_last.tail(1).index, inplace=True)
# 4
data_indexed_5m_change = data_indexed_5m_last - data_indexed_5m_first
'''
########################################################################################
# five minutes current price last & change & max & min
data_cp = pd.DataFrame(data[:, [2, 3]]) # data is np.array()
data_cp = data_cp.set_index(0)
data_cplast = data_cp.resample('5T').last() # 354
inde = data_cplast[data_cplast[1].isna()].index
data_cplast.drop(inde, inplace=True) # 96
# data_cpfirst = data_cp.resample('5T').first() # 354
# inde = data_cpfirst[data_cpfirst[1].isna()].index
# data_cpfirst.drop(inde, inplace=True) # 96
data_cprice_change = data_cplast.diff()
data_cprice_change = data_cprice_change.fillna(method='bfill')
# data_cpratio = data_cplast/data_cpfirst
# move one step forward
# data_cprice_change = data_cprice_change[1][data_cprice_change.index[1:len(data_cprice_change)]]
# cut the last item
# data_cplast = data_cplast[1][data_cplast.index[0:len(data_cplast) - 1]]
data_cpmax = data_cp.resample('5T').max() # 354
indexname = data_cpmax[data_cpmax[1].isna()].index
data_cpmax.drop(indexname, inplace=True) # 96
# ?????? doesn't work: data_cpmax = data_cpmax.resample('5T').agg(lambda x: max(x[i]) for i in range(len(x))) # 354
# cut the last item
# data_cpmax = data_cpmax[1][data_cpmax.index[0:len(data_cpmax)-1]]
data_cpmin = data_cp.resample('5T').min() # 354
ind = data_cpmin[data_cpmin[1].isna()].index
data_cpmin.drop(ind, inplace=True) # 96
# cut the last item
# data_cpmin = data_cpmin[1][data_cpmin.index[0:len(data_cpmin)-1]]
data_cp = pd.DataFrame()
data_cp['max'] = data_cpmax[1]
data_cp['min'] = data_cpmin[1]
data_cp['last'] = data_cplast[1]
data_cp['change'] = data_cprice_change[1]
# data_cp['ratio'] = data_cpratio[1] ratio can not increase forecast accuracy
######################################################################################################
# flag data
data_flag = pd.DataFrame(data[:, [2, 7]])
data_flag = data_flag.set_index(0)
data_flag_plus = data_flag
data_flag_minus = data_flag
# data_flag_plus[1][data_flag_plus[1] == -1] = 0
# data_flag_minus[1][data_flag_minus[1] == 1] = 0
# data_flag_minus[1][data_flag_minus[1] == -1] = 1
# flagtry = list(map(lambda x: x+1 if x == -1 else x, data_flag_plus[1]))
data_flag_plus = data_flag_plus.applymap(lambda x: x + 1 if x == -1 else x)
data_flag_minus = data_flag_minus.applymap(lambda x: x - 1
if x == 1 else x + 2)
data_flag_plsum = data_flag_plus.resample('5T').sum() # 354
data_flag_misum = data_flag_minus.resample('5T').sum() # 354
# index = data_flag_plsum[data_flag_plsum[1] == 0].index
data_flag_plsum.drop(inde, inplace=True) # 96
data_flag_misum.drop(inde, inplace=True) # 96
data_flag_ratio = data_flag_plsum / data_flag_misum # 96
flag = pd.DataFrame()
flag['psum'] = data_flag_plsum[1]
# flag['msum'] = data_flag_misum[1]
flag['ratio'] = data_flag_ratio[1]
# cut the last item
# flag.drop(flag.tail(1).index, inplace=True)
####################################################################################
'''
data_flag = pd.Series(data_flag[1]) # the index is attached automatically
data_flag = data_flag.resample('5T').agg(lambda x: x[-1]/x[1] - 1)
index = data_indexed_5m_first[data_flag[3].isna()].index
data_flag.drop(index, inplace=True)
data_flag = data_flag.fillna(method='backfill')
'''
# standardize money (column five)
data_money = pd.DataFrame(data[:, [2, 5]]) # data is np.array()
data_money = data_money.set_index(0)
data_money = pd.Series(data_money[1])
data_money_5m = data_money.resample('5T').sum() # 354
ind = data_money_5m[data_money_5m == 0.0].index
data_money_5m.drop(ind, inplace=True) # 96
data_money_5m_d = data_money_5m.diff()
data_money_5m_d = data_money_5m_d.fillna(method='backfill')
# cut the last item
# data_money_5m_d.drop(data_money_5m_d.tail(1).index, inplace=True)
data_money = pd.DataFrame()
data_money['money'] = data_money_5m
data_money['diff'] = data_money_5m_d
#########################################################################################################
# MACD
macd = data_indexed[3] # 8503
macd = pd.DataFrame(macd)
macd_12_ema = macd.ewm(span=12, adjust=False).mean()
macd_26_ema = macd.ewm(span=26, adjust=False).mean()
macd_real = macd_12_ema - macd_26_ema
macd_reference = macd_real.ewm(span=9, adjust=False).mean()
# transform to 5 minutes' frequency
mmaa = pd.DataFrame()
mmaa['12'] = macd_12_ema[3]
mmaa['26'] = macd_26_ema[3]
mmaa['real'] = macd_real[3]
mmaa['refer'] = macd_reference[3]
mmaa_5m = mmaa.resample('5T')
mmaa_5m_mean = mmaa_5m.mean() # 354
indd = mmaa_5m_mean[mmaa_5m_mean['12'].isna()].index
mmaa_5m_mean.drop(indd, inplace=True) # 96
# cut the last item
# mmaa_5m_mean.drop(mmaa_5m_mean.tail(1).index, inplace=True) # 95
# macd, macdsignal, macdhist = talib.MACD(close, fastperiod=12,
# slowperiod=26, signalperiod=9)
####################################################################################
high = data_cpmax[1]
low = data_cpmin[1]
close = data_cplast[1]
data_vol = pd.DataFrame(data[:, [2, 6]]) # data is np.array()
data_vol = data_vol.set_index(0)
data_vol = pd.Series(data_vol[1])
data_vol_5m = data_vol.resample('5T').sum() # 354
ind = data_vol_5m[data_vol_5m == 0.0].index
data_vol_5m.drop(ind, inplace=True) # 96
volume = data_vol_5m
####################################################################################
# sma of current price
data_sma_20 = talib.SMA(np.asarray(close), 20)
data_sma_20 = pd.DataFrame(data_sma_20)
data_sma_20 = data_sma_20.fillna(method='backfill')
data_sma_10 = talib.SMA(np.asarray(close), 10)
data_sma_10 = pd.DataFrame(data_sma_10)
data_sma_10 = data_sma_10.fillna(method='backfill')
data_sma_5 = talib.SMA(np.asarray(close), 5)
data_sma_5 = pd.DataFrame(data_sma_5)
data_sma_5 = data_sma_5.fillna(method='backfill')
data_sma = pd.DataFrame()
data_sma['5'] = data_sma_5[0]
data_sma['10'] = data_sma_10[0]
data_sma['20'] = data_sma_20[0]
data_sma['0'] = volume.index
data_sma = data_sma.set_index('0')
####################################################################################
# Average Directional Movement Index
# real = talib.ADX(high, low, close, timeperiod=14)
direct_move = talib.ADX(high, low, close, timeperiod=14)
direct_move = direct_move.fillna(method='backfill')
########################################################################################
# Absolute Price Oscillator
oscillator = talib.APO(close, fastperiod=12, slowperiod=26, matype=0)
oscillator = oscillator.fillna(method='backfill')
########################################################################################
# macd fix
macdfix, macdsignal, macdhist = talib.MACDFIX(close,
signalperiod=9) # nan 40
macdfix.fillna(method='backfill')
#########################################################################################
# money flow
money_flow = talib.MFI(high, low, close, volume, timeperiod=14)
money_flow.fillna(method='backfill')
#########################################################################################
# minus directional indicator
minus_direct = talib.MINUS_DI(high, low, close, timeperiod=14)
minus_direct.fillna(method='backfill')
#########################################################################################
# momentum of the close prices, with a time period of 5:
mom = talib.MOM(close, timeperiod=10)
mom = mom.fillna(method='backfill')
#########################################################################################
# PLUS_DI - Plus Directional Indicator
plus_direct = talib.PLUS_DI(high, low, close, timeperiod=14)
plus_direct = plus_direct.fillna(method='backfill')
#########################################################################################
# PLUS_DM - Plus Directional Movement
plus_direct_move = talib.PLUS_DM(high, low, timeperiod=14)
plus_direct_move = plus_direct_move.fillna(method='backfill')
#########################################################################################
# PPO - Percentage Price Oscillator
percent_oscillator = talib.PPO(close,
fastperiod=12,
slowperiod=26,
matype=0)
percent_oscillator = percent_oscillator.fillna(method='backfill')
#########################################################################################
# ROCP - Rate of change Percentage: (price-prevPrice)/prevPrice
rate_of_change = talib.ROCP(close, timeperiod=10)
rate_of_change = rate_of_change.fillna(method='backfill')
#########################################################################################
# Rate of change ratio
rate_ratio = talib.ROCR(close, timeperiod=10)
rate_ratio = rate_ratio.fillna(method='backfill')
#########################################################################################
# RSI - Relative Strength Index
strength = talib.RSI(close, timeperiod=14)
strength = strength.fillna(method='backfill')
#########################################################################################
# STOCH - Stochastic
slowk, slowd = talib.STOCH(high,
low,
close,
fastk_period=5,
slowk_period=3,
slowk_matype=0,
slowd_period=3,
slowd_matype=0)
slowk = slowk.fillna(method='backfill')
slowd = slowd.fillna(method='backfill')
stochastic = pd.DataFrame()
stochastic['slowk'] = slowk
stochastic['slowd'] = slowd
#########################################################################################
# STOCHF - Stochastic Fast
fastk, fastd = talib.STOCHF(high,
low,
close,
fastk_period=5,
fastd_period=3,
fastd_matype=0)
fastk = fastk.fillna(method='backfill')
fastd = fastd.fillna(method='backfill')
fast = pd.DataFrame()
fast['fastk'] = fastk
fast['fastd'] = fastd
#########################################################################################
# STOCHRSI - Stochastic Relative Strength Index
fastk, fastd = talib.STOCHRSI(close,
timeperiod=14,
fastk_period=5,
fastd_period=3,
fastd_matype=0)
fastk = fastk.fillna(method='backfill')
fastd = fastd.fillna(method='backfill')
relative_strength = pd.DataFrame()
relative_strength['fastk'] = fastk
relative_strength['fastd'] = fastd
#########################################################################################
# ULTOSC - Ultimate Oscillator
ulti_osci = talib.ULTOSC(high,
low,
close,
timeperiod1=7,
timeperiod2=14,
timeperiod3=28)
ulti_osci = ulti_osci.fillna(method='backfill')
#########################################################################################
# WILLR - Williams' %R
willer = talib.WILLR(high, low, close, timeperiod=14)
willer = willer.fillna(method='backfill')
#########################################################################################
# buy sell info factors
buy_sell = data[:, 8:28]
buy_sell = pd.DataFrame(buy_sell, index=data[:, 2])
# col 8-12 buy price, col 13-17 sell price,
# col 18-22 buy quant, col 23-27 sell quant
buy_sell_5m = buy_sell.resample('5T').last() # 354
inde = buy_sell_5m[buy_sell_5m[1].isna()].index
buy_sell_5m.drop(inde, inplace=True) # 96
#########################################################################################
# bollinger bands, with triple exponential moving average:
from talib import MA_Type
upper, middle, lower = talib.BBANDS(close, matype=MA_Type.T3)
#########################################################################################
# generate label y
data_cp2 = pd.DataFrame(data[:, [2, 3]]) # data is np.array()
data_cp2 = data_cp2.set_index(0)
data_cplast = data_cp2.resample('5T').last() # 354
inddd = data_cplast[data_cplast[1].isna()].index
data_cplast.drop(inddd, inplace=True) # 96
pchange = data_cplast.diff() # 96
pchange = pchange.fillna(method='backfill')
# cut the last item
# data_cplast = data_cplast[1][data_cplast.index[0:len(data_cplast) - 1]]
func = lambda x: np.sign(x) if x != 0 else np.sign(x + 1)
pchange['sign'] = list(map(func, pchange[1]))
# pchange.drop(pchange.head(1).index, inplace=True) # 95
####################################################################################
# generate train and label data
X = pd.concat([
mmaa_5m_mean, data_sma, direct_move, oscillator, macdfix, money_flow,
minus_direct, mom, plus_direct, plus_direct_move, percent_oscillator,
rate_of_change, rate_ratio, strength, stochastic, fast,
relative_strength, ulti_osci, willer, data_money, flag, data_cp
],
axis=1)
y = pchange['sign']
price_change = data_cprice_change
return X, y, price_change
def extract_by_type(feature_type, open_prices=None, close_prices=None, high_prices=None, low_prices=None):
if feature_type == 'ROCP':
rocp1 = talib.ROCP(close_prices, timeperiod=1)
rocp2 = talib.ROCP(close_prices, timeperiod=2)
feature.append(rocp1)
feature.append(rocp2)
if feature_type == 'OROCP':
orocp = talib.ROCP(open_prices, timeperiod=1)
feature.append(orocp)
if feature_type == 'HROCP':
hrocp = talib.ROCP(high_prices, timeperiod=1)
feature.append(hrocp)
if feature_type == 'LROCP':
lrocp = talib.ROCP(low_prices, timeperiod=1)
feature.append(lrocp)
if feature_type == 'MACD':
macd, signal, hist = talib.MACD(close_prices, fastperiod=12, slowperiod=26, signalperiod=9)
norm_macd = numpy.nan_to_num(macd) / math.sqrt(numpy.var(numpy.nan_to_num(macd)))
norm_signal = numpy.nan_to_num(signal) / math.sqrt(numpy.var(numpy.nan_to_num(signal)))
norm_hist = numpy.nan_to_num(hist) / math.sqrt(numpy.var(numpy.nan_to_num(hist)))
macdrocp = talib.ROCP(norm_macd + numpy.max(norm_macd) - numpy.min(norm_macd), timeperiod=1)
signalrocp = talib.ROCP(norm_signal + numpy.max(norm_signal) - numpy.min(norm_signal), timeperiod=1)
histrocp = talib.ROCP(norm_hist + numpy.max(norm_hist) - numpy.min(norm_hist), timeperiod=1)
# feature.append(macd / 100.0)
# feature.append(signal / 100.0)
# feature.append(hist / 100.0)
feature.append(norm_macd)
feature.append(norm_signal)
feature.append(norm_hist)
feature.append(macdrocp)
feature.append(signalrocp)
feature.append(histrocp)
if feature_type == 'RSI':
rsi6 = talib.RSI(close_prices, timeperiod=6)
rsi12 = talib.RSI(close_prices, timeperiod=12)
rsi24 = talib.RSI(close_prices, timeperiod=24)
rsi6rocp = talib.ROCP(rsi6 + 100., timeperiod=1)
rsi12rocp = talib.ROCP(rsi12 + 100., timeperiod=1)
rsi24rocp = talib.ROCP(rsi24 + 100., timeperiod=1)
feature.append(rsi6 / 100.0 - 0.5)
feature.append(rsi12 / 100.0 - 0.5)
feature.append(rsi24 / 100.0 - 0.5)
# feature.append(numpy.maximum(rsi6 / 100.0 - 0.8, 0))
# feature.append(numpy.maximum(rsi12 / 100.0 - 0.8, 0))
# feature.append(numpy.maximum(rsi24 / 100.0 - 0.8, 0))
# feature.append(numpy.minimum(rsi6 / 100.0 - 0.2, 0))
# feature.append(numpy.minimum(rsi6 / 100.0 - 0.2, 0))
# feature.append(numpy.minimum(rsi6 / 100.0 - 0.2, 0))
# feature.append(numpy.maximum(numpy.minimum(rsi6 / 100.0 - 0.5, 0.3), -0.3))
# feature.append(numpy.maximum(numpy.minimum(rsi6 / 100.0 - 0.5, 0.3), -0.3))
# feature.append(numpy.maximum(numpy.minimum(rsi6 / 100.0 - 0.5, 0.3), -0.3))
feature.append(rsi6rocp)
feature.append(rsi12rocp)
feature.append(rsi24rocp)
if feature_type == 'UO':
ult_osc = talib.ULTOSC(high_prices, low_prices, close_prices, timeperiod1=7, timeperiod2=14, timeperiod3=28)
feature.append(ult_osc / 100.0 - 0.5)
if feature_type == 'BOLL':
upperband, middleband, lowerband = talib.BBANDS(close_prices, timeperiod=5, nbdevup=2, nbdevdn=2, matype=0)
feature.append((upperband - close_prices) / close_prices)
feature.append((middleband - close_prices) / close_prices)
feature.append((lowerband - close_prices) / close_prices)
if feature_type == 'MA':
ma5 = talib.MA(close_prices, timeperiod=5)
ma10 = talib.MA(close_prices, timeperiod=10)
ma20 = talib.MA(close_prices, timeperiod=20)
ma25 = talib.MA(close_prices, timeperiod=25)
ma30 = talib.MA(close_prices, timeperiod=30)
ma40 = talib.MA(close_prices, timeperiod=40)
ma50 = talib.MA(close_prices, timeperiod=50)
ma60 = talib.MA(close_prices, timeperiod=60)
#ma360 = talib.MA(close_prices, timeperiod=70)
#ma720 = talib.MA(close_prices, timeperiod=720)
ma5rocp = talib.ROCP(ma5, timeperiod=1)
ma10rocp = talib.ROCP(ma10, timeperiod=1)
ma20rocp = talib.ROCP(ma20, timeperiod=1)
ma25rocp = talib.ROCP(ma25, timeperiod=1)
ma30rocp = talib.ROCP(ma30, timeperiod=1)
ma40rocp = talib.ROCP(ma40, timeperiod=1)
ma50rocp = talib.ROCP(ma50, timeperiod=1)
ma60rocp = talib.ROCP(ma60, timeperiod=1)
#ma360rocp = talib.ROCP(ma360, timeperiod=1)
#ma720rocp = talib.ROCP(ma720, timeperiod=1)
feature.append(ma5rocp)
feature.append(ma10rocp)
feature.append(ma20rocp)
feature.append(ma25rocp)
feature.append(ma30rocp)
feature.append(ma40rocp)
feature.append(ma50rocp)
feature.append(ma60rocp)
#feature.append(ma360rocp)
#feature.append(ma720rocp)
feature.append((ma5 - close_prices) / close_prices)
feature.append((ma10 - close_prices) / close_prices)
feature.append((ma20 - close_prices) / close_prices)
feature.append((ma25 - close_prices) / close_prices)
feature.append((ma30 - close_prices) / close_prices)
feature.append((ma40 - close_prices) / close_prices)
feature.append((ma50 - close_prices) / close_prices)
feature.append((ma60 - close_prices) / close_prices)
#feature.append((ma360 - close_prices) / close_prices)
#feature.append((ma720 - close_prices) / close_prices)
if feature_type == 'STOCH':
slow_stoch_k, slow_stoch_d = talib.STOCH(high_prices ,low_prices ,close_prices ,fastk_period=5, slowk_period=3, slowk_matype=0, slowd_period=3, slowd_matype=0)
fast_stoch_k, fast_stoch_d = talib.STOCHF(high_prices , low_prices , close_prices , fastk_period=5, fastd_period=3, fastd_matype=0)
fast_rsi_k, fast_rsi_d = talib.STOCHRSI(close_prices, timeperiod=14, fastk_period=5, fastd_period=3, fastd_matype=0)
feature.append(slow_stoch_k / 100.0 - 0.5)
feature.append(slow_stoch_d / 100.0 - 0.5)
feature.append(fast_stoch_k / 100.0 - 0.5)
feature.append(fast_stoch_d / 100.0 - 0.5)
feature.append(fast_rsi_k / 100.0 - 0.5)
feature.append(fast_rsi_d / 100.0 - 0.5)
if feature_type == 'AO':
median_price = (high_prices + low_prices) / 2
ao = talib.SMA(median_price, 5)-talib.SMA(median_price, 34)
feature.append(ao)
if feature_type == 'ROC':
roc5 = talib.ROC(close_prices, timeperiod=5)
roc10 = talib.ROC(close_prices, timeperiod=10)
roc20 = talib.ROC(close_prices, timeperiod=20)
roc25 = talib.ROC(close_prices, timeperiod=25)
feature.append(roc5)
feature.append(roc10)
feature.append(roc20)
feature.append(roc25)
if feature_type == 'WILLR':
willr = talib.WILLR(high_prices,low_prices,close_prices, timeperiod=14)
feature.append(willr / 100.0 - 0.5)
return feature
def momentum_process(event):
print(event.widget.get())
momentum = event.widget.get()
upperband, middleband, lowerband = ta.BBANDS(close, timeperiod=5, nbdevup=2, nbdevdn=2, matype=0)
fig, axes = plt.subplots(2, 1, sharex=True)
ax1, ax2 = axes[0], axes[1]
axes[0].plot(close, 'rd-', markersize=3)
axes[0].plot(upperband, 'y-')
axes[0].plot(middleband, 'b-')
axes[0].plot(lowerband, 'y-')
axes[0].set_title(momentum, fontproperties="SimHei")
if momentum == '绝对价格振荡器':
real = ta.APO(close, fastperiod=12, slowperiod=26, matype=0)
axes[1].plot(real, 'r-')
elif momentum == '钱德动量摆动指标':
real = ta.CMO(close, timeperiod=14)
axes[1].plot(real, 'r-')
elif momentum == '移动平均收敛/散度':
macd, macdsignal, macdhist = ta.MACD(close, fastperiod=12, slowperiod=26, signalperiod=9)
axes[1].plot(macd, 'r-')
axes[1].plot(macdsignal, 'g-')
axes[1].plot(macdhist, 'b-')
elif momentum == '带可控MA类型的MACD':
macd, macdsignal, macdhist = ta.MACDEXT(close, fastperiod=12, fastmatype=0, slowperiod=26, slowmatype=0, signalperiod=9, signalmatype=0)
axes[1].plot(macd, 'r-')
axes[1].plot(macdsignal, 'g-')
axes[1].plot(macdhist, 'b-')
elif momentum == '移动平均收敛/散度 固定 12/26':
macd, macdsignal, macdhist = ta.MACDFIX(close, signalperiod=9)
axes[1].plot(macd, 'r-')
axes[1].plot(macdsignal, 'g-')
axes[1].plot(macdhist, 'b-')
elif momentum == '动量':
real = ta.MOM(close, timeperiod=10)
axes[1].plot(real, 'r-')
elif momentum == '比例价格振荡器':
real = ta.PPO(close, fastperiod=12, slowperiod=26, matype=0)
axes[1].plot(real, 'r-')
elif momentum == '变化率':
real = ta.ROC(close, timeperiod=10)
axes[1].plot(real, 'r-')
elif momentum == '变化率百分比':
real = ta.ROCP(close, timeperiod=10)
axes[1].plot(real, 'r-')
elif momentum == '变化率的比率':
real = ta.ROCR(close, timeperiod=10)
axes[1].plot(real, 'r-')
elif momentum == '变化率的比率100倍':
real = ta.ROCR100(close, timeperiod=10)
axes[1].plot(real, 'r-')
elif momentum == '相对强弱指数':
real = ta.RSI(close, timeperiod=14)
axes[1].plot(real, 'r-')
elif momentum == '随机相对强弱指标':
fastk, fastd = ta.STOCHRSI(close, timeperiod=14, fastk_period=5, fastd_period=3, fastd_matype=0)
axes[1].plot(fastk, 'r-')
axes[1].plot(fastd, 'r-')
elif momentum == '三重光滑EMA的日变化率':
real = ta.TRIX(close, timeperiod=30)
axes[1].plot(real, 'r-')
plt.show()
def TA(S, s, c):
S = S.fillna(0)
ope = numpy.asfarray(S.Open)
high = numpy.asfarray(S.High)
low = numpy.asfarray(S.Low)
close = numpy.asfarray(S.Close)
volume = numpy.asfarray(S.Volume)
##ROI calculation
ROI = [(close[i + 1] - close[i]) / close[i] for i in range(len(close) - 1)]
ROI.append(0) #add zero value for last day
d = pandas.DataFrame(ROI, index=S.index, columns=['ROI'])
d.to_csv("C:\\Users\\...\\Documents\\Data_TA\\{0}\\ROI_{1}.csv".format(
s, c))
##Baselines
try:
bah.append((S.Close['2013-04-30'] - S.Close['2002-05-01']) /
S.Close['2002-05-01'])
sah.append((-S.Close['2013-04-30'] + S.Close['2002-05-01']) /
S.Close['2002-05-01'])
except:
bah.append((S.Close['2013-04-30'] - S.Close['2002-05-02']) /
S.Close['2002-05-02'])
sah.append((-S.Close['2013-04-30'] + S.Close['2002-05-02']) /
S.Close['2002-05-02'])
rp.append(
numpy.dot(numpy.random.uniform(-1, 1, len(S)), numpy.asfarray(ROI)))
##talib application
#overlap
BBANDS = ta.BBANDS(close)
DEMA = ta.DEMA(close)
EMA = ta.EMA(close)
HT_TRENDLINE = ta.HT_TRENDLINE(close)
KAMA = ta.KAMA(close)
MA = ta.MA(close)
MAMA = ta.MAMA(close)
MIDPOINT = ta.MIDPOINT(close)
MIDPRICE = ta.MIDPRICE(high, low)
SAR = ta.SAR(high, low)
SAREXT = ta.SAREXT(high, low)
SMA = ta.SMA(close)
T3 = ta.T3(close)
TEMA = ta.TEMA(close)
TRIMA = ta.TRIMA(close)
WMA = ta.WMA(close)
#momentum
ADX = ta.ADX(high, low, close)
ADXR = ta.ADXR(high, low, close)
APO = ta.APO(close)
AROON = ta.AROON(high, low)
AROONOSC = ta.AROONOSC(high, low)
BOP = ta.BOP(ope, high, low, close)
CCI = ta.CCI(high, low, close)
CMO = ta.CMO(close)
DX = ta.DX(high, low, close)
MACD = ta.MACD(close)
MACDEXT = ta.MACDEXT(close)
MFI = ta.MFI(high, low, close, volume)
MINUS_DI = ta.MINUS_DI(high, low, close)
MINUS_DM = ta.MINUS_DM(high, low)
MOM = ta.MOM(close)
PLUS_DI = ta.PLUS_DI(high, low, close)
PLUS_DM = ta.PLUS_DM(high, low)
PPO = ta.PPO(close)
ROC = ta.ROC(close)
ROCP = ta.ROCP(close)
ROCR = ta.ROCR(close)
RSI = ta.RSI(close)
STOCH = ta.STOCH(high, low, close)
STOCHF = ta.STOCHF(high, low, close)
STOCHRSI = ta.STOCHRSI(close)
TRIX = ta.TRIX(close)
ULTOSC = ta.ULTOSC(high, low, close)
WILLR = ta.WILLR(high, low, close)
#volume
AD = ta.AD(high, low, close, volume)
ADOSC = ta.ADOSC(high, low, close, volume)
OBV = ta.OBV(close, volume)
#cycle
HT_DCPERIOD = ta.HT_DCPERIOD(close)
HT_DCPHASE = ta.HT_DCPHASE(close)
HT_PHASOR = ta.HT_PHASOR(close)
HT_SINE = ta.HT_SINE(close)
HT_TRENDMODE = ta.HT_TRENDMODE(close)
#price
AVGPRICE = ta.AVGPRICE(ope, high, low, close)
MEDPRICE = ta.MEDPRICE(high, low)
TYPPRICE = ta.TYPPRICE(high, low, close)
WCLPRICE = ta.WCLPRICE(high, low, close)
#volatility
ATR = ta.ATR(high, low, close)
NATR = ta.NATR(high, low, close)
TRANGE = ta.TRANGE(high, low, close)
#pattern
CDL2CROWS = ta.CDL2CROWS(ope, high, low, close)
CDL3BLACKCROWS = ta.CDL3BLACKCROWS(ope, high, low, close)
CDL3INSIDE = ta.CDL3INSIDE(ope, high, low, close)
CDL3LINESTRIKE = ta.CDL3LINESTRIKE(ope, high, low, close)
CDL3OUTSIDE = ta.CDL3OUTSIDE(ope, high, low, close)
CDL3STARSINSOUTH = ta.CDL3STARSINSOUTH(ope, high, low, close)
CDL3WHITESOLDIERS = ta.CDL3WHITESOLDIERS(ope, high, low, close)
CDLABANDONEDBABY = ta.CDLABANDONEDBABY(ope, high, low, close)
CDLADVANCEBLOCK = ta.CDLADVANCEBLOCK(ope, high, low, close)
CDLBELTHOLD = ta.CDLBELTHOLD(ope, high, low, close)
CDLBREAKAWAY = ta.CDLBREAKAWAY(ope, high, low, close)
CDLCLOSINGMARUBOZU = ta.CDLCLOSINGMARUBOZU(ope, high, low, close)
CDLCONCEALBABYSWALL = ta.CDLCONCEALBABYSWALL(ope, high, low, close)
CDLCOUNTERATTACK = ta.CDLCOUNTERATTACK(ope, high, low, close)
CDLDARKCLOUDCOVER = ta.CDLDARKCLOUDCOVER(ope, high, low, close)
CDLDOJI = ta.CDLDOJI(ope, high, low, close)
CDLDOJISTAR = ta.CDLDOJISTAR(ope, high, low, close)
CDLDRAGONFLYDOJI = ta.CDLDRAGONFLYDOJI(ope, high, low, close)
CDLENGULFING = ta.CDLENGULFING(ope, high, low, close)
CDLEVENINGDOJISTAR = ta.CDLEVENINGDOJISTAR(ope, high, low, close)
CDLEVENINGSTAR = ta.CDLEVENINGSTAR(ope, high, low, close)
CDLGAPSIDESIDEWHITE = ta.CDLGAPSIDESIDEWHITE(ope, high, low, close)
CDLGRAVESTONEDOJI = ta.CDLGRAVESTONEDOJI(ope, high, low, close)
CDLHAMMER = ta.CDLHAMMER(ope, high, low, close)
CDLHANGINGMAN = ta.CDLHANGINGMAN(ope, high, low, close)
CDLHARAMI = ta.CDLHARAMI(ope, high, low, close)
CDLHARAMICROSS = ta.CDLHARAMICROSS(ope, high, low, close)
CDLHIGHWAVE = ta.CDLHIGHWAVE(ope, high, low, close)
CDLHIKKAKE = ta.CDLHIKKAKE(ope, high, low, close)
CDLHIKKAKEMOD = ta.CDLHIKKAKEMOD(ope, high, low, close)
CDLHOMINGPIGEON = ta.CDLHOMINGPIGEON(ope, high, low, close)
CDLIDENTICAL3CROWS = ta.CDLIDENTICAL3CROWS(ope, high, low, close)
CDLINNECK = ta.CDLINNECK(ope, high, low, close)
CDLINVERTEDHAMMER = ta.CDLINVERTEDHAMMER(ope, high, low, close)
CDLKICKING = ta.CDLKICKING(ope, high, low, close)
CDLKICKINGBYLENGTH = ta.CDLKICKINGBYLENGTH(ope, high, low, close)
CDLLADDERBOTTOM = ta.CDLLADDERBOTTOM(ope, high, low, close)
CDLLONGLEGGEDDOJI = ta.CDLLONGLEGGEDDOJI(ope, high, low, close)
CDLLONGLINE = ta.CDLLONGLINE(ope, high, low, close)
CDLMARUBOZU = ta.CDLMARUBOZU(ope, high, low, close)
CDLMATCHINGLOW = ta.CDLMATCHINGLOW(ope, high, low, close)
CDLMATHOLD = ta.CDLMATHOLD(ope, high, low, close)
CDLMORNINGDOJISTAR = ta.CDLMORNINGDOJISTAR(ope, high, low, close)
CDLMORNINGSTAR = ta.CDLMORNINGSTAR(ope, high, low, close)
CDLONNECK = ta.CDLONNECK(ope, high, low, close)
CDLPIERCING = ta.CDLPIERCING(ope, high, low, close)
CDLRICKSHAWMAN = ta.CDLRICKSHAWMAN(ope, high, low, close)
CDLRISEFALL3METHODS = ta.CDLRISEFALL3METHODS(ope, high, low, close)
CDLSEPARATINGLINES = ta.CDLSEPARATINGLINES(ope, high, low, close)
CDLSHOOTINGSTAR = ta.CDLSHOOTINGSTAR(ope, high, low, close)
CDLSHORTLINE = ta.CDLSHORTLINE(ope, high, low, close)
CDLSPINNINGTOP = ta.CDLSPINNINGTOP(ope, high, low, close)
CDLSTALLEDPATTERN = ta.CDLSTALLEDPATTERN(ope, high, low, close)
CDLSTICKSANDWICH = ta.CDLSTICKSANDWICH(ope, high, low, close)
CDLTAKURI = ta.CDLTAKURI(ope, high, low, close)
CDLTASUKIGAP = ta.CDLTASUKIGAP(ope, high, low, close)
CDLTHRUSTING = ta.CDLTHRUSTING(ope, high, low, close)
CDLTRISTAR = ta.CDLTRISTAR(ope, high, low, close)
CDLUNIQUE3RIVER = ta.CDLUNIQUE3RIVER(ope, high, low, close)
CDLUPSIDEGAP2CROWS = ta.CDLUPSIDEGAP2CROWS(ope, high, low, close)
CDLXSIDEGAP3METHODS = ta.CDLXSIDEGAP3METHODS(ope, high, low, close)
f = numpy.column_stack(
(ATR, NATR, TRANGE, HT_DCPERIOD, HT_DCPHASE, HT_PHASOR[0],
HT_PHASOR[1], HT_SINE[0], HT_SINE[1], HT_TRENDMODE, AVGPRICE,
MEDPRICE, TYPPRICE, WCLPRICE, ADX, ADXR, APO, AROON[0], AROON[1],
AROONOSC, BOP, CCI, CMO, DX, MACD[0], MACD[1], MACD[2], MACDEXT[0],
MACDEXT[1], MACDEXT[2], MFI, MINUS_DI, MINUS_DM, MOM, PLUS_DI,
PLUS_DM, PPO, ROC, ROCP, ROCR, RSI, STOCH[0], STOCH[1], STOCHF[0],
STOCHF[1], STOCHRSI[0], STOCHRSI[1], TRIX, ULTOSC, WILLR, CDL2CROWS,
CDL3BLACKCROWS, CDL3INSIDE, CDL3LINESTRIKE, CDL3OUTSIDE,
CDL3STARSINSOUTH, CDL3WHITESOLDIERS, CDLABANDONEDBABY,
CDLADVANCEBLOCK, CDLBELTHOLD, CDLBREAKAWAY, CDLCLOSINGMARUBOZU,
CDLCONCEALBABYSWALL, CDLCOUNTERATTACK, CDLDARKCLOUDCOVER, CDLDOJI,
CDLDOJISTAR, CDLDRAGONFLYDOJI, CDLENGULFING, CDLEVENINGDOJISTAR,
CDLEVENINGSTAR, CDLGAPSIDESIDEWHITE, CDLGRAVESTONEDOJI, CDLHAMMER,
CDLHANGINGMAN, CDLHARAMI, CDLHARAMICROSS, CDLHIGHWAVE, CDLHIKKAKE,
CDLHIKKAKEMOD, CDLHOMINGPIGEON, CDLIDENTICAL3CROWS, CDLINNECK,
CDLINVERTEDHAMMER, CDLKICKING, CDLKICKINGBYLENGTH, CDLLADDERBOTTOM,
CDLLONGLEGGEDDOJI, CDLLONGLINE, CDLMARUBOZU, CDLMATCHINGLOW,
CDLMATHOLD, CDLMORNINGDOJISTAR, CDLMORNINGSTAR, CDLONNECK,
CDLPIERCING, CDLRICKSHAWMAN, CDLRISEFALL3METHODS, CDLSEPARATINGLINES,
CDLSHOOTINGSTAR, CDLSHORTLINE, CDLSPINNINGTOP, CDLSTALLEDPATTERN,
CDLSTICKSANDWICH, CDLTAKURI, CDLTASUKIGAP, CDLTHRUSTING, CDLTRISTAR,
CDLUNIQUE3RIVER, CDLUPSIDEGAP2CROWS, CDLXSIDEGAP3METHODS, BBANDS[0],
BBANDS[1], BBANDS[2], DEMA, EMA, HT_TRENDLINE, KAMA, MA, MAMA[0],
MAMA[1], MIDPOINT, MIDPRICE, SAR, SAREXT, SMA, T3, TEMA, TRIMA, WMA,
AD, ADOSC, OBV))
h = numpy.apply_along_axis(nor, 0, f) # normalize columnwise
df = pandas.DataFrame(
h,
index=S.index,
columns=[
'ATR', 'NATR', 'TRANGE', 'HT_DCPERIOD', 'HT_DCPHASE',
'HT_PHASOR[0]', 'HT_PHASOR[1]', 'HT_SINE[0]', 'HT_SINE[1]',
'HT_TRENDMODE', 'AVGPRICE', 'MEDPRICE', 'TYPPRICE', 'WCLPRICE',
'ADX', 'ADXR', 'APO', 'AROON[0]', 'AROON[1]', 'AROONOSC', 'BOP',
'CCI', 'CMO', 'DX', 'MACD[0]', 'MACD[1]', 'MACD[2]', 'MACDEXT[0]',
'MACDEXT[1]', 'MACDEXT[2]', 'MFI', 'MINUS_DI', 'MINUS_DM', 'MOM',
'PLUS_DI', 'PLUS_DM', 'PPO', 'ROC', 'ROCP', 'ROCR', 'RSI',
'STOCH[0]', 'STOCH[1]', 'STOCHF[0]', 'STOCHF[1]', 'STOCHRSI[0]',
'STOCHRSI[1]', 'TRIX', 'ULTOSC', 'WILLR', 'CDL2CROWS',
'CDL3BLACKCROWS', 'CDL3INSIDE', 'CDL3LINESTRIKE', 'CDL3OUTSIDE',
'CDL3STARSINSOUTH', 'CDL3WHITESOLDIERS', 'CDLABANDONEDBABY',
'CDLADVANCEBLOCK', 'CDLBELTHOLD', 'CDLBREAKAWAY',
'CDLCLOSINGMARUBOZU', 'CDLCONCEALBABYSWALL', 'CDLCOUNTERATTACK',
'CDLDARKCLOUDCOVER', 'CDLDOJI', 'CDLDOJISTAR', 'CDLDRAGONFLYDOJI',
'CDLENGULFING', 'CDLEVENINGDOJISTAR', 'CDLEVENINGSTAR',
'CDLGAPSIDESIDEWHITE', 'CDLGRAVESTONEDOJI', 'CDLHAMMER',
'CDLHANGINGMAN', 'CDLHARAMI', 'CDLHARAMICROSS', 'CDLHIGHWAVE',
'CDLHIKKAKE', 'CDLHIKKAKEMOD', 'CDLHOMINGPIGEON',
'CDLIDENTICAL3CROWS', 'CDLINNECK', 'CDLINVERTEDHAMMER',
'CDLKICKING', 'CDLKICKINGBYLENGTH', 'CDLLADDERBOTTOM',
'CDLLONGLEGGEDDOJI', 'CDLLONGLINE', 'CDLMARUBOZU',
'CDLMATCHINGLOW', 'CDLMATHOLD', 'CDLMORNINGDOJISTAR',
'CDLMORNINGSTAR', 'CDLONNECK', 'CDLPIERCING', 'CDLRICKSHAWMAN',
'CDLRISEFALL3METHODS', 'CDLSEPARATINGLINES', 'CDLSHOOTINGSTAR',
'CDLSHORTLINE', 'CDLSPINNINGTOP', 'CDLSTALLEDPATTERN',
'CDLSTICKSANDWICH', 'CDLTAKURI', 'CDLTASUKIGAP', 'CDLTHRUSTING',
'CDLTRISTAR', 'CDLUNIQUE3RIVER', 'CDLUPSIDEGAP2CROWS',
'CDLXSIDEGAP3METHODS', 'BBANDS[0]', 'BBANDS[1]', 'BBANDS[2]',
'DEMA', 'EMA', 'HT_TRENDLINE', 'KAMA', 'MA', 'MAMA[0]', 'MAMA[1]',
'MIDPOINT', 'MIDPRICE', 'SAR', 'SAREXT', 'SMA', 'T3', 'TEMA',
'TRIMA', 'WMA', 'AD', 'ADOSC', 'OBV'
])
df.to_csv("C:\\Users\\...\\Documents\\Data_TA\\{0}\\{1}.csv".format(s, c))
xy = np.loadtxt('shinhan_test.csv', delimiter=',')
t_open = np.transpose(xy[:, [0]])
t_high = np.transpose(xy[:, [1]]) #x,y 행렬 전치 해줌 ( 그래야 talib가 먹힘 )
t_low = np.transpose(xy[:, [2]])
t_vol = np.transpose(xy[:, [3]])
t_y = np.transpose(xy)
open = np.array(t_open[0], dtype='float')
high = np.array(t_high[0], dtype='float')
low = np.array(t_low[0], dtype='float')
volume = np.array(t_vol[0], dtype='float')
close = np.array(t_y[0], dtype='float')
b_upper, b_middle, b_lower = ta.BBANDS(close) #bollinger ban
ma = ta.MA(close, timeperiod=predict_len) #moving average
dmi = ta.DX(high, low, close, timeperiod=predict_len) #direct movement index
macd, macdsignal, macdhist = ta.MACD(
close, fastperiod=predict_len, slowperiod=predict_len * 2,
signalperiod=4) #moving average convergence/divergence
slowk, slowd = ta.STOCH(high,
low,
close,
fastk_period=predict_len,
slowk_period=3,
slowk_matype=0,
slowd_period=3,
slowd_matype=0) #Stochastic
mom = ta.MOM(close, timeperiod=predict_len)
rsi = ta.RSI(close, timeperiod=predict_len * 2) #Relative Strength Index
def get_all_factors(open, high, low, close, volume):
'''get all factors when n minutes as the period'''
SMA = talib.MA(close, 30, matype=0)
EMA = talib.MA(close, 30, matype=1)
WMA = talib.MA(close, 30, matype=2)
DEMA = talib.MA(close, 30, matype=3)
TEMA = talib.MA(close, 30, matype=4)
MA5 = talib.MA(close, 5, matype=0)
MA10 = talib.MA(close, 10, matype=0)
MA20 = talib.MA(close, 20, matype=0)
MA60 = talib.MA(close, 60, matype=0)
MA120 = talib.MA(close, 120, matype=0)
# ======================================== MACD =============================================
macd, macdsignal, macdhist = talib.MACD(close,
fastperiod=12,
slowperiod=26,
signalperiod=9)
# ========================================= RSI =============================================
rsi = talib.RSI(close, timeperiod=6)
# ========================================= KDJ =============================================
slowk, slowd = talib.STOCH(high,
low,
close,
fastk_period=9,
slowk_period=3,
slowd_period=3,
slowk_matype=0,
slowd_matype=0)
# ========================================== BOLLING ========================================
upper, middle, lower = talib.BBANDS(
close,
timeperiod=26,
# number of non-biased standard deviations from the mean
nbdevup=2,
nbdevdn=2,
# Moving average dataType: simple moving average here
matype=0)
# =========================================== SAR ============================================
sar = talib.SAR(high, low, acceleration=0.05, maximum=0.2)
mat = open #0
mat = np.column_stack((mat, high)) #1
mat = np.column_stack((mat, low)) #2
mat = np.column_stack((mat, close)) #3
mat = np.column_stack((mat, volume)) #4
mat = np.column_stack((mat, upper)) #5
mat = np.column_stack((mat, middle)) #6
mat = np.column_stack((mat, lower))
mat = np.column_stack((mat, slowk))
mat = np.column_stack((mat, slowd))
mat = np.column_stack((mat, macd))
mat = np.column_stack((mat, macdhist))
mat = np.column_stack((mat, macdsignal))
mat = np.column_stack((mat, SMA))
mat = np.column_stack((mat, EMA))
mat = np.column_stack((mat, WMA))
mat = np.column_stack((mat, DEMA))
mat = np.column_stack((mat, TEMA))
mat = np.column_stack((mat, rsi))
mat = np.column_stack((mat, sar))
mat = np.column_stack((mat, MA5))
mat = np.column_stack((mat, MA10))
mat = np.column_stack((mat, MA20))
mat = np.column_stack((mat, MA60))
mat = np.column_stack((mat, MA120))
# np.savetxt('1316.csv', mat, delimiter = ',')
# 25
return mat
def get_bbands(df, timeperiod=5, nbdevup=2, nbdevdn=2, matype=0):
return talib.BBANDS(df.close,
timeperiod=timeperiod,
nbdevup=nbdevup,
nbdevdn=nbdevdn,
matype=matype)
def bband_func(x, index):
xc = x.close
return ta.BBANDS(xc)[index]
def technical_indicators_panda_dataframe(df_price):
open = df_price['Open']
high = df_price['High']
low = df_price['Low']
close = df_price['Close']
volume = df_price['Volume']
## Commoditty Channel Index
cci = tb.CCI(high, low, close, timeperiod=20)
## Rate of change
roc = tb.ROC(close, timeperiod=21)
## Momentum Indicators Functions
#Aroon
aroondown, aroonup = tb.AROON(high, low, timeperiod=14)
#Average Directional Movement Index
adx = tb.ADX(high, low, close, timeperiod=14)
di_plus = tb.PLUS_DI(high, low, close, timeperiod=14)
di_minus = tb.MINUS_DI(high, low, close, timeperiod=14)
#Money Flow Index
mfi = tb.MFI(high, low, close, volume, timeperiod=14)
#Relative Strength Index
rsi = tb.RSI(close, timeperiod=14)
#Stochastic
slowk, slowd = tb.STOCH(high,
low,
close,
fastk_period=5,
slowk_period=3,
slowk_matype=0,
slowd_period=3,
slowd_matype=0)
#STOCHF - Stochastic Fast
fastk, fastd = tb.STOCHF(high,
low,
close,
fastk_period=5,
fastd_period=3,
fastd_matype=0)
#Moving Average Convergence/Divergence
macd, macdema, macdhist = tb.MACD(close,
fastperiod=12,
slowperiod=26,
signalperiod=9)
## Overlap Studies Functions
#Bollinger Bands
upperband, middleband, lowerband = tb.BBANDS(close,
timeperiod=5,
nbdevup=2,
nbdevdn=2,
matype=0)
#Weighted Moving Average
wma5 = tb.WMA(close, timeperiod=5)
wma30 = tb.WMA(close, timeperiod=30)
#Exponential Moving Average
ema5 = tb.EMA(close, timeperiod=5)
ema21 = tb.EMA(close, timeperiod=21)
#Double Exponential Moving Average
dema5 = tb.DEMA(close, timeperiod=5)
dema21 = tb.DEMA(close, timeperiod=21)
#Triple Exponential Moving Average
tema = tb.TEMA(close, timeperiod=30)
#Triangular Moving Average
trima5 = tb.DEMA(close, timeperiod=5)
trima30 = tb.TRIMA(close, timeperiod=30)
## Volume indicators Functions
#AD - Chaikin A/D Line
chaikin = tb.AD(high, low, close, volume)
##=> MERGE
kwargs = {"cci":cci, "roc": roc, "aroondown": aroondown, "aroonup": aroonup, "adx": adx, "di_plus":di_plus, "di_minus":di_minus, "mfi": mfi, "rsi": rsi, "slowk": slowk, "slowd": slowd, "fastk": fastk, "fastd": fastd, "macd": macd, "macdema": macdema, "macdhist": macdhist,\
"upperband": upperband , "middleband": middleband , "lowerband": lowerband, "wma5" : wma5, "wma30" : wma30, "ema5" : ema5, "ema21" : ema21, "dema21": dema21, "dema5": dema5, "tema": tema, "trima5": trima5,"trima30": trima30,\
"chaikin": chaikin}
return df_price.assign(**kwargs)
def predict(inputHisDf,
lookAheadDays=3,
windowSize=20,
minCorr=0.9,
onlyPositiveCorr=True):
trendWindow = 5
stdGap = 1.25
hisDf = inputHisDf.set_index('index', drop=False).tail(windowSize + 1)
pU, pM, pL = ta.BBANDS(
hisDf['OrigClose'].head(windowSize).astype(float).values,
timeperiod=trendWindow,
nbdevup=stdGap,
nbdevdn=stdGap)
volU, volM, volL = ta.BBANDS(
hisDf['OrigVolume'].head(windowSize).astype(float).values,
timeperiod=trendWindow,
nbdevup=stdGap,
nbdevdn=stdGap)
preP = hisDf['OrigClose'].iat[-2]
curP = hisDf['OrigClose'].iat[-1]
preV = hisDf['OrigVolume'].iat[-2]
curV = hisDf['OrigVolume'].iat[-1]
pUSlope = _array_slope(pU[-trendWindow:])
pMSlope = _array_slope(pM[-trendWindow:])
volUSlope = _array_slope(volU[-trendWindow:])
volMSlope = _array_slope(volM[-trendWindow:])
if volMSlope > 0: #goes upper with larger std
if curP > pL[-1] and preP < pL[-1]:
return 1
if curP < pU[-1] and preP > pU[-1]:
return -1
'''
if pUSlope > 0 and pMSlope > 0 and pUSlope-pMSlope > 0: #goes upper with larger std
if curP > pU[-1] and preP < pU[-1]:
return 1
elif curP < pU[-1] and preP > pU[-1]:
return -1
elif curP < pL[-1] and preP > pL[-1]:
return -1
elif curP > pL[-1] and preP < pL[-1]:
return 1
elif pUSlope < 0 and pMSlope < 0 and pUSlope-pMSlope < 0: #goes down with small std
if curP > pL[-1] and preP < pL[-1]:
return 1
elif curP < pL[-1] and preP > pL[-1]:
return -1
if volUSlope > 0 and volMSlope > 0 and volUSlope-volMSlope > 0:
if curP > pL[-1] and preP < pL[-1]:
return 1
if curP > pM[-1] and pMSlope > 0:
return 1
if curP < pU[-1] and preP > pU[-1]:
return -1
if curP < pM[-1] and pMSlope < 0:
return -1
elif volUSlope < 0 and volMSlope < 0 and volUSlope-volMSlope < 0:
if curP < pU[-1] and preP > pU[-1]:
return -1
'''
return 0
def _rolling_algo(self, data, n, a1, a2, i):
""" 逐步运行函数。"""
upper, middle, lower = talib.BBANDS(data, n, a1, a2)
return (upper[i], middle[i], lower[i])
close_prices = np.array([x[4] for x in rates_sorted])
# low_prices = np.array([x[1] for x in rates_sorted])
# high_prices = np.array([x[2] for x in rates_sorted])
close_timestamp = rates_sorted[-1][0]
if args.yolo:
print(close_prices)
# Compute Bollinger Bands
# nbdevup: number of non-biased standard deviations from the mean
# nbdevdn:
# matype: moving average type: simple moving average here (0)
upper, middle, lower = talib.BBANDS(
close_prices,
timeperiod=args.periods,
nbdevup=2,
nbdevdn=2,
matype=2)
action = 'NOTHING'
# If the last close price is under the lower band
if close_prices[-1] <= lower[-1]:
action = 'BUY'
# If close_prices is above the recent upper band
elif close_prices[-1] >= upper[-1]:
action = 'SELL'
print('{} >>> {}'.format(close_timestamp, action))
result.append([close_timestamp, action])
