def get_roll_yield(date=None, var="CU", symbol1=None, symbol2=None, df=None):
"""
指定交易日指定品种（主力和次主力）或任意两个合约的展期收益率
Parameters
------
date: string 某一天日期 format： YYYYMMDD
var: string 合约品种如RB、AL等
symbol1: string 合约 1如 rb1810
symbol2: string 合约 2 如 rb1812
df: DataFrame或None 从dailyBar得到合约价格，如果为空就在函数内部抓dailyBar，直接喂给数据可以让计算加快
Return
-------
tuple
roll_yield
near_by
deferred
"""
# date = "20200304"
date = cons.convert_date(
date) if date is not None else datetime.date.today()
if date.strftime("%Y%m%d") not in calendar:
warnings.warn("%s非交易日" % date.strftime("%Y%m%d"))
return None
if symbol1:
var = symbol_varieties(symbol1)
if not isinstance(df, pd.DataFrame):
market = symbol_market(var)
df = get_futures_daily(start_day=date, end_day=date, market=market)
if var:
df = df[~df["symbol"].str.contains(
"efp")] # 20200304 由于交易所获取的数据中会有比如 "CUefp"，所以在这里过滤
df = df[df["variety"] == var].sort_values("open_interest",
ascending=False)
df["close"] = df["close"].astype("float")
symbol1 = df["symbol"].tolist()[0]
symbol2 = df["symbol"].tolist()[1]
close1 = df["close"][df["symbol"] == symbol1.upper()].tolist()[0]
close2 = df["close"][df["symbol"] == symbol2.upper()].tolist()[0]
a = re.sub(r"\D", "", symbol1)
a_1 = int(a[:-2])
a_2 = int(a[-2:])
b = re.sub(r"\D", "", symbol2)
b_1 = int(b[:-2])
b_2 = int(b[-2:])
c = (a_1 - b_1) * 12 + (a_2 - b_2)
if close1 == 0 or close2 == 0:
return False
if c > 0:
return np.log(close2 / close1) / c * 12, symbol2, symbol1
else:
return np.log(close2 / close1) / c * 12, symbol1, symbol2
def _table_cut_cal(table_cut, symbol):
"""
表格切分
:param table_cut: 需要切分的表格
:type table_cut: pandas.DataFrame
:param symbol: 品种
:type symbol: str
:return:
:rtype: pandas.DataFrame
"""
var = symbol_varieties(symbol)
table_cut[intColumns + ['rank']] = table_cut[intColumns + ['rank']].astype(int)
table_cut_sum = table_cut.sum()
table_cut_sum['rank'] = 999
for col in ['vol_party_name', 'long_party_name', 'short_party_name']:
table_cut_sum[col] = None
table_cut = table_cut.append(pd.DataFrame(table_cut_sum).T, sort=True)
table_cut['symbol'] = symbol
table_cut['variety'] = var
table_cut[intColumns + ['rank']] = table_cut[intColumns + ['rank']].astype(int)
return table_cut
def get_rank_sum(date=None, vars_list=cons.contract_symbols):
"""
抓取四个期货交易所前5、前10、前15、前20会员持仓排名数据
注1：由于上期所和中金所只公布每个品种内部的标的排名, 没有公布品种的总排名;
所以函数输出的品种排名是由品种中的每个标的加总获得, 并不是真实的品种排名列表
注2：大商所只公布了品种排名, 未公布标的排名
:param date: 日期 format: YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
:param vars_list: 合约品种如 RB, AL等列表 为空时为所有商品
:return: pd.DataFrame:
展期收益率数据
symbol 标的合约 string
var 商品品种 string
vol_top5 成交量前5会员成交量总和 int
vol_chg_top5 成交量前5会员成交量变化总和 int
long_open_interest_top5 持多单前5会员持多单总和 int
long_open_interest_chg_top5 持多单前5会员持多单变化总和 int
short_open_interest_top5 持空单前5会员持空单总和 int
short_open_interest_chg_top5 持空单前5会员持空单变化总和 int
vol_top10 成交量前10会员成交量总和 int
...
date 日期 string YYYYMMDD
"""
date = cons.convert_date(date) if date is not None else datetime.date.today()
if date.strftime('%Y%m%d') not in calendar:
warnings.warn('%s非交易日' % date.strftime('%Y%m%d'))
return None
dce_var = [i for i in vars_list if i in cons.market_exchange_symbols['dce']]
shfe_var = [i for i in vars_list if i in cons.market_exchange_symbols['shfe']]
czce_var = [i for i in vars_list if i in cons.market_exchange_symbols['czce']]
cffex_var = [i for i in vars_list if i in cons.market_exchange_symbols['cffex']]
big_dict = {}
if len(dce_var) > 0:
data = get_dce_rank_table(date, dce_var)
if data is False:
return False
big_dict.update(data)
if len(shfe_var) > 0:
data = get_shfe_rank_table(date, shfe_var)
if data is False:
return False
big_dict.update(data)
if len(czce_var) > 0:
data = get_czce_rank_table(date, czce_var)
if data is False:
return False
big_dict.update(data)
if len(cffex_var) > 0:
data = get_cffex_rank_table(date, cffex_var)
if data is False:
return False
big_dict.update(data)
records = pd.DataFrame()
for symbol, table in big_dict.items():
table = table.applymap(lambda x: 0 if x == '' else x)
for symbol_inner in set(table['symbol']):
var = symbol_varieties(symbol_inner)
if var in vars_list:
table_cut = table[table['symbol'] == symbol_inner]
table_cut['rank'] = table_cut['rank'].astype('float')
table_cut_top5 = table_cut[table_cut['rank'] <= 5]
table_cut_top10 = table_cut[table_cut['rank'] <= 10]
table_cut_top15 = table_cut[table_cut['rank'] <= 15]
table_cut_top20 = table_cut[table_cut['rank'] <= 20]
big_dict = {'symbol': symbol_inner, 'variety': var,
'vol_top5': table_cut_top5['vol'].sum(), 'vol_chg_top5': table_cut_top5['vol_chg'].sum(),
'long_open_interest_top5': table_cut_top5['long_open_interest'].sum(),
'long_open_interest_chg_top5': table_cut_top5['long_open_interest_chg'].sum(),
'short_open_interest_top5': table_cut_top5['short_open_interest'].sum(),
'short_open_interest_chg_top5': table_cut_top5['short_open_interest_chg'].sum(),
'vol_top10': table_cut_top10['vol'].sum(),
'vol_chg_top10': table_cut_top10['vol_chg'].sum(),
'long_open_interest_top10': table_cut_top10['long_open_interest'].sum(),
'long_open_interest_chg_top10': table_cut_top10['long_open_interest_chg'].sum(),
'short_open_interest_top10': table_cut_top10['short_open_interest'].sum(),
'short_open_interest_chg_top10': table_cut_top10['short_open_interest_chg'].sum(),
'vol_top15': table_cut_top15['vol'].sum(),
'vol_chg_top15': table_cut_top15['vol_chg'].sum(),
'long_open_interest_top15': table_cut_top15['long_open_interest'].sum(),
'long_open_interest_chg_top15': table_cut_top15['long_open_interest_chg'].sum(),
'short_open_interest_top15': table_cut_top15['short_open_interest'].sum(),
'short_open_interest_chg_top15': table_cut_top15['short_open_interest_chg'].sum(),
'vol_top20': table_cut_top20['vol'].sum(),
'vol_chg_top20': table_cut_top20['vol_chg'].sum(),
'long_open_interest_top20': table_cut_top20['long_open_interest'].sum(),
'long_open_interest_chg_top20': table_cut_top20['long_open_interest_chg'].sum(),
'short_open_interest_top20': table_cut_top20['short_open_interest'].sum(),
'short_open_interest_chg_top20': table_cut_top20['short_open_interest_chg'].sum(),
'date': date.strftime('%Y%m%d')
}
records = records.append(pd.DataFrame(big_dict, index=[0]))
if len(big_dict.items()) > 0:
add_vars = [i for i in cons.market_exchange_symbols['shfe'] + cons.market_exchange_symbols['cffex'] if
i in records['variety'].tolist()]
for var in add_vars:
records_cut = records[records['variety'] == var]
var_record = pd.DataFrame(records_cut.sum()).T
var_record['date'] = date.strftime('%Y%m%d')
var_record.loc[:, ['variety', 'symbol']] = var
records = records.append(var_record)
return records.reset_index(drop=True)
def get_czce_rank_table(date=None, vars_list=cons.contract_symbols):
"""
郑州商品交易所前 20 会员持仓排名数据明细
注：该交易所既公布了品种排名, 也公布了标的排名
:param date: 日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
:param vars_list: 合约品种如RB、AL等列表 为空时为所有商品, 数据从20050509开始，每交易日16:30左右更新数据
:return: pd.DataFrame
rank 排名 int
vol_party_name 成交量排序的当前名次会员 string(中文)
vol 该会员成交量 int
vol_chg 该会员成交量变化量 int
long_party_name 持多单排序的当前名次会员 string(中文)
long_open_interest 该会员持多单 int
long_open_interest_chg 该会员持多单变化量 int
short_party_name 持空单排序的当前名次会员 string(中文)
short_open_interest 该会员持空单 int
short_open_interest_chg 该会员持空单变化量 int
symbol 标的合约 string
var 品种 string
date 日期 string YYYYMMDD
"""
date = cons.convert_date(date) if date is not None else datetime.date.today()
if date < datetime.date(2005, 5, 9):
print("czce数据源开始日期为20050509，跳过")
return {}
if date.strftime('%Y%m%d') not in calendar:
warnings.warn('%s非交易日' % date.strftime('%Y%m%d'))
return {}
if date <= datetime.date(2010, 8, 25):
url = cons.CZCE_VOL_RANK_URL_1 % (date.strftime('%Y%m%d'))
data = _czce_df_read(url, skip_rows=0)
r = requests_link(url, 'utf-8')
r.encoding = 'utf-8'
soup = BeautifulSoup(r.text, 'lxml', from_encoding="gb2312")
symbols = []
for link in soup.find_all('b'):
strings = (str(link).split(' '))
if len(strings) > 5:
try:
symbol = chinese_to_english(strings[4])
except:
symbol = strings[4]
symbols.append(symbol)
big_dict = {}
for i in range(len(symbols)):
symbol = symbols[i]
table_cut = data[i + 2]
table_cut.columns = rank_columns
table_cut = table_cut.iloc[:-1, :]
table_cut.loc[:, 'rank'] = table_cut.index
table_cut.loc['合计', 'rank'] = 999
table_cut.loc['合计', ['vol_party_name', 'long_party_name', 'short_party_name']] = None
table_cut.loc[:, 'symbol'] = symbol
table_cut.loc[:, 'variety'] = symbol_varieties(symbol)
table_cut[intColumns] = table_cut[intColumns].fillna(0)
table_cut[intColumns] = table_cut[intColumns].astype(str)
table_cut[intColumns] = table_cut[intColumns].applymap(lambda x: x.replace(',', ''))
table_cut = table_cut.applymap(lambda x: 0 if x == '-' else x)
table_cut[intColumns] = table_cut[intColumns].astype(float)
table_cut[intColumns] = table_cut[intColumns].astype(int)
big_dict[symbol] = table_cut.reset_index(drop=True)
return big_dict
elif date <= datetime.date(2015, 11, 11):
url = cons.CZCE_VOL_RANK_URL_2 % (date.year, date.strftime('%Y%m%d'))
data = _czce_df_read(url, skip_rows=1)[1]
elif date < datetime.date(2017, 12, 28):
url = cons.CZCE_VOL_RANK_URL_3 % (date.year, date.strftime('%Y%m%d'))
data = _czce_df_read(url, skip_rows=1)[0]
else:
url = cons.CZCE_VOL_RANK_URL_3 % (date.year, date.strftime('%Y%m%d'))
data = _czce_df_read(url, skip_rows=0)[0]
if len(data.columns) < 6:
return {}
table = pd.DataFrame(data.iloc[:, :9])
table.columns = rank_columns
table.loc[:, 'rank'] = table.index
table[intColumns] = table[intColumns].astype(str)
table[intColumns] = table[intColumns].applymap(lambda x: x.replace(',', ''))
table = table.applymap(lambda x: 0 if x == '-' else x)
indexes = [i for i in table.index if '合约' in i or '品种' in i]
indexes.insert(0, 0)
big_dict = {}
for i in range(len(indexes)):
if indexes[i] == 0:
table_cut = table.loc[:indexes[i + 1], :]
string = table_cut.index.name
elif i < len(indexes) - 1:
table_cut = table.loc[indexes[i]:indexes[i + 1], :]
string = table_cut.index[0]
else:
table_cut = table.loc[indexes[i]:, :]
string = table_cut.index[0]
if 'PTA' in string:
symbol = 'TA'
else:
try:
symbol = chinese_to_english(find_chinese(re.compile(r'：(.*) ').findall(string)[0]))
except:
symbol = re.compile(r'：(.*) ').findall(string)[0]
var = symbol_varieties(symbol)
if var in vars_list:
table_cut = table_cut.dropna(how='any').iloc[1:, :]
table_cut = table_cut.loc[[x for x in table_cut.index if x in [str(i) for i in range(21)]], :]
table_cut = _table_cut_cal(table_cut, symbol)
big_dict[symbol] = table_cut.reset_index(drop=True)
return big_dict
def get_shfe_rank_table(date=None, vars_list=cons.contract_symbols):
"""
上海期货交易所前 20 会员持仓排名数据明细
注：该交易所只公布每个品种内部的标的排名，没有公布品种的总排名
数据从20020107开始，每交易日16:30左右更新数据
:param date: 日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
:param vars_list: 合约品种如RB、AL等列表 为空时为所有商品
:return: pd.DataFrame
rank 排名 int
vol_party_name 成交量排序的当前名次会员 string(中文)
vol 该会员成交量 int
vol_chg 该会员成交量变化量 int
long_party_name 持多单排序的当前名次会员 string(中文)
long_open_interest 该会员持多单 int
long_open_interest_chg 该会员持多单变化量 int
short_party_name 持空单排序的当前名次会员 string(中文)
short_open_interest 该会员持空单 int
short_open_interest_chg 该会员持空单变化量 int
symbol 标的合约 string
var 品种 string
date 日期 string YYYYMMDD
"""
date = cons.convert_date(date) if date is not None else datetime.date.today()
if date < datetime.date(2002, 1, 7):
print("shfe数据源开始日期为20020107，跳过")
return {}
if date.strftime('%Y%m%d') not in calendar:
warnings.warn('%s非交易日' % date.strftime('%Y%m%d'))
return {}
url = cons.SHFE_VOL_RANK_URL % (date.strftime('%Y%m%d'))
r = requests_link(url, 'utf-8')
try:
context = json.loads(r.text)
except:
return {}
df = pd.DataFrame(context['o_cursor'])
df = df.rename(
columns={'CJ1': 'vol', 'CJ1_CHG': 'vol_chg', 'CJ2': 'long_open_interest', 'CJ2_CHG': 'long_open_interest_chg',
'CJ3': 'short_open_interest',
'CJ3_CHG': 'short_open_interest_chg', 'PARTICIPANTABBR1': 'vol_party_name',
'PARTICIPANTABBR2': 'long_party_name',
'PARTICIPANTABBR3': 'short_party_name', 'PRODUCTNAME': 'product1', 'RANK': 'rank',
'INSTRUMENTID': 'symbol', 'PRODUCTSORTNO': 'product2'})
if len(df.columns) < 3:
return {}
df = df.applymap(lambda x: x.strip() if isinstance(x, str) else x)
df = df.applymap(lambda x: None if x == '' else x)
df['variety'] = df['symbol'].apply(lambda x: symbol_varieties(x))
df = df[df['rank'] > 0]
for col in ['PARTICIPANTID1', 'PARTICIPANTID2', 'PARTICIPANTID3', 'product1', 'product2']:
try:
del df[col]
except:
pass
get_vars = [var for var in vars_list if var in df['variety'].tolist()]
big_dict = {}
for var in get_vars:
df_var = df[df['variety'] == var]
for symbol in set(df_var['symbol']):
df_symbol = df_var[df_var['symbol'] == symbol]
big_dict[symbol] = df_symbol.reset_index(drop=True)
return big_dict