def loadon_data(self, trade_date):
db_polymerize = DBPolymerize(self._name)
max_windows = self._maximization_windows()
begin_date = advanceDateByCalendar('china.sse', trade_date,
'-%sb' % (max_windows + 1))
total_data = db_polymerize.fetch_data(begin_date, trade_date, '1b')
return total_data
def loadon_data(self, begin_date, end_date, benchmark_code_dict, table):
# 需要增加因子方向
db_polymerize = DBPolymerize(self._name)
db_factor = FetchRLFactorEngine(table)
factor_data = db_factor.fetch_factors(begin_date=begin_date, end_date=end_date)
benchmark_data, market_data, exposure_data = db_polymerize.fetch_integrated_data(begin_date, end_date)
# 针对不同的基准
total_data_dict = {}
benchmark_industry_weights_dict = {}
for key, value in benchmark_code_dict.items():
total_data = self.preprocessing(benchmark_data[benchmark_data.index_code == key], market_data, factor_data,
exposure_data)
# 因子标准化
total_data = total_data.sort_values(['trade_date', 'security_code'])
# 因子中性化
total_data = total_data.groupby(['trade_date']).apply(self._factor_preprocess)
total_data.loc[:, self._factor_columns] = total_data.loc[:, self._factor_columns].fillna(0)
benchmark_industry_weights = benchmark_data[benchmark_data.index_code == key].groupby(
['trade_date', 'industry_code']).apply(lambda x: x['weighing'].sum())
benchmark_industry_weights = benchmark_industry_weights.unstack().fillna(0)
total_data_dict[value] = total_data
benchmark_industry_weights_dict[value] = benchmark_industry_weights
return total_data_dict, benchmark_industry_weights_dict
def loadon_data(self, trade_date):
db_polymerize = DBPolymerize(self._name)
max_windows = self._maximization_windows()
begin_date = advanceDateByCalendar('china.sse', trade_date,
'-%sb' % (max_windows + 1))
market_data, index_data = db_polymerize.fetch_volatility_value_data(
begin_date, trade_date, '1b')
# market_data, index_data = db_polymerize.fetch_volatility_value_data('2018-08-15', trade_date, '1b')
return market_data, index_data
def loadon_data(self, begin_date, trade_date, table):
# 确定获取的日期范围
freq = '1m'
benchmark_code_dict = {
'000905.XSHG': '2070000187',
'000300.XSHG': '2070000060'
}
db_factor = FetchRLFactorEngine(table)
factor_data = db_factor.fetch_factors(begin_date=begin_date,
end_date=trade_date,
freq=freq)
db_polymerize = DBPolymerize(self._name)
benchmark_data, index_data, market_data, exposure_data = db_polymerize.fetch_performance_data(
benchmark_code_dict.keys(), begin_date, trade_date, freq)
# 针对不同的基准
total_data_dict = {}
index_rets_dict = {}
benchmark_industry_weights_dict = {}
for key, value in benchmark_code_dict.items():
total_data, index_rets, coverage_rate = self.performance_preprocessing(
benchmark_data[benchmark_data.index_code == key],
index_data[index_data.security_code == value], market_data,
factor_data, exposure_data)
# 中性化处理，因子值填充待修改
total_data = total_data.sort_values(
['trade_date', 'security_code'])
total_data = total_data.groupby(['trade_date'
]).apply(self._factor_preprocess)
total_data.loc[:, self.
_factor_columns] = total_data.loc[:, self.
_factor_columns].fillna(
0)
benchmark_industry_weights = benchmark_data[
benchmark_data.index_code == key].groupby(
['trade_date',
'industry_code']).apply(lambda x: x['weighing'].sum())
benchmark_industry_weights = benchmark_industry_weights.unstack(
).fillna(0)
total_data_dict[value] = total_data
index_rets_dict[value] = index_rets
benchmark_industry_weights_dict[value] = benchmark_industry_weights
return total_data_dict, benchmark_industry_weights_dict, index_rets_dict
def loadon_data(self, trade_date):
benchmark = ['000905.XSHG', '000300.XSHG']
db_polymerize = DBPolymerize(self._name)
benchmark_data, index_data, market_data, factor_data, exposure_data = db_polymerize.fetch_performance_data(
benchmark, '2018-08-19', trade_date, '1b')
total_data, index_se_dict = self.performance_preprocessing(
benchmark_data, index_data, market_data, factor_data,
exposure_data)
#中性化处理
total_data = total_data.sort_values(['trade_date', 'security_code'])
total_data = total_data.groupby(['trade_date'
]).apply(self._factor_preprocess)
benchmark_industry_weights = benchmark_data.groupby(
['trade_date',
'industry_code']).apply(lambda x: x['weighing'].sum())
benchmark_industry_weights = benchmark_industry_weights.unstack(
).fillna(0)
return total_data, benchmark_industry_weights, index_se_dict,
def loadon_data(self, begin_date, end_date, benchmark_code_dict, table):
# 需要增加因子方向
db_polymerize = DBPolymerize(self._name)
db_factor = FetchRLFactorEngine(table)
factor_data = db_factor.fetch_factors(begin_date=begin_date,
end_date=end_date)
benchmark_data, market_data, exposure_data = db_polymerize.fetch_integrated_data(
begin_date, end_date)
# 获取因子列表
self._factor_columns = [
i for i in factor_data.columns
if i not in ['id', 'trade_date', 'security_code']
]
# 针对不同的基准
total_data_dict = {}
benchmark_industry_weights_dict = {}
coverage_rate_dict = {}
factor_direction_dict = {}
# 获取因子方向
db_url = '''mysql+mysqlconnector://{0}:{1}@{2}:{3}/{4}'''.format(
config.rl_db_user, config.rl_db_pwd, config.rl_db_host,
config.rl_db_port, config.rl_db_database)
def get_factor_direction(factors_string):
destination = sa.create_engine(db_url)
sql = """select factor_name, universe, factor_direction from factor_performance_ic_ir_sub_test
where time_type='10' and factor_name in ( {0} );""".format(
factors_string)
try:
factor_sets = pd.read_sql(sql, destination)
return factor_sets
except:
print("failed to get the factor directions!")
factor_str = ','.join(
["'" + str(factor) + "'" for factor in self._factor_columns])
factor_directions = get_factor_direction(factor_str)
for key, value in benchmark_code_dict.items():
total_data, coverage_rate = self.preprocessing(
benchmark_data[benchmark_data.index_code == key], market_data,
factor_data, exposure_data)
# 因子标准化
total_data = total_data.sort_values(
['trade_date', 'security_code'])
total_data = total_data.groupby(['trade_date'
]).apply(self._factor_preprocess)
total_data.loc[:, self.
_factor_columns] = total_data.loc[:, self.
_factor_columns].fillna(
0)
benchmark_industry_weights = benchmark_data[
benchmark_data.index_code == key].groupby(
['trade_date',
'industry_code']).apply(lambda x: x['weighing'].sum())
benchmark_industry_weights = benchmark_industry_weights.unstack(
).fillna(0)
total_data_dict[value] = total_data
benchmark_industry_weights_dict[value] = benchmark_industry_weights
coverage_rate_dict[value] = coverage_rate
factor_direction_dict[value] = factor_directions[
factor_directions.universe == key]
return total_data_dict, benchmark_industry_weights_dict, coverage_rate_dict, factor_direction_dict