def evaluate(self, db, args):
dates_expr = f'trade_dates.{args.date}.2.lag'
ma_l = stocklab.evaluate(
f'moving_average.{args.stock_id}.(${dates_expr}).8')
ma_s = stocklab.evaluate(
f'moving_average.{args.stock_id}.(${dates_expr}).3')
ml = ma_l[1] - ma_l[0]
ms = ma_s[1] - ma_s[0]
cross = (ma_l[1] - ma_s[1]) * (ma_l[0] - ma_s[0]) <= 0
bear = ml > ms
bull = ml < ms
return (0 if not cross else 1 if bull else -1 if bear else 0)
def evaluate(self, db, args):
n = args.n
days = args.days + n
prices = np.asarray(stocklab.evaluate(
f'ohlc.{args.stock_id}.($trade_dates.{args.date}.{days}.lag)'
))[:,3]
deltas = np.diff(prices)
seed = deltas[:n - 1]
up = seed[seed > 0].sum() / n
down = -seed[seed < 0].sum() / n
rs = up / down
rsi = np.zeros_like(prices)
rsi[:n] = 100. - 100. / (1. + rs)
for i in range(n, len(prices)):
delta = deltas[i - 1] # cause the diff is 1 shorter
if delta > 0:
upval = delta
downval = 0.
else:
upval = 0.
downval = -delta
up = (up * (n - 1) + upval) / n
down = (down * (n - 1) + downval) / n
rs = up / down
rsi[i] = 100. - 100. / (1. + rs)
return rsi[n:]
def evaluate(self, db, args):
brokers = stocklab.evaluate(
f'stock_brokers.{args.stock_id}._.{args.date}.{args.n_days}')
force_list = np.asarray(
[b['buy_amt'] - b['sell_amt'] for bid, b in brokers.items()])
if 'buy' == args.buy_or_sell:
return force_list[force_list > 0]
else:
return -force_list[force_list < 0]
def evaluate(self, db, args):
bsf = stocklab.evaluate(
f'tendency.{args.mode}.{args.stock_id}.{args.date}.{args.days}')
bsf = bsf[bsf > args.min_force]
def normalize(v):
norm = np.linalg.norm(v)
if norm == 0:
return v
return v / norm
return normalize(bsf).var() if len(bsf) else 0.0
def _eval(self, path, peek=False):
use_cache = 'disable_cache' not in self.spec or not self.spec['disable_cache']
if path in self.cache and use_cache and not peek:
return self.cache[path]
assert not re.search(r'\(((?!\)).)*?\(', path), f'expression cannot have nested parentheses: {path}'
splited = [e.strip('()') for e in re.split(r'(\(.*?\)|\.)', path) if e and e != '.']
mod_name = splited[0]
assert self.name == mod_name, f'{splited}'
arg_spec = self.spec['args']
assert self.args_validator(arg_spec)
assert len(arg_spec) == len(splited) - 1, f'invalid expression: {path}'
from collections.abc import Iterable
_replace = lambda t, i, elem: t[:i] + [elem] + t[i+1:]
for i in range(len(splited)):
field = splited[i]
try:
float(field)
continue
except ValueError:
pass
if '$' != field[0]:
continue
values = stocklab.metaevaluate(field.lstrip('$'))
assert isinstance(values, Iterable)
return [
stocklab.evaluate('.'.join(
[f'({s})' for s in _replace(splited, i, str(val))]
))
for val in values]
args_list = tuple(splited[1:])
args = stocklab.Args(args_list, arg_spec)
result = self._eval_in_context(args, peek=peek)
if not peek:
assert self.type_validator(result)
if use_cache:
self.cache[path] = result
return result
rand_sim = random_simulate(
date,
days,
verbose=True,
dates_expr=lambda d, n: f'trade_dates.{d}.{n}.lag',
ohlc_expr=lambda s, d: f'ohlc.{s}.{d}',
sign_expr=lambda s, d: f'my_sign.{s}.{d}',
#stocks_expr='stocks.1702'
)
stock_id, gain = next(rand_sim)
print(f'stock_id: {stock_id}, gain: {gain:.3}')
# plot it
dates_expr = f'trade_dates.{date}.{days}.lag'
dates = stocklab.metaevaluate(dates_expr)
ohlcs = np.asarray(stocklab.evaluate(f'ohlc.{stock_id}.(${dates_expr})'))
prices = ohlcs[:, 3]
signs = stocklab.evaluate(f'my_sign.{stock_id}.(${dates_expr})')
plot(dates,
ohlcs,
signs,
aux={
'ma_l':
stocklab.evaluate(f'moving_average.{stock_id}.(${dates_expr}).8'),
'ma_s':
stocklab.evaluate(f'moving_average.{stock_id}.(${dates_expr}).3')
},
top_aux={
'volumn': stocklab.evaluate(f'volumn.{stock_id}.(${dates_expr})'),
})
def evaluate(self, db, args):
return np.asarray(
stocklab.evaluate(
f'ohlc.{args.stock_id}.($trade_dates.{args.date}.{args.N}.lag)'
))[:, 3].mean()
def _eval(expr):
if self.meta:
return stocklab.metaevaluate(str(expr))
else:
return stocklab.evaluate(str(expr))
import numpy as np
import logging
import stocklab
from stocklab.datetime import Date
from stocklab.utils import *
#stocklab.change_log_level(logging.DEBUG)
date = stocklab.metaevaluate('nearest.20200424')
days = 100
stock_id = 3034
# plot it
dates_expr = f'trade_dates.{date}.{days}.lag'
dates = stocklab.metaevaluate(dates_expr)
ohlcs = np.asarray(stocklab.evaluate(f'ohlc.{stock_id}.(${dates_expr})'))
ma_l = np.asarray(
stocklab.evaluate(f'moving_average.{stock_id}.(${dates_expr}).8'))
ma_s = np.asarray(
stocklab.evaluate(f'moving_average.{stock_id}.(${dates_expr}).3'))
plot(dates,
ohlcs,
aux={
'ma_l': ma_l,
'ma_s': ma_s
},
top_aux={'macd': ma_l - ma_s})
def _sim(stock_id):
ohlcs = np.asarray(
stocklab.evaluate(ohlc_expr(stock_id, f'(${dates_expr})')))
prices = ohlcs[:, 3]
signs = stocklab.evaluate(sign_expr(stock_id, f'(${dates_expr})'))
return simulate(dates, signs, prices, verbose)