def do_action(self, action, pre_portfolio_value, only_update):
sell_price, buy_price = self.get_action_price(action)
sell_prices, buy_prices = [sell_price], [buy_price]
if only_update:
sell_prices, buy_prices = [0] * self.n, [0] * self.n
divide_rate = self.market.get_divide_rate(self.code, self.current_date)
logger.debug("divide_rate: %.4f" % divide_rate)
self.portfolio.update_before_trade(divide_rate)
cash_change = 0
if not only_update:
# 全仓卖出
sell_cash_change, ok = self.sell(0, sell_price, 0.0)
# 全仓买进
buy_cash_change, ok = self.buy(0, buy_price, 1.0)
cash_change = buy_cash_change + sell_cash_change
logger.debug("do_action: time_id: %d, %s, cash_change: %.1f" %
(self.current_time_id, self.code, cash_change))
close_price = self.market.get_close_price(self.code, self.current_date)
self.portfolio.update_after_trade(
close_price=close_price,
cash_change=cash_change,
pre_portfolio_value=pre_portfolio_value)
return sell_prices, buy_prices
def sell(self, id, price, target_pct):
# id: code id
code = self.codes[id]
logger.debug("sell %s, bid price: %.2f" % (code, price))
ok, price = self.market.sell_check(code=code,
datestr=self.current_date,
bid_price=price)
if ok:
# 全仓卖出
cash_change, price, vol = self.portfolios[id].order_target_percent(
percent=target_pct,
price=price,
pre_portfolio_value=self.portfolio_value,
current_cash=self.cash)
self.cash += cash_change
if vol != 0:
self.info["orders"].append([
"sell", code,
round(cash_change, 1),
round(price, 2), vol
])
logger.debug("sell %s target_percent: 0, cash_change: %.3f" %
(code, cash_change))
return cash_change, ok
return 0, ok
def buy(self, id, price, target_pct):
# id: code id
code = self.codes[id]
logger.debug("buy %s, bid_price: %.2f" % (code, price))
ok, price = self.market.buy_check(code=code,
datestr=self.current_date,
bid_price=price)
pre_cash = self.cash
if ok:
# 分仓买进
cash_change, price, vol = self.portfolios[id].order_target_percent(
percent=target_pct,
price=price,
pre_portfolio_value=self.portfolio_value,
current_cash=self.cash)
self.cash += cash_change
if vol != 0:
self.info["orders"].append(
["buy", code,
round(cash_change, 1),
round(price, 2), vol])
logger.debug("buy %s cash: %.1f, cash_change: %1.f" %
(code, pre_cash, cash_change))
return cash_change, ok
return 0, ok
def get_action_price(self, v_price, code):
pre_close = self.market.get_pre_close_price(code, self.current_date)
logger.debug("%s %s pre_close: %.2f" %
(self.current_date, code, pre_close))
# scale [-1, 1] to [-0.1, 0.1]
pct = v_price * 0.1
price = round(pre_close * (1 + pct), 2)
return price
def get_action_price(self, action, code):
pre_close = self.market.get_pre_close_price(code, self.current_date)
logger.debug("%s %s pre_close: %.2f" %
(self.current_date, code, pre_close))
[v_sell, v_buy] = action
# scale [-1, 1] to [-0.1, 0.1]
pct_sell, pct_buy = v_sell * 0.1, v_buy * 0.1
sell_price = round(pre_close * (1 + pct_sell), 2)
buy_price = round(pre_close * (1 + pct_buy), 2)
return sell_price, buy_price
def do_action(self, action, pre_portfolio_value, only_update):
"""
only_update: 仅更新Portfolio, 不做操作, 即:buy_and_hold策略
"""
cash_change = 0
# 更新拆分信息
for i in range(self.n):
code = self.codes[i]
divide_rate = self.market.get_divide_rate(code, self.current_date)
self.portfolios[i].update_before_trade(divide_rate)
sell_prices, buy_prices = [], []
if only_update:
sell_prices, buy_prices = [0] * self.n, [0] * self.n
if not only_update:
# 卖出
for i in range(self.n):
code = self.codes[i]
act_i = action[4 * i:4 * (i + 1)]
sell_price = self.get_action_price(act_i[0], code)
sell_prices.append(sell_price)
target_pct = self.get_action_target_pct(act_i[1])
sell_cash_change, ok = self.sell(i, sell_price, target_pct)
cash_change += sell_cash_change
# 买进
for i in range(self.n):
code = self.codes[i]
act_i = action[4 * i:4 * (i + 1)]
buy_price = self.get_action_price(act_i[2], code)
buy_prices.append(buy_price)
target_pct = self.get_action_target_pct(act_i[3])
buy_cash_change, ok = self.buy(i, buy_price, target_pct)
cash_change += buy_cash_change
logger.debug("do_action: time_id: %d, cash_change: %.1f" %
(self.current_time_id, cash_change))
# update
for i in range(self.n):
code = self.codes[i]
close_price = self.market.get_close_price(code, self.current_date)
self.portfolios[i].update_after_trade(
close_price=close_price,
cash_change=cash_change,
pre_portfolio_value=pre_portfolio_value)
return sell_prices, buy_prices
def do_action(self, action, pre_portfolio_value, only_update):
cash_change = 0
# 更新拆分信息
for i in range(self.n):
code = self.codes[i]
divide_rate = self.market.get_divide_rate(code, self.current_date)
self.portfolios[i].update_before_trade(divide_rate)
sell_prices, buy_prices = [], []
if only_update:
sell_prices, buy_prices = [0] * self.n, [0] * self.n
if not only_update:
# 卖出
for i in range(self.n):
code = self.codes[i]
act_i = action[2 * i:2 * (i + 1)]
sell_price, _ = self.get_action_price(act_i, code)
sell_prices.append(sell_price)
sell_cash_change, ok = self.sell(i, sell_price, 0)
cash_change += sell_cash_change
# 买进
for i in range(self.n):
code = self.codes[i]
act_i = action[2 * i:2 * (i + 1)]
_, buy_price = self.get_action_price(act_i, code)
buy_prices.append(buy_price)
buy_cash_change, ok = self.buy(i, buy_price, self.avg_percent)
cash_change += buy_cash_change
logger.debug("do_action: time_id: %d, cash_change: %.1f" %
(self.current_time_id, cash_change))
# update
for i in range(self.n):
code = self.codes[i]
close_price = self.market.get_close_price(code, self.current_date)
self.portfolios[i].update_after_trade(
close_price=close_price,
cash_change=cash_change,
pre_portfolio_value=pre_portfolio_value)
return sell_prices, buy_prices
def buy_check(self, code='', datestr='', bid_price=None):
# 返回：OK, 成交价
ok = False
# 停牌
if self.is_suspended(code, datestr):
return ok, 0
# 获取当天标的信息
[open, high, low, pct_change
] = self.codes_history[code].loc[datestr,
["open", "high", "low", "pct_chg"]]
# 涨停封板, 无法买入
if low == high and pct_change > self.top_pct_change:
logger.debug(u"sell_check %s %s 涨停法买进" % (code, datestr))
return ok, 0
# 买入竞价低于最低价，不能成交
if bid_price < low:
return ok, 0
# 买入竞价高于最低价， 可以成交
if bid_price >= low:
return True, min(bid_price, high)
def step(self, action, only_update=False):
"""
only_update为True时，表示buy_and_hold策略，可作为一种baseline策略
"""
self.action = action
self.info = {"orders": []}
logger.debug("=" * 50 + "%s" % self.current_date + "=" * 50)
logger.debug("current_time_id: %d, portfolio: %.1f" %
(self.current_time_id, self.portfolio_value))
logger.debug("step action: %s" % str(action))
# 到最后一天
if self.current_date == self.dates[-1]:
self.done = True
pre_portfolio_value = self.portfolio_value
sell_prices, buy_prices = self.do_action(action, pre_portfolio_value,
only_update)
self.update_portfolio()
self.update_value_percent()
self.update_reward(sell_prices, buy_prices)
self.obs = self._next()
self.info = {
"orders": self.info["orders"],
"current_date": self.current_date,
"portfolio_value": round(self.portfolio_value / self.investment,
3),
"daily_pnl": round(self.daily_pnl, 1),
"reward": self.reward
}
return self.obs, self.reward, self.done, self.info, self.rewards
def sell_check(self, code='', datestr='', bid_price=None):
# 返回：OK, 成交价
ok = False
# 停牌
if self.is_suspended(code, datestr):
return ok, 0
# 获取当天标的信息
[open, high, low, pct_change
] = self.codes_history[code].loc[datestr,
["open", "high", "low", "pct_chg"]]
# 跌停封板， 不能卖出
if low == high and pct_change < -self.top_pct_change:
logger.debug(u"sell_check %s %s 跌停无法卖出" % (code, datestr))
return ok, 0
# 卖出竞价高于最高价，不可以成交
if bid_price > high:
return ok, 0
# 卖出竞价在最低最高价之间， 可以成交，按出价成交
# NOTE: 这里卖出竞价低于最低价时，可以成交，按最低价成交
if bid_price <= high:
ok = True
return ok, max(bid_price, low)