def get_start_date(
self,
code_list: Union[str,
Tuple[str],
List[str]],
factor_time_range: list
):
"""
获取上市时间
"""
# stock_list = utils.QA_fmt_code_list(code_list, style="jq")
# df_local = jqdatasdk.get_all_securities(types="stock")
stock_list = utils.QA_fmt_code_list(code_list)
df_local = QA_fetch_stock_basic(status=None).set_index("code")
intersection = list(df_local.index.intersection(stock_list))
ss = df_local.loc[intersection]["list_date"]
# ss.index = ss.index.map(lambda x: x[:6])
# 日期处理
date_range = list(map(lambda x: x.date(), factor_time_range))
multiindex = pd.MultiIndex.from_product(
[date_range,
utils.QA_fmt_code_list(intersection)]
)
values = multiindex.map(lambda x: ss.loc[x[1]]).tolist()
df_local = pd.Series(index=multiindex, data=values)
df_local.index.names = ["date", "code"]
return df_local
def get_groupby(
self,
code_list: Union[str, Tuple[str], List[str]],
factor_time_range: list = None,
industry_cls: str = None,
industry_data: Union[dict, pd.Series, pd.DataFrame] = None,
frequence: str = 'DAY',
detailed: bool = False,
) -> pd.Series:
"""
获取行业信息
参数
---
:param code_list: 股票代码
:param factor_time_range: 因子时间范围
:param industry_cls: 行业分类
:param industry_data: 行业信息
:param detailed: 如果按日期获取行业数据，对速度影响很大,
设置该参数为 False, 则默认按照 end_time 的行业分类作为全部日期的行业信息
返回
---
:return: 以 ['日期' '股票'] 为索引的行业信息
"""
# 1. 股票代码
if isinstance(code_list, tuple):
code_list = list(code_list)
# 2. 日期处理
if not detailed:
date_range = [pd.Timestamp(max(factor_time_range))]
else:
date_range = factor_time_range
# 3. 行业处理
if not industry_cls:
industry_cls = self.industry_cls
if not industry_data:
industry_data = self.industry_data
if (not industry_cls) and (not industry_data):
warnings.warn("没有指定行业分类方式，也没有输入行业分类信息", UserWarning)
return pd.Series(
index=pd.MultiIndex.from_product([date_range, code_list]),
name="group",
data="NA",
)
# 4. 具体行业
if not industry_cls: # 没有输入行业分类信息
df_local = pd.DataFrame()
if isinstance(industry_data, dict):
df_local = pd.DataFrame(index=date_range,
data=[industry_data] *
len(date_range)).stack(level=-1)
if isinstance(industry_data, pd.DataFrame):
df_local = industry_data.stack(level=-1)
else:
df_local = industry_data
df_local.index.names = ["date", "code"]
return df_local
# 如果输入了行业分类信息，按行业分类信息进行处理
# FIXME: 暂时使用聚宽的行业数据
stock_list = utils.QA_fmt_code_list(code_list, style="jq")
df_local = pd.DataFrame()
industries = map(partial(jqdatasdk.get_industry, stock_list),
date_range)
industries = {
d: {
s: ind.get(s).get(industry_cls,
dict()).get("industry_name", "NA")
for s in stock_list
}
for d, ind in zip(date_range, industries)
}
df_local = pd.DataFrame(industries).T.sort_index()
df_local.columns = df_local.columns.map(lambda x: x[0:6])
df_local = df_local.stack(level=-1)
df_local.index.names = ["date", "code"]
return df_local
def __gen_clean_factor_and_forward_returns(self):
"""
格式化因子数据，附加因子远期收益，分组，权重信息
"""
factor_data = self.factor
# 股票代码: 默认转换为 QA 支持格式
code_list = utils.QA_fmt_code_list(
list(factor_data.index.get_level_values("code").drop_duplicates())
)
# 因子日期
factor_time_range = list(factor_data.index.levels[0].drop_duplicates())
start_time = min(factor_time_range)
end_time = max(factor_time_range)
# 附加数据
if hasattr(self.prices, "__call__"):
prices = self.prices(
code_list=code_list,
start_time=start_time,
end_time=end_time,
frequence=self.frequence,
)
prices = prices.loc[~prices.index.duplicated()]
else:
prices = self.prices
self.prices = prices
if hasattr(self.groupby, "__call__"):
groupby = self.groupby(
code_list=code_list,
factor_time_range=factor_time_range
)
else:
groupby = self.groupby
self.groupby = groupby
if hasattr(self.stock_start_date, "__call__"):
stock_start_date = self.stock_start_date(
code_list=code_list,
factor_time_range=factor_time_range
)
else:
stock_start_date = self.stock_start_date
self.stock_start_date = stock_start_date
if hasattr(self.weights, "__call__"):
weights = self.weights(
code_list=code_list,
factor_time_range=factor_time_range,
frequence=self.frequence
)
else:
weights = self.weights
self.weights = weights
# 周期处理
# self.interval = utils.get_interval(self.frequence)
# 4. 因子处理
self._clean_factor_data = get_clean_factor_and_forward_returns(
factor=factor_data,
prices=self.prices,
groupby=self.groupby,
stock_start_date=self.stock_start_date,
weights=self.weights,
binning_by_group=self.binning_by_group,
quantiles=self.quantiles,
bins=self.bins,
periods=self.periods,
max_loss=self.max_loss,
zero_aware=self.zero_aware,
frequence=self.frequence,
)
def get_groupby(
self,
code_list: Union[str,
Tuple[str],
List[str]],
factor_time_range: list = None,
industry_cls: str = None,
industry_data: Union[dict,
pd.Series,
pd.DataFrame] = None,
frequence: str = 'DAY',
detailed: bool = False,
) -> pd.Series:
"""
获取行业信息
参数
---
:param code_list: 股票代码
:param factor_time_range: 因子时间范围
:param industry_cls: 行业分类
:param industry_data: 行业信息
:param detailed: 如果按日期获取行业数据，对速度影响很大,
设置该参数为 False, 则默认按照 end_time 的行业分类作为全部日期的行业信息
返回
---
:return: 以 ['日期' '股票'] 为索引的行业信息
"""
# 1. 股票代码
if isinstance(code_list, tuple):
code_list = list(code_list)
# 2. 日期处理
if not detailed:
date_range = [pd.Timestamp(max(factor_time_range)).date()]
else:
date_range = list(map(lambda x: x.date(), factor_time_range))
# 3. 行业处理
if not industry_cls:
industry_cls = self.industry_cls
if not industry_data:
industry_data = self.industry_data
if (industry_cls is None) and (industry_data is None):
warnings.warn("没有指定行业分类方式，也没有输入行业分类信息", UserWarning)
return pd.Series(
index=pd.MultiIndex.from_product([date_range,
code_list]),
name="group",
data="NA",
)
# 4. 具体行业
if not industry_cls: # 没有输入行业分类信息
df_local = pd.DataFrame()
if isinstance(industry_data, dict):
df_local = pd.DataFrame(
index=date_range,
data=[industry_data] * len(date_range)
).stack(level=-1)
if isinstance(industry_data, pd.DataFrame):
df_local = industry_data.stack(level=-1)
else:
df_local = industry_data
df_local.index.names = ["date", "code"]
return df_local
# 如果输入了行业分类信息，按行业分类信息进行处理
# FIXME: 暂时使用聚宽的行业数据
# stock_list = utils.QA_fmt_code_list(code_list, style="jq")
stock_list = utils.QA_fmt_code_list(code_list)
df_local = pd.DataFrame()
for cursor_date in date_range:
df_tmp = QA_fetch_industry_adv(
code=code_list,
cursor_date = cursor_date,
levels=industry_cls.split('_')[1],
src=industry_cls.split("_")[0])[["code", "industry_name"]]
df_tmp["date"] = cursor_date
df_local = df_local.append(df_tmp)
# industries = map(
# partial(jqdatasdk.get_industry,
# stock_list),
# date_range
# )
# industries = {
# d: {
# s: ind.get(s).get(industry_cls,
# dict()).get("industry_name",
# "NA")
# for s in stock_list
# }
# for d,
# ind in zip(date_range,
# industries)
# }
# df_local = pd.DataFrame(industries).T.sort_index()
# df_local.columns = df_local.columns.map(lambda x: x[0:6])
# df_local = df_local.stack(level=-1)
# df_local.index.names = ["date", "code"]
df_local = df_local.set_index(["date", "code"])
return df_local["industry_name"]
