def _weekly_start_end_date_handler(start_date, end_date):
start_date = to_date(start_date)
monday = start_date - datetime.timedelta(days=start_date.weekday())
first_trading_day_in_week = monday if is_trading_date(monday) else get_next_trading_date(monday)
if first_trading_day_in_week < start_date:
start_date = monday + datetime.timedelta(weeks=1)
end_date = to_date(end_date)
if end_date > datetime.date.today():
end_date = datetime.date.today()
friday = end_date - datetime.timedelta(days=end_date.weekday()) + datetime.timedelta(days=4)
last_trading_day_in_week = friday if is_trading_date(friday) else get_previous_trading_date(friday)
if last_trading_day_in_week > end_date:
end_date = friday - datetime.timedelta(weeks=1)
return to_date_int(start_date), to_date_int(end_date)
def get_factor_covariance(date, horizon='daily'):
""" 获取因子协方差矩阵
:param date: str 日期（例如：‘2017-03-20’）
:param horizon: str 预测期限。默认为日度（'daily'），可选月度（‘monthly’）或季度（'quarterly'）。
:return: pandas.DataFrame，其中 index 和 column 均为因子名称。
"""
date = get_previous_trading_date(get_next_trading_date(date))
date = ensure_date_int(date)
ensure_string_in(horizon, HORIZON_CONTAINER, 'horizon')
df = get_client().execute('get_factor_covariance', date, horizon)
if not df:
return
df = pd.DataFrame(df)
df.drop("date", axis=1, inplace=True)
return df.reindex(columns=df.index)
def get_current_trading_date(dt):
if 7 <= dt.hour < 18:
return datetime.datetime(year=dt.year, month=dt.month, day=dt.day)
return get_next_trading_date(dt - datetime.timedelta(hours=4))
def get_tick_price_multi_df(order_book_ids, start_date, end_date, fields, market):
ins_list = ensure_instruments(order_book_ids)
order_book_ids = [ins.order_book_id for ins in ins_list]
types = {ins.type for ins in ins_list}
start_date, end_date = ensure_date_range(start_date, end_date, datetime.timedelta(days=3))
if "Future" in types or "Option" in types:
base_fields = FUTURE_TICK_FIELDS
base_columns = FUTURE_TICK_COLUMNS
else:
base_fields = EQUITIES_TICK_FIELDS
base_columns = EQUITIES_TICK_COLUMNS
if fields:
fields = ensure_list_of_string(fields, "fields")
check_items_in_container(fields, base_fields, "fields")
columns = [f for f in base_columns if f in fields]
else:
fields = base_fields
columns = base_columns
gtw_fields = set(fields) | {"date", "time"}
if set(fields) & RELATED_DABAR_FIELDS:
gtw_fields.update({"volume", "last"})
history_df = get_history_tick(order_book_ids, start_date, end_date, list(gtw_fields), columns, market)
history_latest_date = 0 if history_df is None else date_to_int8(get_current_trading_date(
history_df.index.get_level_values(1).max()))
today = today_int()
next_trading_date = date_to_int8(get_next_trading_date(today, market=market))
if history_latest_date >= end_date or start_date > next_trading_date or end_date < today:
return history_df
if end_date >= next_trading_date and (start_date > today or history_latest_date >= today):
live_date = next_trading_date
else:
live_date = today
if history_latest_date >= live_date:
return history_df
live_dfs = []
for ins in ins_list:
try:
live_df = get_ticks(ins.order_book_id, start_date=live_date, end_date=live_date, expect_df=True,
market=market)
except (PermissionDenied, MarketNotSupportError):
pass
else:
if live_df is None:
continue
if "trading_date" not in live_df.columns:
live_df["trading_date"] = int8_to_datetime(live_date)
else:
live_df["trading_date"] = live_df["trading_date"].apply(to_datetime)
if ins.type in ("Future", "Option"):
live_df["change_rate"] = live_df["last"] / live_df["prev_settlement"] - 1
else:
live_df["change_rate"] = live_df["last"] / live_df["prev_close"] - 1
live_df = live_df.reindex(columns=columns)
live_dfs.append(live_df)
if not live_dfs:
return history_df
if history_df is None:
return pd.concat(live_dfs)
return pd.concat([history_df] + live_dfs)