def get_field(self,
contract: str = 'main',
price: str = 'close',
rebalance_num: int = 1,
field: str = 'continuous_price') -> DataFrame:
"""
获取连续合约指定字段的数据
Parameters
----------
contract: str
合约种类，目前可选有main和active_near,main表示主力合约,active_near表示活跃近月
rebalance_num: int, default = 1
换仓天数，可选1，3，5
field: str, default = 'close'
字段，continuous_open或者continuous_close
Returns
-------
df: DataFrame
连续合约field字段数据，一般是开盘价或收盘价
"""
data = self.get_continuous_contract_data(contract=contract,
price=price,
rebalance_num=rebalance_num)
df = stack_dataframe_by_fields(data=data,
index_field='datetime',
column_field='underlying_symbol',
data_field=field)
return df
def get_field(self, symbol: str, field: str) -> DataFrame:
"""
获取单品种收盘价DataFrame
Parameters
________
symbol: string
品种代码
field: string
要获取的字段，如open, high, low, close
Returns
-------
单品种不同合约的收盘价，index是交易日期，columns是合约代码，data是字段值
Example:
________
>>> self = DailyDataManager()
>>> df = self.get_field("A", "close")
>>>>print(df)
contract A0901 A0903 A0905 A0907 ... A2105 A2107 A2109 A2111
datetime ...
2009-01-05 3660.0 3598.0 3413.0 3308.0 ... NaN NaN NaN NaN
2009-01-06 3679.0 3630.0 3459.0 3360.0 ... NaN NaN NaN NaN
2009-01-07 3599.0 3610.0 3448.0 3360.0 ... NaN NaN NaN NaN
2009-01-08 3600.0 3623.0 3420.0 3348.0 ... NaN NaN NaN NaN
2009-01-09 3600.0 3623.0 3462.0 3407.0 ... NaN NaN NaN NaN
...
"""
# 预先检查
if field not in self.daily_data.columns.drop(['contract', 'datetime']):
raise NameError(f"No field named {field}")
if symbol not in self.daily_data.underlying_symbol.tolist():
raise NameError(f"No symbol named {symbol}")
# 如果已有数据，则直接返回
if field in self.fields_dict:
if symbol in self.fields_dict[field]:
return self.fields_dict[field][symbol]
daily_data = self.daily_data
daily_data = daily_data[daily_data['underlying_symbol'] == symbol]
data = stack_dataframe_by_fields(data=daily_data,
index_field='datetime',
column_field='contract',
data_field=field)
self.fields_dict[field][symbol] = data
return data
def get_near_contract_except_main(daily_data: DataFrame,
main_contract_df: DataFrame) -> DataFrame:
"""
获取单品种除主力合约以外的近月合约
:param daily_data: 单品种日线行情数据
:param main_contract_df: 单品种每日主力合约,注意是shift前的main_contract
:type daily_data: DataFrame
:type main_contract_df: DataFrame
:return: 单品种每日近月合约，注意是shift前的近月合约，Series，index是时间，data是shift前的近月合约
"""
try:
close = stack_dataframe_by_fields(data=daily_data,
index_field='datetime',
column_field='contract',
data_field='close')
except KeyError:
close = daily_data.copy()
main_contract_list = main_contract_df['main_contract'].tolist()
near_contract_list = []
for i in range(len(close)):
if i != len(close) - 1:
# 可选合约池:今天有交易且排除掉主力合约
if main_contract_list[i] in close.iloc[i].dropna().index:
pool = close.iloc[i].dropna().index.drop(main_contract_list[i])
# 如果没有可选合约，则返回缺失值
if len(pool) == 0:
print(main_contract_list[i], close.index[i])
near_contract_list.append(np.nan)
else:
near_contract_list.append(pool.min())
else:
near_contract_list.append(np.nan)
else:
pool = close.iloc[i].dropna().index.drop(main_contract_list[i])
near_contract_list.append(pool.min())
near_contract_df = pd.Series(near_contract_list,
index=close.index).reset_index()
near_contract_df.columns = ['datetime', 'near_contract']
return near_contract_df
def compute_factor(self) -> DataFrame:
"""
计算因子，通过对compute_single_factor实现
Returns
-------
因子值: DataFrame
index为datetime, columns为underlying_symbol, data为factor
"""
symbol_list = self.get_symbol_list()
factor_list = []
for symbol in tqdm(symbol_list):
factor = self.compute_single_factor(symbol)
factor_list.append(factor)
factor = pd.concat(factor_list, axis=0)
factor = stack_dataframe_by_fields(data=factor,
index_field='datetime',
column_field='underlying_symbol',
data_field='factor')
factor = factor.rolling(window=self.window, min_periods=1).mean()
self.factor_value = factor
return factor