def test_bei_chi():
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
kline.loc[:, "dt"] = pd.to_datetime(kline.dt)
ka = KlineAnalyze(kline, name="日线", max_xd_len=10, verbose=False)
bi1 = {
"start_dt": ka.bi_list[-11]['dt'],
"end_dt": ka.bi_list[-10]['dt'],
"direction": "down"
}
bi2 = {
"start_dt": ka.bi_list[-13]['dt'],
"end_dt": ka.bi_list[-12]['dt'],
"direction": "down"
}
x1 = ka.is_bei_chi(bi1, bi2, mode="bi", adjust=0.9)
xd1 = {
"start_dt": ka.xd_list[-2]['dt'],
"end_dt": ka.xd_list[-1]['dt'],
"direction": "down"
}
xd2 = {
"start_dt": ka.xd_list[-4]['dt'],
"end_dt": ka.xd_list[-3]['dt'],
"direction": "down"
}
x2 = ka.is_bei_chi(xd1, xd2, mode='xd', adjust=0.9)
print('背驰计算结果：{}，{}'.format(x1, x2))
def test_update_ta():
ka = KlineAnalyze(kline, name="日线", max_raw_len=2000, verbose=False)
ma_x1 = dict(ka.ma[-1])
macd_x1 = dict(ka.macd[-1])
ka.update(kline.iloc[-1].to_dict())
ma_x2 = dict(ka.ma[-1])
macd_x2 = dict(ka.macd[-1])
assert ma_x1['dt'] == ma_x2['dt']
assert [round(x, 2) for x in ma_x1.values() if isinstance(x, float)] == \
[round(x, 2) for x in ma_x2.values() if isinstance(x, float)]
assert macd_x1['dt'] == macd_x2['dt']
assert [round(x, 2) for x in macd_x1.values() if isinstance(x, float)] == \
[round(x, 2) for x in macd_x2.values() if isinstance(x, float)]
def test_kline_analyze():
ka = KlineAnalyze(kline, name="日线", max_raw_len=2000)
# 测试绘图
file_img = "kline.png"
ka.to_image(file_img, max_k_count=5000)
assert os.path.exists(file_img)
os.remove(file_img)
file_html = "kline.html"
ka.to_html(file_html)
assert os.path.exists(file_html)
os.remove(file_html)
# 测试分型识别结果
assert ka.fx_list[-1]['fx_mark'] == 'g' and ka.fx_list[-1]['fx'] == 3456.97
assert ka.fx_list[-5]['fx_mark'] == 'g' and ka.fx_list[-5]['fx'] == 2983.44
# 测试笔识别结果
assert ka.bi_list[-1]['fx_mark'] == 'g' and ka.bi_list[-1]['bi'] == 3456.97
assert ka.bi_list[-4]['fx_mark'] == 'd' and ka.bi_list[-4]['bi'] == 2646.8
# 测试线段识别结果
assert ka.xd_list[-2]['fx_mark'] == 'g' and ka.xd_list[-2]['xd'] == 3288.45
assert ka.xd_list[-3]['fx_mark'] == 'd' and ka.xd_list[-3]['xd'] == 2440.91
# 测试增量更新
ka_raw_len = len(ka.kline_raw)
for x in [2890, 2910, 2783, 3120]:
k = dict(ka.kline_raw[-1])
k['close'] = x
ka.update(k)
assert len(ka.kline_raw) == ka_raw_len
assert ka.kline_raw[-1]['close'] == x
def selector(symbols: List):
"""输入股票列表，输入符合买点定义的股票"""
res = []
for symbol in tqdm(symbols, desc="缠论选股"):
try:
kline = get_kline(symbol=symbol, end_date=datetime.now(), freq="30min", count=1000)
ka = KlineAnalyze(kline, ma_params=(5, 34, 60, 250), bi_mode="new")
if ka.ma[-1]['ma60'] >= ka.latest_price >= ka.ma[-1]['ma250']:
# print("{} 满足条件1：ma60 > close > ma233".format(symbol))
points = ka.bi_list[-7:]
if len(points) == 7 and points[-1]['fx_mark'] == 'd':
zs_g = min([x['bi'] for x in points[2:6] if x['fx_mark'] == 'g'])
zs_d = max([x['bi'] for x in points[2:6] if x['fx_mark'] == 'd'])
if zs_g > zs_d:
# print("{} 满足条件2：向下中枢完成".format(symbol))
date_span = [points[-5]['dt'], points[-1]['dt']]
low = [x['low'] for x in ka.kline_raw if date_span[1] >= x['dt'] >= date_span[0]]
ma_ = [x['ma250'] for x in ka.ma if date_span[1] >= x['dt'] >= date_span[0]]
num = cross_number(low, ma_)
res.append({"symbol": symbol, "cross_num": num})
except:
print("{} 分析失败".format(symbol))
traceback.print_exc()
return res
def test_ka_update():
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
kline.loc[:, "dt"] = pd.to_datetime(kline.dt)
kline1 = kline.iloc[:2000]
kline2 = kline.iloc[2000:]
ka1 = KlineAnalyze(kline, name="日线", max_raw_len=5000, verbose=False)
ka2 = KlineAnalyze(kline1, name="日线", max_raw_len=5000, verbose=False)
for _, row in kline2.iterrows():
ka2.update(row.to_dict())
assert len(ka1.kline_new) == len(ka2.kline_new)
assert len(ka1.fx_list) == len(ka2.fx_list)
assert len(ka1.bi_list) == len(ka2.bi_list)
def test_update_ta():
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
kline.loc[:, "dt"] = pd.to_datetime(kline.dt)
ka = KlineAnalyze(kline, name="日线", max_xd_len=10, verbose=False)
ma_x1 = dict(ka.ma[-1])
macd_x1 = dict(ka.macd[-1])
ka.update(kline.iloc[-1].to_dict())
ma_x2 = dict(ka.ma[-1])
macd_x2 = dict(ka.macd[-1])
assert ma_x1['dt'] == ma_x2['dt']
assert [round(x, 2) for x in ma_x1.values() if isinstance(x, float)] == \
[round(x, 2) for x in ma_x2.values() if isinstance(x, float)]
assert macd_x1['dt'] == macd_x2['dt']
assert [round(x, 2) for x in macd_x1.values() if isinstance(x, float)] == \
[round(x, 2) for x in macd_x2.values() if isinstance(x, float)]
def test_calculate_power():
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
kline.loc[:, "dt"] = pd.to_datetime(kline.dt)
ka = KlineAnalyze(kline, name="日线", max_raw_len=5000, verbose=False)
# 测试 macd 力度
last_xd_power = ka.calculate_macd_power(start_dt=ka.xd_list[-2]['dt'], end_dt=ka.xd_list[-1]['dt'],
mode='xd', direction="up" if ka.xd_list[-1]['fx_mark'] == 'g' else "down")
last_bi_power = ka.calculate_macd_power(start_dt=ka.bi_list[-2]['dt'], end_dt=ka.bi_list[-1]['dt'], mode='bi')
assert int(last_xd_power) == 389
assert int(last_bi_power) == 300
# 测试 vol 力度
last_xd_power = ka.calculate_vol_power(start_dt=ka.xd_list[-2]['dt'], end_dt=ka.xd_list[-1]['dt'])
last_bi_power = ka.calculate_vol_power(start_dt=ka.bi_list[-2]['dt'], end_dt=ka.bi_list[-1]['dt'])
assert int(last_xd_power) == 13329239053
assert int(last_bi_power) == 9291793337
def test_kline_analyze():
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
kline.loc[:, "dt"] = pd.to_datetime(kline.dt)
ka = KlineAnalyze(kline,
name="日线",
max_count=1000,
use_xd=True,
verbose=False)
# 测试绘图
file_img = "kline.png"
ka.to_image(file_img, max_k_count=5000)
assert os.path.exists(file_img)
# 测试分型识别结果
assert ka.fx_list[-1]['fx_mark'] == 'g'
assert ka.fx_list[-5]['fx_mark'] == 'g'
# 测试笔识别结果
assert ka.bi_list[-1]['fx_mark'] == 'g'
assert ka.bi_list[-4]['fx_mark'] == 'd'
# 测试线段识别结果
assert ka.xd_list[-2]['fx_mark'] == 'g'
assert ka.xd_list[-3]['fx_mark'] == 'd'
# 测试增量更新
for x in [2890, 2910, 2783, 3120]:
k = dict(ka.kline_raw[-1])
k['close'] = x
ka.update(k)
assert ka.kline_raw[-1]['close'] == x
def test_ka_update():
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
kline.loc[:, "dt"] = pd.to_datetime(kline.dt)
kline1 = kline.iloc[:2000]
kline2 = kline.iloc[2000:]
ka1 = KlineAnalyze(kline,
name="日线",
max_count=1000,
use_xd=True,
verbose=False)
ka2 = KlineAnalyze(kline1,
name="日线",
max_count=1000,
use_xd=True,
verbose=False)
for _, row in kline2.iterrows():
ka2.update(row.to_dict())
assert ka1.kline_new[-1]['dt'] == ka2.kline_new[-1]['dt']
assert ka1.fx_list[-1]['dt'] == ka2.fx_list[-1]['dt']
assert ka1.bi_list[-1]['dt'] == ka2.bi_list[-1]['dt']
assert ka1.xd_list[-1]['dt'] == ka2.xd_list[-1]['dt']
ka3 = KlineAnalyze(kline,
name="日线",
max_count=1000,
use_xd=False,
verbose=False)
assert ka3.kline_new[-1]['dt'] == ka2.kline_new[-1]['dt']
assert ka3.fx_list[-1]['dt'] == ka2.fx_list[-1]['dt']
assert ka3.bi_list[-1]['dt'] == ka2.bi_list[-1]['dt']
assert not ka3.xd_list
def test_get_sub_section():
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
kline.loc[:, "dt"] = pd.to_datetime(kline.dt)
ka = KlineAnalyze(kline, name="日线", max_xd_len=10, verbose=False)
sub_kn = ka.get_sub_section(ka.fx_list[-2]['dt'],
ka.fx_list[-1]['dt'],
mode='kn',
is_last=True)
assert sub_kn[0]['dt'] == ka.fx_list[-2]['dt'] and sub_kn[-1][
'dt'] == ka.fx_list[-1]['dt']
sub_fx = ka.get_sub_section(ka.bi_list[-2]['dt'],
ka.bi_list[-1]['dt'],
mode='fx',
is_last=True)
assert sub_fx[0]['dt'] == ka.bi_list[-2]['dt'] and sub_fx[-1][
'dt'] == ka.bi_list[-1]['dt']
sub_bi = ka.get_sub_section(ka.xd_list[-2]['dt'],
ka.xd_list[-1]['dt'],
mode='bi',
is_last=True)
assert sub_bi[0]['dt'] == ka.xd_list[-2]['dt'] and sub_bi[-1][
'dt'] == ka.xd_list[-1]['dt']
sub_xd = ka.get_sub_section(ka.xd_list[-4]['dt'],
ka.xd_list[-1]['dt'],
mode='xd',
is_last=True)
assert sub_xd[0]['dt'] == ka.xd_list[-4]['dt'] and sub_xd[-1][
'dt'] == ka.xd_list[-1]['dt']
def singal_selector(symbols: List):
res = []
for symbol in symbols:
try:
kline = get_kline(symbol=symbol, end_date=datetime.now(), freq="30min", count=1000)
ka = KlineAnalyze(kline, ma_params=(5, 34, 60, 250), bi_mode="new")
if ka.ma[-1]['ma60'] >= ka.latest_price >= ka.ma[-1]['ma250']:
type = check_bei_chi(ka.xd_list[-5], ka.xd_list[-4], ka.xd_list[-3], ka.xd_list[-2], ka.xd_list[-1]).get("bc")
if type in ["向下趋势背驰", "向下盘整背驰"]:
res.append({"symbol": symbol, "beici": num})
except:
print("{} 分析失败".format(symbol))
traceback.print_exc()
return res
def test_kline_pro():
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
bars = kline.to_dict("records")
ka = KlineAnalyze(bars)
bs = []
for x in ka.xd_list:
if x['fx_mark'] == 'd':
mark = "buy"
else:
mark = "sell"
bs.append({"dt": x['dt'], "mark": mark, mark: x['xd']})
chart = plot.kline_pro(ka.kline_raw, fx=ka.fx_list, bi=ka.bi_list, xd=ka.xd_list, bs=bs)
chart.render()
def selector(symbols: List):
"""输入股票列表，输入符合买点定义的股票"""
res = []
for symbol in tqdm(symbols, desc="缠论日线笔中枢三买选股"):
try:
kline = get_kline(symbol=symbol, end_date=datetime.now(), freq="D", count=1000)
ka = KlineAnalyze(kline, ma_params=(5, 34, 120, 233), bi_mode="new")
points = ka.bi_list[-6:]
if len(points) == 6 and points[-1]['fx_mark'] == "d":
zs_g = min([x['bi'] for x in points[:4] if x['fx_mark'] == 'g'])
zs_d = max([x['bi'] for x in points[:4] if x['fx_mark'] == 'd'])
if points[-1]['bi'] > zs_g > zs_d:
res.append(symbol)
except:
print("{} 分析失败".format(symbol))
traceback.print_exc()
return res
def use_large_df():
symbol = "*****@*****.**"
freq = '5min'
file_csv = f"{symbol}_kline_{freq}.csv"
start_dt = datetime(2017, 1, 1, 6, 0, 0)
end_dt = datetime(2020, 5, 1, 6, 0, 0)
freq_dur_sec = {"1min": 60, '5min': 300, '30min': 1800, 'D': 3600 * 24}
freq_delta = {
"1min": timedelta(days=20),
'5min': timedelta(days=100),
'30min': timedelta(days=300),
'D': timedelta(days=3000)
}
api = TqApi()
k = DataDownloader(api,
symbol_list=symbol,
dur_sec=freq_dur_sec[freq],
start_dt=start_dt - freq_delta[freq],
end_dt=end_dt,
csv_file_name=file_csv)
with closing(api):
while not k.is_finished():
api.wait_update()
print("download progress: %.2f%%" % k.get_progress())
kline = pd.read_csv(file_csv)
kline.columns = [x.replace(symbol + ".", "") for x in kline.columns]
kline.rename({"volume": "vol"}, axis=1, inplace=True)
kline.loc[:, "symbol"] = symbol
kline.loc[:, "dt"] = kline['datetime'].apply(lambda x: x.split(".")[0])
kline = kline[['symbol', 'dt', 'open', 'close', 'high', 'low', 'vol']]
print(kline.shape)
ka = KlineAnalyze(kline)
return ka
sys.path.insert(0, '..')
import os
import pandas as pd
import czsc
from czsc.analyze import KlineAnalyze, find_zs
warnings.warn("czsc version is {}".format(czsc.__version__))
cur_path = os.path.split(os.path.realpath(__file__))[0]
# cur_path = "./test"
file_kline = os.path.join(cur_path, "data/000001.SH_D.csv")
kline = pd.read_csv(file_kline, encoding="utf-8")
kline.loc[:, "dt"] = pd.to_datetime(kline.dt)
kline1 = kline.iloc[:2000]
kline2 = kline.iloc[2000:]
ka = KlineAnalyze(kline1, name="日线", max_raw_len=2000, verbose=True)
def test_update():
for _, row in kline2.iterrows():
ka.update(row.to_dict())
assert ka.kline_raw[-1]['dt'] == row['dt']
def test_update_ta():
ma_x1 = dict(ka.ma[-1])
macd_x1 = dict(ka.macd[-1])
ka.update(kline.iloc[-1].to_dict())
ma_x2 = dict(ka.ma[-1])
macd_x2 = dict(ka.macd[-1])
assert ma_x1['dt'] == ma_x2['dt']