def get_max_up_stock(data_cen):
"""
日常函数，用于获取连续N天涨停股票
:param data_cen:
:return:find_max_start_max_down_with_buy
"""
result = Calculator.get_max_up_stock(data_cen)
FileOutput.csv_output(None, result, 'two_up_stock.csv')
def multi_process_simulate(self, stock_codes, judge_time, out_dir_name, target_function=None):
"""
多进程进行模拟计算，充分发挥多核性能
:param target_function: 目标调用函数
:param out_dir_name:
:param stock_codes: 所有的股票编码
:param judge_time: 判定时机的函数
:return:
"""
cpu_num = multiprocessing.cpu_count()
each_cpu_load = len(stock_codes) // cpu_num
queue = queues.Queue(20, ctx=multiprocessing)
if target_function is None:
raise Exception('没有传入模拟函数！！！')
final_sum_dict = {}
for item in self.sum_field:
final_sum_dict[item] = 0
for item in self.max_field:
final_sum_dict[item] = 0
for item in self.min_field:
final_sum_dict[item] = 0
for i in range(cpu_num):
temp_stock_code = []
for j in range(i * each_cpu_load, (i + 1) * each_cpu_load):
temp_stock_code.append(stock_codes[j])
if i == cpu_num - 1:
for j in range(cpu_num * each_cpu_load, len(stock_codes)):
temp_stock_code.append(stock_codes[j])
process_i = multiprocessing.Process(target=target_function, args=(temp_stock_code, judge_time, queue,
out_dir_name))
process_i.start()
process_i.join()
for i in range(cpu_num):
temp_dict = queue.get()
for item in self.sum_field:
final_sum_dict[item] = final_sum_dict[item] + temp_dict[item]
for item in self.max_field:
final_sum_dict[item] = temp_dict[item] \
if final_sum_dict[item] < temp_dict[item] \
else final_sum_dict[item]
for item in self.min_field:
final_sum_dict[item] = temp_dict[item] \
if final_sum_dict[item] is None or final_sum_dict[item] > temp_dict[item] \
else final_sum_dict[item]
all_field = self.sum_field
all_field.extend(self.max_field)
all_field.extend(self.min_field)
sum_frame = pandas.DataFrame(columns=tuple(all_field))
sum_frame = sum_frame.append(final_sum_dict, ignore_index=True)
FileOutput.csv_output(None, sum_frame, "final_sum_info", spe_dir_name=out_dir_name)
def __call__(self, *args, **kwargs):
self.data_center = DataCenter()
if self.stock_codes is None or len(self.stock_codes) == 0:
return
# 统计所有的股票，在:param peirod期间之内的盈利情况
sum_dict = {
'win_pct_less_than_50': 0,
'win_pct_bigger_than_50': 0,
'win_pct_bigger_than_60': 0,
'win_pct_bigger_than_70': 0,
}
for stock_code in self.stock_codes:
base_infos = self.data_center.fetch_base_data(stock_code)
shift_close = base_infos['close'].shift(self.period)
win_pct = (shift_close - base_infos['close']) / base_infos['close']
base_infos['win_pct'] = win_pct
base_infos.drop(['vol', 'amount', 'pre_close', 'change'], axis=1, inplace=True)
if not base_infos.empty:
count_percent = len(base_infos[base_infos['win_pct'] > 0]) / len(base_infos)
sum_dict['win_pct_less_than_50'] = sum_dict['win_pct_less_than_50'] + 1 if count_percent < 0.5 \
else sum_dict['win_pct_less_than_50']
sum_dict['win_pct_bigger_than_50'] = sum_dict['win_pct_bigger_than_50'] + 1 if count_percent >= 0.5 \
else sum_dict['win_pct_bigger_than_50']
sum_dict['win_pct_bigger_than_60'] = sum_dict['win_pct_bigger_than_60'] + 1 if count_percent >= 0.6 \
else sum_dict['win_pct_bigger_than_60']
sum_dict['win_pct_bigger_than_70'] = sum_dict['win_pct_bigger_than_70'] + 1 if count_percent >= 0.7 \
else sum_dict['win_pct_bigger_than_70']
sum_str = "win count percent is " + str(count_percent)
file_name = "period_simulate_" + stock_code
now_time = datetime.datetime.now()
now_time_str = now_time.strftime('%Y%m%d')
file_name += '_' + now_time_str
file_name += '.csv'
FileOutput.csv_output(None, base_infos, file_name, spe_dir_name=self.judge_out_name,
extra_content=sum_str)
return sum_dict
def get_max_up_stock(up_days=2):
result = Calculator.get_max_up_stock(data_center, up_days=up_days)
FileOutput.csv_output(None, result, 'two_up_stock_20190304.csv')
def __call__(self, *args, **kwargs):
# TODO -- 看一下Python的数值计算方式，是否有小数？？？
self.data_center = DataCenter()
final_sum_dict = {
'win_num': 0,
'lose_num': 0,
'max_win_pct': 0,
'max_lost_pct': 0,
'ave_win_mny': 0
}
trade_rst = pandas.DataFrame(columns=('ts_code', 'trade_times', 'final_win', 'win_pct'))
for stock_code in self.stock_codes:
trade_rst_dict = {
'ts_code': stock_code,
'trade_times': 0,
'final_win': 0,
'win_pct': 0
}
base_infos = self.data_center.fetch_base_data(stock_code)
if base_infos is None or len(base_infos) == 0:
continue
ret_time = []
for item in self.judge_time:
ret_time.append(item(base_infos))
if ret_time is None or len(ret_time) == 0:
continue
detail_trade_info = pandas.DataFrame(
columns=('ts_code', 'curr_close', 'trade_date', 'trade_num', 'hold_num', 'hold_mny',
'total_mny'))
if len(ret_time[0]) != len(base_infos):
print("Not equals!!!")
for i in range(len(ret_time[0])):
# 判定是否是买入时机
operate_flag = self.DO_NOTHING
temp_buy_pct = 0.1
for item in ret_time:
# 从当前往前5天之内是否有发出过买入信号，如果有，就算有
start_index = i - self.MULTI_INDI_BETWEEN
start_index = 0 if start_index < 0 else start_index
temp_val = self.DO_NOTHING
for j in range(start_index, i + 1):
if item.at[j, 'flag'] == self.BUY_FLAG:
temp_val = self.BUY_FLAG
temp_buy_pct = item.at[j, 'percent'] if 0 < item.at[j, 'percent'] < temp_buy_pct \
else temp_buy_pct
break
elif item.at[j, 'flag'] == self.SOLD_FLAG:
temp_val = self.SOLD_FLAG
temp_buy_pct = item.at[j, 'percent'] if 0 < item.at[j, 'percent'] < temp_buy_pct \
else temp_buy_pct
break
if operate_flag == self.DO_NOTHING or operate_flag == temp_val:
operate_flag = temp_val
continue
else:
operate_flag = self.DO_NOTHING
break
if operate_flag == self.BUY_FLAG:
# 默认买入10%
buy_pct = temp_buy_pct
buy_mny = self.initial_mny * buy_pct
buy_mny = self.left_mny if self.left_mny < buy_mny else buy_mny
buy_num = buy_mny / base_infos.at[i, 'close']
self.hold_num = self.hold_num + buy_num
self.left_mny = self.left_mny - buy_mny
hold_mny = self.hold_num * base_infos.at[i, 'close']
# 记录买入信息
temp_dict = {
'ts_code': stock_code,
'trade_date': base_infos.at[i, 'trade_date'],
'trade_num': buy_num,
'hold_num': self.hold_num,
'hold_mny': hold_mny,
'total_mny': self.left_mny + hold_mny,
'curr_close': base_infos.at[i, 'close']
}
detail_trade_info = detail_trade_info.append(temp_dict, ignore_index=True)
trade_rst_dict['trade_times'] = trade_rst_dict['trade_times'] + 1
elif operate_flag == self.SOLD_FLAG:
# 默认卖出持仓数量的10%
sold_pct = temp_buy_pct
sold_num = self.hold_num * sold_pct
sold_num = sold_num / 100 * 100
self.hold_num = self.hold_num - sold_num
sold_mny = sold_num * base_infos.at[i, 'close']
self.left_mny = self.left_mny + sold_mny
hold_mny = self.hold_num * base_infos.at[i, 'close']
# 记录卖出信息
temp_dict = {
'ts_code': stock_code,
'trade_date': base_infos.at[i, 'trade_date'],
'trade_num': sold_num,
'hold_num': self.hold_num,
'hold_mny': hold_mny,
'total_mny': self.left_mny + hold_mny,
'curr_close': base_infos.at[i, 'close']
}
detail_trade_info = detail_trade_info.append(temp_dict, ignore_index=True)
trade_rst_dict['trade_times'] = trade_rst_dict['trade_times'] + 1
if not detail_trade_info.empty:
file_name = "simulate_" + str(stock_code)
now_time = datetime.datetime.now()
now_time_str = now_time.strftime('%Y%m%d')
file_name += '_' + now_time_str
file_name += '.csv'
FileOutput.csv_output(None, detail_trade_info, file_name, spe_dir_name=self.judge_out_name)
else:
print("no such stock!")
self.reset()
if detail_trade_info.empty:
continue
# 统计该只股票的最终盈利
last_win = detail_trade_info.at[(len(detail_trade_info) - 1), 'total_mny'] - self.initial_mny
last_win_pct = last_win / self.initial_mny
trade_rst_dict['final_win'] = last_win
trade_rst_dict['win_pct'] = last_win_pct
trade_rst = trade_rst.append(trade_rst_dict, ignore_index=True)
# 统计多只股票的汇总
final_sum_dict['win_num'] = final_sum_dict['win_num'] + 1 if last_win > 0 else final_sum_dict['win_num']
final_sum_dict['lose_num'] = final_sum_dict['lose_num'] + 1 if last_win <= 0 else final_sum_dict['lose_num']
final_sum_dict['max_win_pct'] = last_win_pct if last_win_pct > final_sum_dict['max_win_pct'] else \
final_sum_dict['max_win_pct']
final_sum_dict['max_lost_pct'] = last_win_pct if last_win_pct <= final_sum_dict['max_lost_pct'] else \
final_sum_dict['max_lost_pct']
sum_str = "获利数量：" + str(final_sum_dict['win_num']) + " 损失数量：" + str(final_sum_dict['lose_num']) \
+ " 最大获利百分比：" + str(final_sum_dict['max_win_pct']) + " 最大损失百分比：" + \
str(final_sum_dict['max_lost_pct'])
if not trade_rst.empty:
file_name = "trade_group_" + self.stock_codes[0]
now_time = datetime.datetime.now()
now_time_str = now_time.strftime('%Y%m%d')
file_name += '_' + now_time_str
file_name += '.csv'
FileOutput.csv_output(None, trade_rst, file_name, spe_dir_name=self.judge_out_name, extra_content=sum_str)
return final_sum_dict