def vote(self, code, day):
features = []
values = db_helper.select(query_kpi_fn, (code, day, 30))
if len(values) < 30:
return features
values = np.array(values)[::-1]
rsi_6_lines = values[:, 1].astype(float)
rsi_12_lines = values[:, 2].astype(float)
rsi_24_lines = values[:, 3].astype(float)
if rsi_6_lines[-1] <= 20:
features.append(self.OVER_SOLD)
if rsi_6_lines[-1] < 50 and rsi_12_lines[-1] < 50 and rsi_24_lines[
-1] < 50:
if ins.match(rsi_6_lines[-3:],
rsi_12_lines[-3:]) == 1 and ins.match(
rsi_6_lines[-3:], rsi_24_lines[-3:]):
features.append(self.GOOD_CROSS)
if trend.is_uptrend(rsi_6_lines[-10:]):
k_values = db_helper.select(query_k_fn, (code, day, 10))
if len(k_values) == 10:
k_values = np.array(k_values)[::-1]
close_values = k_values[:, 0].astype(float)
if trend.is_downtrend(close_values):
features.append(self.BOTTOM_DEVIATION)
return features
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) == 0:
return
data = np.array(data)
close = np.float(data[0][1])
last_ax, last_bx, last_trix = None, None, None
last_data = db_helper.select(query_trix_fn, (code, day, 1))
if len(last_data) == 1:
last_ax = np.float(last_data[0][0])
last_bx = np.float(last_data[0][1])
last_trix = np.float(last_data[0][2])
ax, bx, tx = _calc_trix(last_ax, last_bx, last_trix, close)
trix = TRIX()
trix.code = code
trix.time_date = day
trix.ax = ax
trix.bx = bx
trix.trix = tx
last_tma_list = db_helper.select(query_trix_fn, (code, day, TRIX_M - 1))
last_tma_list = np.array(last_tma_list)
last_tma_list = last_tma_list[:, 2]
trix.tma = _get_tma(last_tma_list, tx)
return trix
def his(code, day, day_len, forward):
out = PrettyTable(['交易日', '收盘价', '涨跌幅', '评分', '特征'])
query_data = db_helper.select(
query_data_his_fn, (code, day, day_len, TradingDataDaily, forward))
query_data = np.array(query_data)
close_values = np.array(list(map(lambda x: np.float(x.close), query_data)))
open_values = np.array(list(map(lambda x: np.float(x.open), query_data)))
days = np.array(list(map(lambda x: x.time_date, query_data)))
query_data = db_helper.select(query_eval_his_fn,
(code, day, day_len, forward))
query_data = np.array(query_data)
days2 = np.array(list(map(lambda x: x.time_date, query_data)))
score_values = np.array(list(map(lambda x: np.float(x.score), query_data)))
feature_values = np.array(list(map(lambda x: x.feature, query_data)))
for i in range(len(days)):
d = days[i]
close = close_values[i]
open_price = open_values[i]
delta = np.float((close - open_price) / open_price * 100)
index = -1
for j in range(len(days2)):
if days[i] == days2[j]:
index = j
score, feature = 0, ''
if index >= 0:
score = score_values[index]
feature = feature_values[index]
out.add_row([d, _f(close), _f(delta, '%.2f%%'), _f(score), feature])
print(out)
pass
def vote(self, code, day):
features = []
values = db_helper.select(query_kdj_fn, (code, day, 30))
if len(values) < 30:
return features
values = np.array(values)[::-1]
k_values = values[:, 1].astype(float)
d_values = values[:, 2].astype(float)
j_values = values[:, 3].astype(float)
if k_values[-1] <= 20 and d_values[-1] <= 20 and j_values[-1] <= 20:
features.append(self.OVER_SOLD)
if ins.match(k_values[-3:], d_values[-3:]) == 1 and ins.match(
j_values[-3:], d_values[-3:]):
features.append(self.GOOD_CROSS)
s = 0
if trend.is_uptrend(k_values[-10:]):
s = s + 1
if trend.is_uptrend(d_values[-10:]):
s = s + 1
if trend.is_uptrend(j_values[-10:]):
s = s + 1
if s >= 2:
k_values = db_helper.select(query_k_fn, (code, day, 10))
k_values = np.array(k_values)[::-1]
close_values = k_values[:, 0].astype(float)
if trend.is_downtrend(close_values):
features.append(self.BOTTOM_DEVIATION)
return features
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < KDJ_N:
return None
last_kdj = db_helper.select(query_kdj_fn, (code, day))
last_k, last_d = 50, 50
if len(last_kdj) == 1:
last_kdj = np.array(last_kdj)
last_k, last_d = last_kdj[0][0], last_kdj[0][1]
return _calc_kdj(code, data, last_k, last_d)
def pick():
time_date = request.args.get('time_date')
pl = request.args.get('type')
data = None
if pl == 'SCORE':
data = db_helper.select(query_data_fn, (time_date, ))
else:
data = db_helper.select(query_pl_data_fn, (time_date, pl))
pass
return jsonify({'list': data})
pass
def trace(vote_day, count=10, days=5):
# str = vote_day[0:4] + vote_day[5:7] + vote_day[8:10]
# file_name = DATA_PATH + 'R_%s.txt' % str
# stock_codes = []
# with open(file_name, 'r') as file_to_read:
# while True:
# lines = file_to_read.readline() # 整行读取数据
# if not lines:
# break
# stock_codes.append(lines.split()[0])
query_data = db_helper.select(query_eval_fn, (vote_day, count))
query_data = np.array(query_data)
stock_codes = query_data[:, 0]
scores = query_data[:, 1].astype(float)
features = query_data[:, 2]
# print(stock_codes)
out = PrettyTable([
'股票代码', '评分', '当日收盘价', '次日涨跌幅', '期内最高价', '期内最低价', '期后收盘价', '期内最大涨幅',
'期内最大跌幅', '期后涨跌幅', '特征'
])
for i in range(len(stock_codes)):
values = db_helper.select(
query_data_fn,
(stock_codes[i], vote_day, days + 1, TradingDataDaily))
values = np.array(values)
close_values = np.array(list(map(lambda x: np.float(x.close), values)))
high_values = np.array(list(map(lambda x: np.float(x.high), values)))
low_values = np.array(list(map(lambda x: np.float(x.low), values)))
close = np.float(close_values[0])
delta_next = np.float(
(np.float(close_values[1]) - close) / close * 100)
high = np.float(np.max(high_values[2:]))
low = np.float(np.max(low_values[2:]))
delta_high = np.float((high - close) / close * 100)
last = np.float(close_values[-1])
delta_low = np.max([np.float((close - low) / close * 100), 0])
delta_close = np.float((last - close) / close * 100)
out.add_row([
stock_codes[i], scores[i],
_f(close),
_f(delta_next, '%.2f%%'),
_f(high),
_f(low),
_f(last),
_f(delta_high, '%.2f%%'),
_f1(delta_low),
_f(delta_close, '%.2f%%'), features[i]
])
print('跟踪%s评估的%d只股票%d日内的价格情况' % (vote_day, count, days))
print(out)
pass
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < KDJ_N:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
last_kdj = db_helper.select(query_kdj_fn, (code, day))
last_k, last_d = 50, 50
if len(last_kdj) == 1:
last_kdj = np.array(last_kdj)
last_k, last_d = np.float(last_kdj[0][0]), np.float(last_kdj[0][1])
return _calc_kdj(code, data, last_k, last_d)
def eva(vote_day, count=10, days=5):
query_data = db_helper.select(query_eval_fn, (vote_day, count))
if len(query_data) == 0:
elect(vote_day, count)
query_data = db_helper.select(query_eval_fn, (vote_day, count))
query_data = np.array(query_data)
stock_codes = query_data[:, 0]
scores = query_data[:, 1].astype(float)
features = query_data[:, 2]
# print(stock_codes)
out = PrettyTable([
'股票代码', '评分', '当日收盘价', '次日涨跌幅', '最高价', '最低价', '最后收盘价', '最大涨幅', '最大跌幅',
'后涨跌幅', '特征'
])
for i in range(len(stock_codes)):
values = db_helper.select(
query_data_fn,
(stock_codes[i], vote_day, days + 1, TradingDataDaily))
values = np.array(values)
close_values = np.array(list(map(lambda x: np.float(x.close), values)))
high_values = np.array(list(map(lambda x: np.float(x.high), values)))
low_values = np.array(list(map(lambda x: np.float(x.low), values)))
close = np.float(close_values[0])
delta_next, high, low, delta_high, last, delta_low, delta_close = 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0
if len(values) > 1:
delta_next = np.float(
(np.float(close_values[1]) - close) / close * 100)
last = np.float(close_values[-1])
delta_close = np.float((last - close) / close * 100)
high = np.float(np.max(high_values[1:]))
low = np.float(np.max(low_values[1:]))
delta_high = np.float((high - close) / close * 100)
delta_low = np.max([np.float((close - low) / close * 100), 0])
out.add_row([
stock_codes[i], scores[i],
_f(close),
_f(delta_next, '%.2f%%'),
_f(high),
_f(low),
_f(last),
_f(delta_high, '%.2f%%'),
_f1(delta_low),
_f(delta_close, '%.2f%%'), features[i]
])
print('跟踪%s评估的%d只股票%d日内的价格情况' % (vote_day, count, days))
print(out)
pass
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < 2:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
his = db_helper.select(query_emv_fn, (code, day))
his = np.array(his)
his = his[::-1]
em_his = np.array(his[1 - EMV_N:, 0].astype(float))
emv_his = np.array(his[1 - EMV_M:, 0].astype(float))
emv, em_his, emv_his = _emv(code, data, em_his, emv_his)
return emv
def detail():
code = request.args.get('code')
print(code)
data = db_helper.select(query_detail_data_fn, (code, ))
data = np.array(data)
# data = [['2018-11-11',5.5,'aaa']]
text = htmle_head()
text = text + "<h5>%s</h5>" % code
text = text + """
<div>
<ul>
"""
for i in range(len(data)):
text = text + """
<li>
<a>%d</a>
<h3>%s</h3>
<p>%s</p>
</li>
""" % (data[i][1], data[i][0], data[i][2])
text = text + """
</ul>
</div>
</body>
</html>
"""
return text
pass
def _get_close_price(self, code, start_day, end_day):
data = db_helper.select(
query_single_fn,
(code, start_day, end_day, TradingDataDaily.close))
if len(data) == 0:
return None
else:
return np.array(data)[:, 0]
def dig(step=30, slide_days=10, turn_days=2):
day = time.strftime("%Y-%m-%d", time.localtime())
code_list = db_helper.select(query_code_fn, (day, ))
rst = []
for i in range(len(code_list)):
code = code_list[i][0]
is_rebound, ma_values = _is_rebound(code, step, slide_days, turn_days)
print("分析[%s,%s]:%r" % ('ma', code, is_rebound))
if is_rebound:
data = db_helper.select(query_data_fn, (code, day, turn_days))
data = np.array(data)
data = data[:, 1]
ma_values = ma_values[-turn_days:]
is_less = np.min(np.less(ma_values, data))
if is_less:
rst.append(code)
return rst
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < ARBR_N + 1:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
return _arbr(code, data)
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < BIAS_N4:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
rsi = _bias(code, data)
return rsi
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < DMA_N + DMA_M - 1:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
dma = _dma(code, data)
return dma
def vote(self, code, day):
features = []
values = db_helper.select(query_wr_fn, (code, day, 30))
if len(values) < 30:
return features
values = np.array(values)[::-1]
wr_6_values = values[:, 1].astype(float)
if trend.is_uptrend(wr_6_values[-10:]):
k_values = db_helper.select(query_k_fn, (code, day, 10))
if len(k_values) == 10:
k_values = np.array(k_values)[::-1]
close_values = k_values[:, 0].astype(float)
if trend.is_downtrend(close_values):
features.append(self.BOTTOM_DEVIATION)
if wr_6_values[-1] > 80:
features.append(self.OVER_SOLD)
return features
def _init_all(code):
data = db_helper.select(query_data_fn, (code, None))
if len(data) < ARBR_N + 1:
return None
data = np.array(data)
arbr_list = []
for i in range(len(data) - ARBR_N - 1):
last_data = data[i:i + ARBR_N + 1]
arbr_list.append(_arbr(code, last_data))
return arbr_list
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < BOLL_N + 1:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
boll = _boll(code, data)
return boll
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < ROC_N + ROC_M - 1:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
roc = _roc(code, data)
return roc
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < WR_40:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
entity = _wr(code, data)
return entity
def vote(self, code, day):
features = []
values = db_helper.select(self.query_kpi_fn, (code, day, 30, BOLL))
k_values = db_helper.select(self.query_kpi_fn,
(code, day, 30, TradingDataDaily))
if len(values) < 30 or len(k_values) < 30 or len(values) != len(
k_values):
return features
values = np.array(values)[::-1]
k_values = np.array(k_values)[::-1]
md_values = np.array(list(map(lambda x: np.float(x.md), values)))
close_values = np.array(
list(map(lambda x: np.float(x.close), k_values)))
bt, l = zero_line.breakthrough(close_values, md_values)
if bt and l <= 3:
features.append(self.BREAK_ZERO)
return features
def _init_day(code, day):
data = db_helper.select(query_data_fn, (code, day))
if len(data) < RSI_N3 + 1:
return None
data = np.array(data)
if data[0][0] != day:
return None
data = data[::-1]
rsi = _rsi(code, data)
return rsi
def _init_all(code):
data = db_helper.select(query_data_fn, (code, None))
if len(data) < RSI_N3 + 1:
return None
data = np.array(data)
rsi_list = []
for i in range(len(data) - RSI_N3):
last_data = data[i:i + (RSI_N3 + 1)]
rsi = _rsi(code, last_data)
rsi_list.append(rsi)
return rsi_list
def _init_all(code):
data = db_helper.select(query_data_fn, (code, None))
if len(data) < BOLL_N + 1:
return None
data = np.array(data)
boll_list = []
for i in range(len(data) - BOLL_N):
last_data = data[i:i + (BOLL_N + 1)]
boll = _boll(code, last_data)
boll_list.append(boll)
return boll_list
def vote(self, code, day):
features = []
arbr_values = db_helper.select(query_arbr_fn, (code, day, 14))
if len(arbr_values) < 14:
return features
arbr_values = np.array(arbr_values)[::-1]
ar_values = arbr_values[:, 1].astype(float)
br_values = arbr_values[:, 2].astype(float)
if ins.match(br_values[-3:], ar_values[-3:]) == 1:
features.append(self.GOOD_CROSS)
if trend.is_uptrend(ar_values) and trend.is_uptrend(br_values):
k_values = db_helper.select(query_k_fn, (code, day, 14))
if len(k_values) == 14:
k_values = np.array(k_values)[::-1]
close_values = k_values[:, 0].astype(float)
if trend.is_downtrend(close_values):
features.append(self.BOTTOM_DEVIATION)
return features
def _init_all(code):
data = db_helper.select(query_data_fn, (code, None))
if len(data) < WR_40:
return None
data = np.array(data)
entity_list = []
for i in range(len(data) - WR_40 + 1):
last_data = data[i:i + WR_40]
entity = _wr(code, last_data)
entity_list.append(entity)
return entity_list
def _init_all(code):
data = db_helper.select(query_data_fn, (code, None))
if len(data) < BIAS_N4:
return None
data = np.array(data)
rsi_list = []
for i in range(len(data) - BIAS_N4 + 1):
last_data = data[i:i + BIAS_N4]
rsi = _bias(code, last_data)
rsi_list.append(rsi)
return rsi_list
def eye():
time_date = db_helper.select(query_date, (1))[0][0]
data = db_helper.select(query_data_fn, (time_date, ))
data = np.array(data)
text = htmle_head()
text = text + "<h5>%s</h5>" % time_date
text = text + """
<div>
<ul>
"""
for i in range(len(data)):
text = text + build_html_with_stock_data(data[i][0], data[i][1],
data[i][2])
text = text + """
</ul>
</div>
</body>
</html>
"""
return text
def _init_all(code):
data = db_helper.select(query_data_fn, (code, None))
if len(data) < DMA_N + DMA_M - 1:
return None
data = np.array(data)
roc_list = []
for i in range(len(data) - (DMA_N + DMA_M - 1)):
last_data = data[i:i + (DMA_N + DMA_M)]
roc = _dma(code, last_data)
roc_list.append(roc)
return roc_list
