def pred(self, daybefore, sample_size, sign_name, cache_size=1024):
sign_num = alphatree.getSignNum(daybefore, sample_size,
c_char_p(sign_name))
alpha_cache = (c_float * sign_num)()
alphatree.predAlphaGBDF(daybefore, sample_size, c_char_p(sign_name),
alpha_cache, cache_size)
return np.array([alpha_cache[i] for i in range(sign_num)])
def get_trend(self, sign_name, elements, daybefore, sample_size, column,
max_std_scale):
element_cache = self.decode_features(elements)
pic_size = (column * len(elements) * 3 * alphatree.getSignNum(
daybefore, sample_size, c_char_p(sign_name.encode('utf-8'))))
pic_cache = (c_float * pic_size)()
alphatree.getTrendPic(self.picid, c_char_p(sign_name.encode('utf-8')),
element_cache, len(elements), daybefore,
sample_size, pic_cache, column,
c_float(max_std_scale))
# trend = np.zeros((pic_size,))
# for i in range(pic_size):
# trend[i] = pic_cache[i]
trend = np.array(pic_cache)
return trend.reshape([-1, column, len(elements) * 3, 1])
def pred(self, sign_name, daybefore, sample_size):
sign_num = alphatree.getSignNum(daybefore, sample_size,
c_char_p(sign_name.encode()))
alpha_cache = (c_float * sign_num)()
min_open_value_cache = (c_float *
sign_num)() if self.is_open_buy else None
max_open_value_cache = (c_float *
sign_num)() if self.is_open_buy else None
alphatree.predAlphaFilter(c_char_p(sign_name.encode()), daybefore,
sample_size, alpha_cache,
min_open_value_cache, max_open_value_cache)
#res = np.array([alpha_cache[i] for i in range(sign_num)])
if self.is_open_buy:
return np.array(alpha_cache), np.array(
min_open_value_cache), np.array(max_open_value_cache)
return np.array(alpha_cache)
def get_k_line(self, sign_name, open_elements, high_elements, low_elements,
close_elements, daybefore, sample_size, column,
max_std_scale):
open_cache = self.decode_features(open_elements)
high_cache = self.decode_features(high_elements)
low_cache = self.decode_features(low_elements)
close_cache = self.decode_features(close_elements)
pic_size = (column * len(open_elements) * 3 * alphatree.getSignNum(
daybefore, sample_size, c_char_p(sign_name.encode('utf-8'))))
pic_cache = (c_float * pic_size)()
alphatree.getKLinePic(self.picid, c_char_p(sign_name.encode('utf-8')),
open_cache, high_cache, low_cache, close_cache,
len(open_elements), daybefore, sample_size,
pic_cache, column, c_float(max_std_scale))
# kline = np.zeros((pic_size,))
# for i in range(pic_size):
# kline[i] = pic_cache[i]
kline = np.array(pic_cache)
return kline.reshape([-1, column, len(open_elements) * 3, 1])
def testRead(self,
daybefore,
sample_size,
feature_num,
target,
sign_name,
cache_size=2048):
sign_num = alphatree.getSignNum(daybefore, sample_size,
c_char_p(sign_name))
target_cache = (c_float * sign_num)()
feature_cache = (c_float * (sign_num * feature_num))()
data_num = alphatree.testReadGBDT(daybefore, sample_size,
c_char_p(target),
c_char_p(sign_name), feature_cache,
target_cache, cache_size)
assert data_num == sign_num
t = [target_cache[i] for i in range(sign_num)]
f = []
for i in range(sign_num):
for j in range(feature_num):
f.append(feature_cache[i * feature_num + j])
#f = [[feature_cache[i * feature_num + j] for j in range(feature_num)] for i in range(sign_num)]
return np.array(f), np.array(t)