def generate_3mindata(self, am:ArrayManager, bar:BarData):
offset = -self.offset
offset_m = int(offset / 2)
calc_nums = np.array(self.ma_tag[-offset:-1])
# var_val = np.var(calc_nums)
std_val = np.std(calc_nums)
std_val2 = np.std(np.array(self.ma_tag[-10:-1]))
std_val3 = np.std(np.array(am.range[-30:-10]))
ma = self.ma_tag[-1]
mean_val = np.mean(calc_nums)
mean_val2 = np.mean(np.array(self.ma_tag[-5:-1]))
mean_val3 = np.mean(np.array(self.ma_tag[-20:-1]))
mean_val4 = np.mean(np.array(self.ma_tag[-30:-5]))
kdj_val = am.kdj()
deg1 = calc_regress_deg(am.close[offset : offset_m], False)
deg2 = calc_regress_deg(am.close[offset_m :], False)
deg3 = calc_regress_deg(am.close[-10 :], False)
deg_full = calc_regress_deg(am.close[offset :], False)
wave = self.wave(am.close[-30:])
wave_r_sum = np.sum(wave["range"])
macd=am.macd(20,40, 16)
calc_data = (dict(
kdj=[round(kdj_val["k"][-1],2),round(kdj_val["d"][-1],2),round(kdj_val["j"][-1],2)],
cci_20=am.cci(20),rsi=am.rsi(20),adx=am.adx(20),boll=am.boll(20, 3.4),
macd=[round(macd[0],2),round(macd[1],2),round(macd[2],2)],
deg40_20=round(deg1,2), deg20_0=round(deg2,2), deg20_10=round(calc_regress_deg(am.close[-20:-10], False),2), deg10_0=round(deg3,2),
deg30_15=round(calc_regress_deg(am.close[-30:-15], False),2), deg15_0=round(calc_regress_deg(am.close[-15:], False),2),deg_f=round(deg_full,2),
atr=round(am.atr(10, length=15), 3), tr=round(am.atr(1, length=2), 3),atr_40=round(am.atr(40, length=42), 3),
time=bar.datetime, price=bar.close_price, ma=round(ma, 2),
std_40=round(std_val, 2),mean40=round(mean_val,2), mean_std=np.mean(self.std_range.data[-5:]),
std_10=round(std_val2,2), mean30_10=round(mean_val4,2), mean10=round(mean_val2,2),
vol=am.volume[-1], std_range=self.std_range.data[-1:-5:-1], range=am.range[-1:-5:-1].tolist(),
range_sum=np.sum(am.range[-5:]),
pattern=list(map(lambda x: KLINE_PATTERN_CHINESE[x], self.pattern_record.keys())),
ma120t=self.ma120_track,
ma120t_list=self.ma120_track_list[-1:-10:-1],
ma120t_sort=sorted(self.ma120_track_list[-20:-1], key=abs),
ma120t_sum=np.sum(self.ma120_track_list[-20:-1] + [self.ma120_track]),
ma120t_mean=np.mean(self.ma120_track_list[-20:-1] + [self.ma120_track]),
ma120t_std=np.std(self.ma120_track_list[-20:-1] + [self.ma120_track]),
ma_info=list(map(lambda x:x["std"], self.ma_info[-1:])),
wave_cnt=len(wave), wave_r_sum=wave_r_sum, atr_mean=np.mean(am.atr(20, array=True,length=240)[-200:])
))
return calc_data
def generate_data(self, am: ArrayManager, bar: BarData):
offset = -self.offset
offset_m = int(offset / 2)
std_val3 = np.std(np.array(am.range[-30:-10]))
kdj_val = am.kdj()
has_kdj_recore = False
k = kdj_val["k"]
d = kdj_val["d"]
j = kdj_val["j"]
if (k[-1] > 75 and d[-1] > 75 and j[-1] > 75) or \
(k[-1] < 25 and d[-1] < 25 and j[-1] < 75):
if (j[-2] < k[-2] or j[-2] < d[-2]) and (j[-1] > k[-1] and j[-1] > d[-1]) \
or \
(j[-2] > k[-2] or j[-2] > d[-2]) and (j[-1] < k[-1] and j[-1] < d[-1]):
has_kdj_recore = True
t = bar.datetime
self.kdj_record.append(
(t.strftime("%H:%M:%S"), round(k[-1], 3), round(d[-1], 3),
round(j[-1], 3)))
deg1 = calc_regress_deg(am.close[offset:offset_m], False)
deg2 = calc_regress_deg(am.close[offset_m:], False)
deg3 = calc_regress_deg(am.close[-10:], False)
deg_full = calc_regress_deg(am.close[offset:], False)
macd = am.macd(20, 40, 16)
calc_data = (dict(
kdj=[
round(kdj_val["k"][-1], 2),
round(kdj_val["d"][-1], 2),
round(kdj_val["j"][-1], 2)
],
cci_20=am.cci(20),
rsi=am.rsi(20),
adx=am.adx(20),
boll=am.boll(20, 3.4),
macd=[round(macd[0], 2),
round(macd[1], 2),
round(macd[2], 2)],
deg40_20=round(deg1, 2),
deg20_0=round(deg2, 2),
deg20_10=round(calc_regress_deg(am.close[-20:-10], False), 2),
deg30_15=round(calc_regress_deg(am.close[-30:-15], False), 2),
deg15_0=round(calc_regress_deg(am.close[-15:], False), 2),
deg_f=round(deg_full, 2),
deg30_10=round(calc_regress_deg(am.close[-30:-10], False), 2),
deg10_0=round(deg3, 2),
atr=round(am.atr(10, length=15), 3),
tr=round(am.atr(1, length=2), 3),
atr_40=round(am.atr(40, length=42), 3),
time=bar.datetime,
price=bar.close_price,
mean_std=np.mean(self.std_range.data[-5:]),
vol=am.volume[-1],
std_range=self.std_range.data[-1:-5:-1],
range=am.range[-1:-5:-1].tolist(),
range_sum=np.sum(am.range[-5:]),
pattern=list(
map(lambda x: KLINE_PATTERN_CHINESE[x],
self.pattern_record.keys())),
atr_mean=np.mean(am.atr(20, array=True, length=240)[-200:]),
))
if self.ma_info.info.index.size >= 31:
ma5 = self.ma_info.info[5][-31:]
x = AnalyseWave(ma5)
calc_data["ma5_info"] = x.optimize
ma10 = self.ma_info.info[10][-31:]
x = AnalyseWave(ma10)
calc_data["ma10_info"] = x.optimize
return calc_data
class MaLevelTrackStrategy(CtaTemplate):
author = "用Python的交易员"
ma_level = [5, 10, 20, 30, 120]
ma_tag = []
bd = []
fast_ma0 = 0.0
fast_ma1 = 0.0
slow_ma0 = 0.0
slow_ma1 = 0.0
request_order = []
bar_identify = []
parameters = ["ma_level"]
variables = ["fast_ma0", "fast_ma1", "slow_ma0", "slow_ma1"]
def __init__(self, cta_engine, strategy_name, vt_symbol, setting):
""""""
super(MaLevelTrackStrategy, self).__init__(cta_engine, strategy_name,
vt_symbol, setting)
self.bg = BarGenerator(self.on_bar, 15, self.on_1min_bar)
self.am = ArrayManager(400)
self.am3 = ArrayManager(150)
self.bg3 = BarGenerator(self.on_bar, 3, self.on_3min_bar)
self.am5 = ArrayManager(120)
self.bg5 = BarGenerator(self.on_bar, 5, self.on_5min_bar)
self.order_data = None
self.positions = Position(self)
self.std_range = IntervalGen(np.std, 5)
self.std_range3 = IntervalGen(np.std, 5)
self.std_range5 = IntervalGen(np.std, 5)
self.pattern_record = PatternRecord()
# self.pattern_record.set_expiry([KlinePattern.CDLEVENINGSTAR], 3)
self.pattern_record.set_expiry(list(KlinePattern), 1)
five_min_open_5 = partial(self.reverse_shape_strategy,
setting={
"atr": 10,
"atr_valve": 0.8,
"deg1": (10, 5),
"deg2": 5
})
self.open_strategy = {
"1": [self.reverse_shape_strategy],
"5": [five_min_open_5],
}
self.offset = 40
self.ma120_track = None
self.ma120_track_list = []
def on_init(self):
"""
Callback when strategy is inited.
"""
self.write_log("策略初始化")
self.load_bar(10)
def on_start(self):
"""
Callback when strategy is started.
"""
self.write_log("策略启动")
self.put_event()
def on_stop(self):
"""
Callback when strategy is stopped.
"""
self.write_log("策略停止")
self.put_event()
def on_tick(self, tick: TickData):
"""
Callback of new tick data update.
"""
self.bg.update_tick(tick)
self.bg3.update_tick(tick)
self.bg5.update_tick(tick)
def on_3min_bar(self, bar: BarData):
self.am3.update_bar(bar)
self.std_range3.update(self.am3.range[-1])
if not self.am.inited or not self.trading:
return
pattern = self.am3.pattern(
[KlinePattern.CDLEVENINGSTAR, KlinePattern.CDL2CROWS])
if len(pattern) > 0:
print(pattern)
self.pattern_record.add_pattern(pattern)
# deg = calc_regress_deg(self.am3.close[-20:])
def wave(self, data, window=0.0002):
if len(data) <= 0:
return
# r = array[::-1]
result = {"value": [], "range": [], "pos": [], "length": []}
r = data
l = len(data) - 1
now = r[0]
# v_list.append(now)
# p_list.append(0)
pos = 1
vol = 0
u_tag = None
d_tag = None
end_tag = None
start_pos = 0
while pos < l:
if math.isnan(now):
now = r[pos]
pos += 1
continue
else:
start_pos = pos - 1
break
while pos < l:
if now < r[pos]:
u_tag = pos
if d_tag:
diff = r[start_pos] - r[d_tag]
if abs(diff /
r[start_pos]) > window and d_tag - start_pos > 1:
end_tag = d_tag
elif now > r[pos]:
d_tag = pos
if u_tag:
diff = r[start_pos] - r[u_tag]
if abs(diff /
r[start_pos]) > window and d_tag - start_pos > 1:
end_tag = u_tag
if end_tag is not None:
result["range"].append(r[end_tag] / r[start_pos] - 1)
result["length"].append(end_tag - start_pos)
start_pos = end_tag
result["value"].append(r[end_tag])
result["pos"].append(end_tag)
end_tag = None
vol += r[pos] - now
now = r[pos]
pos += 1
return pd.DataFrame(result)
def mode_identify(self, bar: BarData):
self.bar_identify = []
hl_scale = round(bar.high_price / bar.low_price - 1, 4)
if hl_scale > 0.001:
diff = bar.high_price - bar.low_price
diff_up = bar.low_price + diff / 2 * 1.20
diff_down = bar.low_price + diff / 2 * 0.80
close = bar.close_price
if bar.open_price < diff_up and bar.open_price > diff_down and \
bar.close_price < diff_up and bar.close_price > diff_down:
if bar.close_price > bar.open_price:
print("绿十字星", bar.datetime, bar.high_price, bar.low_price,
diff, diff_up, diff_down, bar.open_price,
bar.close_price)
else:
print("红十字星", bar.datetime, bar.high_price, bar.low_price,
diff, diff_up, diff_down, bar.open_price,
bar.close_price)
def on_5min_bar(self, bar: BarData):
self.std_range5.update(self.am5.range[-1])
self.am5.update_bar(bar)
if not self.am.inited or not self.trading:
return
self.on_strategy(self.am5, bar, self.open_strategy["5"])
# pattern_list = [KlinePattern.CDLEVENINGSTAR, KlinePattern.CDL2CROWS, KlinePattern.CDLCONCEALBABYSWALL, KlinePattern.CDLEVENINGDOJISTAR]
# pattern = self.am5.pattern(list(KlinePattern))
# if len(pattern) > 0:
# print(list(map(lambda x: (KLINE_PATTERN_CHINESE[x[0]],x[1]), pattern)))
# self.pattern_record.add_pattern(pattern)
# deg_full = calc_regress_deg(self.am.close[-40 :], False)
# print("deg:",deg_full)
# self.pattern_record.update()
def open_v3(self, am: ArrayManager, bar: BarData):
std_val2 = np.std(np.array(self.ma_tag[-10:-1]))
mean_val2 = np.mean(np.array(self.ma_tag[-10:-1]))
mean = np.mean(np.array(self.ma_tag[-30:-10]))
if std_val2 < 0.2:
if mean_val2 > 3:
if mean_val2 >= (mean + 1):
return self.buy(bar.close_price, 1, type=OrderType.MARKET)
elif mean_val2 < 2:
if mean_val2 <= (mean - 1):
return self.short(bar.close_price,
1,
type=OrderType.MARKET)
def open_v1(self, am: ArrayManager, bar: BarData):
offset = -40
offset_m = int(offset / 2)
calc_nums = np.array(self.ma_tag[-offset:-1])
mean_val = np.mean(calc_nums)
# var_val = np.var(calc_nums)
std_val = np.std(calc_nums)
if std_val < 1 and mean_val < 2 and self.ma_tag[-1] >= (mean_val + 2):
return self.buy(bar.close_price, 1, type=OrderType.MARKET)
elif std_val < 1 and mean_val > 3 and self.ma_tag[-1] <= (mean_val -
2):
return self.short(bar.close_price, 1, type=OrderType.MARKET)
def open_v2(self, am: ArrayManager, bar: BarData):
std_val2 = np.std(np.array(self.ma_tag[-10:-1]))
mean_val2 = np.mean(np.array(self.ma_tag[-10:-1]))
mean = np.mean(np.array(self.ma_tag[-30:-10]))
if std_val2 < 0.2:
if mean_val2 > 2.5:
if mean_val2 >= (mean + 1):
return self.buy(bar.close_price, 1, type=OrderType.MARKET)
elif mean_val2 < 2.5:
if mean_val2 <= (mean - 1):
return self.short(bar.close_price,
1,
type=OrderType.MARKET)
def open2(self, am: ArrayManager, bar: BarData, calc_data):
deg = calc_data["deg20"]
ma = self.ma_tag[-1]
if deg > 0.5 and ma > 3 and self.am5.range[-1] > -0.002:
return self.buy(bar.close_price, 1, type=OrderType.MARKET)
elif deg < -0.5 and ma < 2 and self.am5.range[-1] < 0.002:
return self.short(bar.close_price, 1, type=OrderType.MARKET)
def open1(self, am: ArrayManager, bar: BarData, calc_data):
mean = calc_data["mean30_10"]
mean_val2 = calc_data["mean10"]
# if std_val2 < 0.2:
if mean_val2 > 3.5 and mean_val2 >= (mean + 2):
return self.buy(bar.close_price, 1, type=OrderType.MARKET)
elif mean_val2 < 1.5 and mean_val2 <= (mean - 2):
return self.short(bar.close_price, 1, type=OrderType.MARKET)
# v形反转捕获
def reverse_shape_strategy(self,
am: ArrayManager,
bar: BarData,
calc_data,
setting={
"atr": 40,
"atr_valve": 0.8,
"deg1": (40, 20),
"deg2": (20, 0),
}):
deg1 = calc_data["deg40_20"]
deg2 = calc_data["deg20_0"]
kdj = calc_data["kdj"]
atr = self.am.atr(40)
if atr < 0.08:
return
if deg1 > 0 and deg2 > 0 or \
deg1 < 0 and deg2 < 0:
return
if not (abs(deg1) > 0.15 and abs(deg2) > 0.1 and
(abs(deg1) + abs(deg2)) > 0.3):
return
close = am.close[-40:]
min_val = np.min(close)
max_val = np.max(close)
mid_val = max_val if deg1 > 0 else min_val
mid_pos = np.where(close == mid_val)[0][0]
if mid_pos < 10 or mid_pos > 30:
return
start_val = np.min(close[:mid_pos]) if deg1 > 0 else np.max(
close[:mid_pos])
start_pos = np.where(close == start_val)[0][0]
l = mid_pos - start_pos
# pos2 = np.where(close == min_val)[0][0]
x_fit = reg_util.regress_y_polynomial(close[:mid_pos], zoom=True)
deg1_remake = calc_regress_deg(x_fit[:abs(mid_pos)], False)
y_fit = reg_util.regress_y_polynomial(close[mid_pos:], zoom=True)
deg2_remake = calc_regress_deg(y_fit[:abs(mid_pos)], False)
print(start_pos, mid_pos, deg1, deg2, deg1_remake, deg2_remake, l,
start_val, mid_val)
if deg2 < 0:
if kdj[0] < 20 and kdj[1] < 10 and kdj[2] < 10:
# if kdj[2] < 10:
return self.short(bar.close_price, 1, type=OrderType.MARKET)
else:
if kdj[0] > 80 and kdj[1] > 90 and kdj[2] > 90:
# if kdj[2] > 90:
return self.buy(bar.close_price, 1, type=OrderType.MARKET)
# print("找到大v形:", deg1, deg2 )
def open5(self, am: ArrayManager, bar: BarData, calc_data):
ma = self.ma_tag[-1]
mean = calc_data["mean30_10"]
atr = self.am.atr(10, array=True, length=20)
tr = self.am.atr(1, array=True, length=11)
# self.ma120_track
ma120 = self.am.sma(120)
# if std_val2 < 0.2:
mean_std = calc_data["mean_std"]
if mean_std < 0.8 and tr[-1] > 0.1 and tr[-1] / tr[-10] > 3 and tr[
-1] / atr[-1] >= 1.7 and tr[-10] / atr[-10] < 1:
if np.sum(self.am.range[-10:]) > 0 and self.ma120_track > 0:
return self.buy(bar.close_price, 1, type=OrderType.MARKET)
elif self.ma120_track < 0:
return self.short(bar.close_price, 1, type=OrderType.MARKET)
def open_kline1(self, am: ArrayManager, bar: BarData, calc_data):
if KlinePattern.CDLEVENINGSTAR not in self.pattern_record:
return
# if std_val2 < 0.2:
deg = calc_regress_deg(self.am.close[-5:], False)
print("kline_strategy", deg)
if deg < -0.1:
return self.short(bar.close_price, 1, type=OrderType.MARKET)
def generate_data(self, bar: BarData):
offset = -self.offset
offset_m = int(offset / 2)
calc_nums = np.array(self.ma_tag[-offset:-1])
# var_val = np.var(calc_nums)
std_val = np.std(calc_nums)
std_val2 = np.std(np.array(self.ma_tag[-10:-1]))
std_val3 = np.std(np.array(self.am.range[-30:-10]))
ma = self.ma_tag[-1]
mean_val = np.mean(calc_nums)
mean_val2 = np.mean(np.array(self.ma_tag[-5:-1]))
mean_val3 = np.mean(np.array(self.ma_tag[-20:-1]))
mean_val4 = np.mean(np.array(self.ma_tag[-30:-5]))
kdj_val = self.am.kdj()
deg1 = calc_regress_deg(self.am.close[offset:offset_m], False)
deg2 = calc_regress_deg(self.am.close[offset_m:], False)
deg3 = calc_regress_deg(self.am.close[-10:], False)
deg_full = calc_regress_deg(self.am.close[offset:], False)
wave = self.wave(self.am.close[-30:])
wave_r_sum = np.sum(wave["range"])
macd = self.am.macd(20, 40, 16)
calc_data = (dict(
kdj=[
round(kdj_val["k"][-1], 2),
round(kdj_val["d"][-1], 2),
round(kdj_val["j"][-1], 2)
],
cci_20=self.am.cci(20),
rsi=self.am.rsi(20),
adx=self.am.adx(20),
boll=self.am.boll(20, 3.4),
macd=[round(macd[0], 2),
round(macd[1], 2),
round(macd[2], 2)],
deg40_20=round(deg1, 2),
deg20_0=round(deg2, 2),
deg20_10=round(calc_regress_deg(self.am.close[-20:-10], False), 2),
deg10_0=round(deg3, 2),
deg30_15=round(calc_regress_deg(self.am.close[-30:-15], False), 2),
deg15_0=round(calc_regress_deg(self.am.close[-15:], False), 2),
deg_f=round(deg_full, 2),
atr=round(self.am.atr(10, length=15), 3),
tr=round(self.am.atr(1, length=2), 3),
atr_40=round(self.am.atr(40, length=42), 3),
time=bar.datetime,
price=bar.close_price,
ma=round(ma, 2),
std_40=round(std_val, 2),
mean40=round(mean_val, 2),
mean_std=np.mean(self.std_range.data[-5:]),
std_10=round(std_val2, 2),
mean30_10=round(mean_val4, 2),
mean10=round(mean_val2, 2),
vol=self.am.volume[-1],
std_range=self.std_range.data[-1:-5:-1],
range=self.am.range[-1:-5:-1].tolist(),
range_sum=np.sum(self.am.range[-5:]),
pattern=list(
map(lambda x: KLINE_PATTERN_CHINESE[x],
self.pattern_record.keys())),
ma120t=self.ma120_track,
ma120t_list=self.ma120_track_list[-1:-10:-1],
ma120t_sort=sorted(self.ma120_track_list[-20:-1], key=abs),
ma120t_sum=np.sum(self.ma120_track_list[-20:-1] +
[self.ma120_track]),
ma120t_mean=np.mean(self.ma120_track_list[-20:-1] +
[self.ma120_track]),
ma120t_std=np.std(self.ma120_track_list[-20:-1] +
[self.ma120_track]),
wave_cnt=len(wave),
wave_r_sum=wave_r_sum,
atr_mean=np.mean(self.am.atr(20, array=True, length=240)[-200:])))
return calc_data
def on_strategy(self, am: ArrayManager, bar: BarData, strategy_list):
calc_data = self.generate_data(bar)
order_id = None
if self.pos == 0:
for open_strategy in strategy_list:
if order_id is not None:
break
order_id = open_strategy(am, bar, calc_data)
else:
order_id = self.positions.on_strategy(bar, calc_data)
if order_id is not None:
offset = -self.offset
offset_m = int(offset / 2)
self.tracker["trade_info"].append(
(self.am.time_array[offset], self.am.time_array[offset_m],
bar.datetime, calc_data["deg40_20"], calc_data["deg20_0"]))
self.request_order.extend(order_id)
if self.tracker is not None:
self.tracker["ma_tag_ls"].append(calc_data)
def on_1min_bar(self, bar: BarData):
self.am.update_bar(bar)
am = self.am
max_len = self.ma_level[-1] + 20
data = self.am.close[-max_len:-1]
ma_lvl = []
for i in self.ma_level:
ma = self.am.sma(i, True)[-1]
ma_lvl.append(ma)
l = len(ma_lvl)
ma_lvl_tag = []
now = bar.close_price
direction = 1 if now > ma_lvl[0] else 0
ma_lvl_tag.append(direction)
for i in range(l - 1):
val = 1 if ma_lvl[i] > ma_lvl[i + 1] else 0
ma_lvl_tag.append(val)
bincount_val = np.bincount(np.array(ma_lvl_tag))
tag_val = 0
if len(bincount_val) == 2:
tag_val = bincount_val[1]
if len(self.ma_tag) < 200:
self.ma_tag.append(tag_val)
else:
self.ma_tag[:-1] = self.ma_tag[1:]
self.ma_tag[-1] = tag_val
if self.tracker is not None:
self.tracker["bar_data"].append(bar)
self.std_range.update(self.am.range[-1])
ma120 = self.am.sma(120)
if bar.close_price >= ma120:
if self.ma120_track is None:
self.ma120_track = 1
elif self.ma120_track > 0:
self.ma120_track += 1
else:
self.ma120_track_list.append(self.ma120_track)
self.ma120_track = 1
elif bar.close_price < ma120:
if self.ma120_track is None:
self.ma120_track = -1
elif self.ma120_track < 0:
self.ma120_track -= 1
else:
self.ma120_track_list.append(self.ma120_track)
self.ma120_track = -1
if not am.inited or not self.trading:
return
self.on_strategy(am, bar, self.open_strategy["1"])
# median_val = np.median(calc_nums)
self.put_event()
def on_bar(self, bar: BarData):
"""
Callback of new bar data update.
"""
self.bg3.update_bar(bar)
self.bg5.update_bar(bar)
self.bg.update_bar(bar)
# def init_order_data(self):
# self.order_data = np.array([])
def on_order(self, order: OrderData):
"""
Callback of new order data update.
"""
print("{}产生了{},价格为{},笔数为{},交易{},pos={}".format(
order.datetime.strftime("%m/%d %H:%M:%S"),
order.offset.value + order.direction.value, order.price,
order.volume, order.status.value, self.pos))
if order.vt_orderid in self.request_order:
self.positions.on_order(order)
if order.status == Status.ALLTRADED or order.status == Status.CANCELLED or order.status == Status.REJECTED:
self.request_order.remove(order.vt_orderid)
# if order.status == Status.ALLTRADED or order.status == Status.PARTTRADED:
# if order.direction == Direction.LONG:
# if self.positions.volumn == 0:
# self.positions.close_price = round(order.price * 0.995)
# self.positions.volumn += order.volume
# elif order.direction == Direction.SHORT:
# self.positions.volumn -= order.volume
# elif order.direction == Direction.NET:
# self.positions.volumn = order.volume
def on_trade(self, trade: TradeData):
"""
Callback of new trade data update.
"""
self.put_event()
def on_stop_order(self, stop_order: StopOrder):
"""
Callback of stop order update.
"""
pass
def generate_data(self, am:ArrayManager, bar:BarData):
offset = -self.offset
offset_m = int(offset / 2)
calc_nums = np.array(self.ma_tag[-offset:-1])
# var_val = np.var(calc_nums)
std_val = np.std(calc_nums)
std_val2 = np.std(np.array(self.ma_tag[-10:-1]))
std_val3 = np.std(np.array(am.range[-30:-10]))
ma = self.ma_tag[-1]
mean_val = np.mean(calc_nums)
mean_val2 = np.mean(np.array(self.ma_tag[-5:-1]))
mean_val3 = np.mean(np.array(self.ma_tag[-20:-1]))
mean_val4 = np.mean(np.array(self.ma_tag[-30:-5]))
kdj_val = am.kdj()
has_kdj_recore = False
k = kdj_val["k"]
d = kdj_val["d"]
j = kdj_val["j"]
if (k[-1] > 75 and d[-1] > 75 and j[-1] > 75) or \
(k[-1] < 25 and d[-1] < 25 and j[-1] < 75):
if (j[-2] < k[-2] or j[-2] < d[-2]) and (j[-1] > k[-1] and j[-1] > d[-1]) \
or \
(j[-2] > k[-2] or j[-2] > d[-2]) and (j[-1] < k[-1] and j[-1] < d[-1]):
has_kdj_recore = True
t = local_to_eastern(bar.datetime.timestamp())
self.kdj_record.append((t.strftime("%H:%M:%S"), round(k[-1], 3), round(d[-1], 3), round(j[-1], 3)))
deg1 = calc_regress_deg(am.close[offset : offset_m], False)
deg2 = calc_regress_deg(am.close[offset_m :], False)
deg3 = calc_regress_deg(am.close[-10 :], False)
deg_full = calc_regress_deg(am.close[offset :], False)
wave = self.wave(am.close[-30:])
wave_r_sum = np.sum(wave["range"])
macd=am.macd(20,40, 16)
calc_data = (dict(
ma_info=self.ma_info[-1:],
kdj=[round(kdj_val["k"][-1],2),round(kdj_val["d"][-1],2),round(kdj_val["j"][-1],2)],
cci_20=am.cci(20),rsi=am.rsi(20),adx=am.adx(20),boll=am.boll(20, 3.4),
macd=[round(macd[0],2),round(macd[1],2),round(macd[2],2)],
deg40_20=round(deg1,2), deg20_0=round(deg2,2), deg20_10=round(calc_regress_deg(am.close[-20:-10], False),2),
deg30_10=round(calc_regress_deg(am.close[-30:-10], False),2),deg10_0=round(deg3,2),
deg30_15=round(calc_regress_deg(am.close[-30:-15], False),2), deg15_0=round(calc_regress_deg(am.close[-15:], False),2),deg_f=round(deg_full,2),
atr=round(am.atr(10, length=15), 3), tr=round(am.atr(1, length=2), 3),atr_40=round(am.atr(40, length=42), 3),
time=bar.datetime, price=bar.close_price, ma=round(ma, 2),
std_40=round(std_val, 2),mean40=round(mean_val,2), mean_std=np.mean(self.std_range.data[-5:]),
std_10=round(std_val2,2), mean30_10=round(mean_val4,2), mean10=round(mean_val2,2),
vol=am.volume[-1], std_range=self.std_range.data[-1:-5:-1], range=am.range[-1:-5:-1].tolist(),
range_sum=np.sum(am.range[-5:]),
pattern=list(map(lambda x: KLINE_PATTERN_CHINESE[x], self.pattern_record.keys())),
ma120t=self.ma120_track,
ma120t_list=self.ma120_track_list[-1:-10:-1],
ma120t_sort=sorted(self.ma120_track_list[-20:-1], key=abs),
ma120t_sum=np.sum(self.ma120_track_list[-20:-1] + [self.ma120_track]),
ma120t_mean=np.mean(self.ma120_track_list[-20:-1] + [self.ma120_track]),
ma120t_std=np.std(self.ma120_track_list[-20:-1] + [self.ma120_track]),
wave_cnt=len(wave), wave_r_sum=wave_r_sum, atr_mean=np.mean(am.atr(20, array=True,length=240)[-200:]),
kdj_record=self.kdj_record[-10:],
))
if self.ma_info[-1]["ma5"] <= 0.16:
calc_data["kdj_key"] = True
return calc_data
