def get_futures_daily(start_day=None,
end_day=None,
market="CFFEX",
index_bar=False):
"""
获取交易所日交易数据
Parameters
------
start_day: 开始日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
end_day: 结束数据 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
market: 'CFFEX' 中金所, 'CZCE' 郑商所, 'SHFE' 上期所, 'DCE' 大商所 之一。默认为中金所
index_bar: bool 是否合成指数K线
Return
-------
DataFrame
中金所日交易数据(DataFrame):
symbol 合约代码
date 日期
open 开盘价
high 最高价
low 最低价
close 收盘价
volume 成交量
open_interest 持仓量
turnover 成交额
settle 结算价
pre_settle 前结算价
variety 合约类别
或 None(给定日期没有交易数据)
"""
if market.upper() == "CFFEX":
f = get_cffex_daily
elif market.upper() == "CZCE":
f = get_czce_daily
elif market.upper() == "SHFE":
f = get_shfe_daily
elif market.upper() == "DCE":
f = get_dce_daily
else:
print("Invalid Market Symbol")
return
start_day = (cons.convert_date(start_day)
if start_day is not None else datetime.date.today())
end_day = (cons.convert_date(end_day)
if end_day is not None else cons.convert_date(
cons.get_latest_data_date(datetime.datetime.now())))
df_list = list()
while start_day <= end_day:
df = f(start_day)
if df is not None:
df_list.append(df)
if index_bar:
df_list.append(get_futures_index(df))
start_day += datetime.timedelta(days=1)
if len(df_list) > 0:
return pd.concat(df_list).reset_index(drop=True)
def get_czce_receipt_1(date: str = None,
vars_list: List = cons.contract_symbols):
"""
抓取郑州商品交易所注册仓单数据
适用20080222至20100824(包括)
:param date: 开始日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
:type date: str
:param vars_list: list
:type vars_list: 合约品种如CF、TA等列表 为空时为所有商品
:return: 展期收益率数据
:rtype: pandas.DataFrame
var 商品品种 string
receipt 仓单数 int
date 日期 string YYYYMMDD
"""
date = cons.convert_date(date).strftime(
'%Y%m%d') if date is not None else datetime.date.today()
if date not in calendar:
warnings.warn('%s非交易日' % date.strftime('%Y%m%d'))
return None
if date == '20090820':
return pd.DataFrame()
url = cons.CZCE_RECEIPT_URL_1 % date
r = requests_link(url, encoding='utf-8')
context = r.text
data = pd.read_html(context)[1]
records = pd.DataFrame()
indexes = [x for x in data.index if '品种：' in str(data[0].tolist()[x])]
ends = [x for x in data.index if '总计' in str(data[0].tolist()[x])]
for i in list(range(len(indexes))):
if i != len(indexes) - 1:
data_cut = data.loc[indexes[i]:ends[i], :]
data_cut = data_cut.fillna(method='pad')
else:
data_cut = data.loc[indexes[i]:, :]
data_cut = data_cut.fillna(method='pad')
if 'PTA' in data_cut[0].tolist()[0]:
var = 'TA'
else:
var = chinese_to_english(
re.sub(r'[A-Z]+', '', data_cut[0].tolist()[0][3:]))
if var == 'CF':
receipt = data_cut[6].tolist()[-1]
receipt_chg = data_cut[7].tolist()[-1]
else:
receipt = data_cut[5].tolist()[-1]
receipt_chg = data_cut[6].tolist()[-1]
data_dict = {
'var': var,
'receipt': int(receipt),
'receipt_chg': int(receipt_chg),
'date': date
}
records = records.append(pd.DataFrame(data_dict, index=[0]))
if len(records.index) != 0:
records.index = records['var']
vars_in_market = [i for i in vars_list if i in records.index]
records = records.loc[vars_in_market, :]
return records.reset_index(drop=True)
def get_czce_receipt_2(date: str = None,
vars_list: List = cons.contract_symbols):
"""
抓取郑州商品交易所注册仓单数据
适用20100825(包括)至20151111(包括)
Parameters
------
date: 开始日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
vars_list: 合约品种如CF、TA等列表 为空时为所有商品
Return
-------
DataFrame:
展期收益率数据(DataFrame):
var 商品品种 string
receipt 仓单数 int
date 日期 string YYYYMMDD
"""
date = cons.convert_date(date).strftime(
'%Y%m%d') if date is not None else datetime.date.today()
if date not in calendar:
warnings.warn('%s非交易日' % date.strftime('%Y%m%d'))
return None
url = cons.CZCE_RECEIPT_URL_2 % (date[:4], date)
r = requests.get(url)
r.encoding = 'utf-8'
data = pd.read_html(r.text)[3:]
records = pd.DataFrame()
for data_cut in data:
if len(data_cut.columns) > 3:
last_indexes = [
x for x in data_cut.index
if '注：' in str(data_cut[0].tolist()[x])
]
if len(last_indexes) > 0:
last_index = last_indexes[0] - 1
data_cut = data_cut.loc[:last_index, :]
if 'PTA' in data_cut[0].tolist()[0]:
var = 'TA'
else:
strings = data_cut[0].tolist()[0]
string = strings.split(' ')[0][3:]
var = chinese_to_english(re.sub(r'[A-Z]+', '', string))
data_cut.columns = data_cut.T[1].tolist()
receipt = data_cut['仓单数量'].tolist()[-1]
receipt_chg = data_cut['当日增减'].tolist()[-1]
data_dict = {
'var': var,
'receipt': int(receipt),
'receipt_chg': int(receipt_chg),
'date': date
}
records = records.append(pd.DataFrame(data_dict, index=[0]))
if len(records.index) != 0:
records.index = records['var']
vars_in_market = [i for i in vars_list if i in records.index]
records = records.loc[vars_in_market, :]
return records.reset_index(drop=True)
def get_spot_price(date=None, vars_list=cons.contract_symbols):
"""
获取某个交易日大宗商品现货价格及相应基差
:param date: 开始日期 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: pandas.DataFrame
展期收益率数据:
var 商品品种 string
sp 现货价格 float
near_symbol 临近交割合约 string
near_price 临近交割合约结算价 float
dom_symbol 主力合约 string
dom_price 主力合约结算价 float
near_basis 临近交割合约相对现货的基差 float
dom_basis 主力合约相对现货的基差 float
near_basis_rate 临近交割合约相对现货的基差率 float
dom_basis_rate 主力合约相对现货的基差率 float
date 日期 string YYYYMMDD
"""
date = cons.convert_date(
date) if date is not None else datetime.date.today()
if date < datetime.date(2011, 1, 4):
raise Exception("数据源开始日期为 20110104, 请将获取数据时间点设置在 20110104 后")
if date.strftime("%Y%m%d") not in calendar:
warnings.warn(f"{date.strftime('%Y%m%d')}非交易日")
return None
u1 = cons.SYS_SPOT_PRICE_LATEST_URL
u2 = cons.SYS_SPOT_PRICE_URL.format(date.strftime("%Y-%m-%d"))
i = 1
while True:
for url in [u2, u1]:
try:
r = pandas_read_html_link(url)
string = r[0].loc[1, 1]
news = "".join(re.findall(r"[0-9]", string))
if news[3:11] == date.strftime("%Y%m%d"):
records = _check_information(r[1], date)
records.index = records["symbol"]
var_list_in_market = [
i for i in vars_list if i in records.index
]
temp_df = records.loc[var_list_in_market, :]
temp_df.reset_index(drop=True, inplace=True)
return temp_df
else:
time.sleep(3)
except:
print(
f"{date.strftime('%Y-%m-%d')}日生意社数据连接失败，第{str(i)}次尝试，最多5次")
i += 1
if i > 5:
print(
f"{date.strftime('%Y-%m-%d')}日生意社数据连接失败, 如果当前交易日是 2018-09-12, 由于生意社源数据缺失, 无法访问, 否则为重复访问已超过5次，您的地址被网站墙了，请保存好返回数据，稍后从该日期起重试"
)
return False
def get_shfe_receipt_1(date: str = None,
vars_list: List = cons.contract_symbols):
"""
抓取上海期货交易所注册仓单数据, 适用20081006至20140518(包括) 20100126、20101029日期交易所格式混乱，直接回复脚本中DataFrame, 20100416、20130821日期交易所数据丢失
:param date: 开始日期 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(DataFrame):
var 商品品种 string
receipt 仓单数 int
date 日期 string YYYYMMDD
"""
date = cons.convert_date(date).strftime(
'%Y%m%d') if date is not None else datetime.date.today()
if date not in calendar:
warnings.warn(f"{date.strftime('%Y%m%d')}非交易日")
return None
if date == '20100126':
shfe_20100126['date'] = date
return shfe_20100126
elif date == '20101029':
shfe_20101029['date'] = date
return shfe_20101029
elif date in ['20100416', '20130821']:
return warnings.warn('20100416、20130821日期交易所数据丢失')
else:
var_list = [
'天然橡胶', '沥青仓库', '沥青厂库', '热轧卷板', '燃料油', '白银', '线材', '螺纹钢', '铅', '铜',
'铝', '锌', '黄金', '锡', '镍'
]
url = cons.SHFE_RECEIPT_URL_1 % date
data = pandas_read_html_link(url)[0]
indexes = [x for x in data.index if (data[0].tolist()[x] in var_list)]
last_index = [
x for x in data.index if '注' in str(data[0].tolist()[x])
][0] - 1
records = pd.DataFrame()
for i in list(range(len(indexes))):
if i != len(indexes) - 1:
data_cut = data.loc[indexes[i]:indexes[i + 1] - 1, :]
else:
data_cut = data.loc[indexes[i]:last_index, :]
data_cut = data_cut.fillna(method='pad')
data_dict = dict()
data_dict['var'] = chinese_to_english(data_cut[0].tolist()[0])
data_dict['receipt'] = int(data_cut[2].tolist()[-1])
data_dict['receipt_chg'] = int(data_cut[3].tolist()[-1])
data_dict['date'] = date
records = records.append(pd.DataFrame(data_dict, index=[0]))
if len(records.index) != 0:
records.index = records['var']
vars_in_market = [i for i in vars_list if i in records.index]
records = records.loc[vars_in_market, :]
return records.reset_index(drop=True)
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 get_rank_sum_daily(start_day=None, end_day=None, vars_list=cons.contract_symbols):
"""
采集四个期货交易所前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 start_day: 开始日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
:param end_day: 结束数据 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(DataFrame):
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
"""
start_day = cons.convert_date(start_day) if start_day is not None else datetime.date.today()
end_day = cons.convert_date(end_day) if end_day is not None else cons.convert_date(
cons.get_latest_data_date(datetime.datetime.now()))
records = pd.DataFrame()
while start_day <= end_day:
print(start_day)
if start_day.strftime('%Y%m%d') in calendar:
data = get_rank_sum(start_day, vars_list)
if data is False:
print(f"{start_day.strftime('%Y-%m-%d')}日交易所数据连接失败，已超过20次，您的地址被网站墙了，请保存好返回数据，稍后从该日期起重试")
return records.reset_index(drop=True)
records = records.append(data)
else:
warnings.warn(f"{start_day.strftime('%Y%m%d')}非交易日")
start_day += datetime.timedelta(days=1)
return records.reset_index(drop=True)
def get_spot_price_daily(start_day=None,
end_day=None,
vars_list=cons.contract_symbols):
"""
获取某段时间大宗商品现货价格及相应基差
:param start_day: str 开始日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象; 默认为当天
:param end_day: str 结束数据 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象; 默认为当天
:param vars_list: list 合约品种如 [RB, AL]; 默认参数为所有商品
:return: pandas.DataFrame
展期收益率数据:
var 商品品种 string
sp 现货价格 float
near_symbol 临近交割合约 string
near_price 临近交割合约结算价 float
dom_symbol 主力合约 string
dom_price 主力合约结算价 float
near_basis 临近交割合约相对现货的基差 float
dom_basis 主力合约相对现货的基差 float
near_basis_rate 临近交割合约相对现货的基差率 float
dom_basis_rate 主力合约相对现货的基差率 float
date 日期 string YYYYMMDD
"""
start_day = (cons.convert_date(start_day)
if start_day is not None else datetime.date.today())
end_day = (cons.convert_date(end_day)
if end_day is not None else cons.convert_date(
cons.get_latest_data_date(datetime.datetime.now())))
df_list = []
while start_day <= end_day:
print(start_day)
temp_df = get_spot_price(start_day, vars_list)
if temp_df is False:
return pd.concat(df_list).reset_index(drop=True)
elif temp_df is not None:
df_list.append(temp_df)
start_day += datetime.timedelta(days=1)
if len(df_list) > 0:
temp_df = pd.concat(df_list)
temp_df.reset_index(drop=True, inplace=True)
return temp_df
def get_dce_receipt(date: str = None,
symbol_list: List = cons.contract_symbols):
"""
完成
采集大连商品交易所注册仓单数据
:param date: format 开始日期: YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象, 为空时为当天
:param symbol_list: 合约品种如 RB, AL等列表, 为空时为所有商品数据从 20060106开始，每周五更新仓单数据。直到20090407起，每交易日都更新仓单数据
:return: pd.DataFrame
展期收益率数据(DataFrame):
var 商品品种 string
receipt 仓单数 int
date 日期 string YYYYMMDD
"""
if not isinstance(symbol_list, list):
return warnings.warn(f"symbol_list: 必须是列表")
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(f"{date.strftime('%Y%m%d')}非交易日")
return None
payload = {
"weekQuotes.variety": "all",
"year": date.year,
"month": date.month - 1, # 网站月份描述少 1 个月, 属于网站问题
"day": date.day
}
data = pandas_read_html_link(cons.DCE_RECEIPT_URL,
method="post",
data=payload,
headers=cons.dce_headers)[0]
records = pd.DataFrame()
for x in data.to_dict(orient='records'):
if isinstance(x['品种'], str):
if x['品种'][-2:] == '小计':
var = x['品种'][:-2]
temp_data = {
'var': chinese_to_english(var),
'receipt': int(x['今日仓单量']),
'receipt_chg': int(x['增减']),
'date': date.strftime('%Y%m%d')
}
records = records.append(pd.DataFrame(temp_data, index=[0]))
if len(records.index) != 0:
records.index = records['var']
vars_in_market = [i for i in symbol_list if i in records.index]
records = records.loc[vars_in_market, :]
return records.reset_index(drop=True)
def get_shfe_receipt_2(date: str = None,
vars_list: List = cons.contract_symbols):
"""
抓取上海商品交易所注册仓单数据
适用20140519(包括)至今
Parameters
------
date: 开始日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
vars_list: 合约品种如RB、AL等列表 为空时为所有商品
Return
-------
DataFrame:
展期收益率数据(DataFrame):
var 商品品种 string
receipt 仓单数 int
date 日期 string YYYYMMDD
"""
date = cons.convert_date(date).strftime(
'%Y%m%d') if date is not None else datetime.date.today()
if date not in calendar:
warnings.warn('%s非交易日' % date.strftime('%Y%m%d'))
return None
url = cons.SHFE_RECEIPT_URL_2 % date
r = requests_link(url, encoding='utf-8')
try:
context = r.json()
except:
return pd.DataFrame()
data = pd.DataFrame(context['o_cursor'])
if len(data.columns) < 1:
return pd.DataFrame()
records = pd.DataFrame()
for var in set(data['VARNAME'].tolist()):
data_cut = data[data['VARNAME'] == var]
data_dict = {
'var': chinese_to_english(re.sub(r"\W|[a-zA-Z]", "", var)),
'receipt': int(data_cut['WRTWGHTS'].tolist()[-1]),
'receipt_chg': int(data_cut['WRTCHANGE'].tolist()[-1]),
'date': date
}
records = records.append(pd.DataFrame(data_dict, index=[0]))
if len(records.index) != 0:
records.index = records['var']
vars_in_market = [i for i in vars_list if i in records.index]
records = records.loc[vars_in_market, :]
return records.reset_index(drop=True)
def get_cffex_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：该交易所既公布品种排名，也公布标的排名
: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, 数据从20100416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
"""
vars_list = [i for i in vars_list if i in cons.market_exchange_symbols['cffex']]
date = cons.convert_date(date) if date is not None else datetime.date.today()
if date < datetime.date(2010, 4, 16):
print(Exception("cffex数据源开始日期为20100416，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 {}
big_dict = {}
for var in vars_list:
url = cons.CFFEX_VOL_RANK_URL % (date.strftime('%Y%m'), date.strftime('%d'), var)
r = requests_link(url, encoding='gbk')
if not r:
return False
if '网页错误' not in r.text:
table = pd.read_csv(StringIO(r.text.split('\n交易日,')[1]))
table = table.dropna(how='any')
table = table.applymap(lambda x: x.strip() if isinstance(x, str) else x)
for symbol in set(table['合约']):
table_cut = table[table['合约'] == symbol]
table_cut.columns = ['symbol', 'rank'] + rank_columns
table_cut = _table_cut_cal(pd.DataFrame(table_cut), symbol)
big_dict[symbol] = table_cut.reset_index(drop=True)
return big_dict
def get_shfe_v_wap(date=None):
"""
获取上期所日成交均价数据
Parameters
------
date: 日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
Return
-------
DataFrame
郑商所日交易数据(DataFrame):
symbol 合约代码
date 日期
time_range v_wap时段，分09:00-10:15和09:00-15:00两类
v_wap 加权平均成交均价
或 None(给定日期没有数据)
"""
day = cons.convert_date(
date) if date is not None else datetime.date.today()
if day.strftime("%Y%m%d") not in calendar:
warnings.warn("%s非交易日" % day.strftime("%Y%m%d"))
return None
try:
json_data = json.loads(
requests_link(
cons.SHFE_V_WAP_URL % (day.strftime("%Y%m%d")),
headers=cons.headers,
encoding="utf-8",
).text)
except requests.HTTPError as reason:
if reason.response not in [404, 403]:
print(cons.SHFE_DAILY_URL % (day.strftime("%Y%m%d")), reason)
return None
if len(json_data["o_currefprice"]) == 0:
return None
try:
df = pd.DataFrame(json_data["o_currefprice"])
df["INSTRUMENTID"] = df["INSTRUMENTID"].str.strip()
df[":B1"].astype("int16")
return df.rename(columns=cons.SHFE_V_WAP_COLUMNS)[list(
cons.SHFE_V_WAP_COLUMNS.values())]
except:
return None
def get_ine_daily(date="20200312"):
"""
上海国际能源交易中心-日频率-量价数据
上海国际能源交易中心: 原油期货(上市时间: 20180326); 20号胶期货(上市时间: 20190812)
trade_price: http://www.ine.cn/statements/daily/?paramid=kx
trade_note: http://www.ine.cn/data/datanote.dat
:param date: 日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象，默认为当前交易日
:type date: str or datetime.date
:return: 上海国际能源交易中心-日频率-量价数据
:rtype: pandas.DataFrame or None
"""
day = cons.convert_date(
date) if date is not None else datetime.date.today()
if day.strftime("%Y%m%d") not in calendar:
warnings.warn(f"{day.strftime('%Y%m%d')}非交易日")
return None
url = f"http://www.ine.cn/data/dailydata/kx/kx{day.strftime('%Y%m%d')}.dat"
r = requests.get(url)
result_df = pd.DataFrame()
temp_df = pd.DataFrame(r.json()["o_curinstrument"]).iloc[:-1, :]
temp_df = temp_df[temp_df["DELIVERYMONTH"] != "小计"]
result_df["symbol"] = temp_df["PRODUCTID"].str.upper().str.split(
"_", expand=True)[0] + temp_df["DELIVERYMONTH"]
result_df["date"] = day.strftime("%Y%m%d")
result_df["open"] = temp_df["OPENPRICE"]
result_df["high"] = temp_df["HIGHESTPRICE"]
result_df["low"] = temp_df["LOWESTPRICE"]
result_df["close"] = temp_df["CLOSEPRICE"]
result_df["volume"] = temp_df["VOLUME"]
result_df["open_interest"] = temp_df["OPENINTEREST"]
result_df["turnover"] = 0
result_df["settle"] = temp_df["SETTLEMENTPRICE"]
result_df["pre_settle"] = temp_df["PRESETTLEMENTPRICE"]
result_df["variety"] = temp_df["PRODUCTID"].str.upper().str.split(
"_", expand=True)[0]
return result_df
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_dce_rank_table(date=None, vars_list=cons.contract_symbols):
"""
大连商品交易所前 20 会员持仓排名数据明细
注: 该交易所既公布品种排名, 也公布标的排名
: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, 数据从20060104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(2006, 1, 4):
print(Exception("dce数据源开始日期为20060104，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 {}
vars_list = [i for i in vars_list if i in cons.market_exchange_symbols['dce']]
big_dict = {}
for var in vars_list:
url = cons.DCE_VOL_RANK_URL % (var.lower(), var.lower(), date.year, date.month - 1, date.day)
list_60_name = []
list_60 = []
list_60_chg = []
rank = []
texts = requests_link(url).content.splitlines()
if not texts:
return False
if len(texts) > 30:
for text in texts:
line = text.decode("utf-8")
string_list = line.split()
try:
if int(string_list[0]) <= 20:
list_60_name.append(string_list[1])
list_60.append(string_list[2])
list_60_chg.append(string_list[3])
rank.append(string_list[0])
except:
pass
table_cut = pd.DataFrame({'rank': rank[0:20],
'vol_party_name': list_60_name[0:20],
'vol': list_60[0:20],
'vol_chg': list_60_chg[0:20],
'long_party_name': list_60_name[20:40],
'long_open_interest': list_60[20:40],
'long_open_interest_chg': list_60_chg[20:40],
'short_party_name': list_60_name[40:60],
'short_open_interest': list_60[40:60],
'short_open_interest_chg': list_60_chg[40:60]
})
table_cut = table_cut.applymap(lambda x: x.replace(',', ''))
table_cut = _table_cut_cal(table_cut, var)
big_dict[var] = table_cut.reset_index(drop=True)
return big_dict
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_shfe_daily(date=None):
"""
上海期货交易所-日频率-量价数据
:param date: 日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象, 默认为当前交易日
:type date: str or datetime.date
:return: 上海期货交易所-日频率-量价数据
:rtype: pandas.DataFrame or None
上期所日交易数据(DataFrame):
symbol 合约代码
date 日期
open 开盘价
high 最高价
low 最低价
close 收盘价
volume 成交量
open_interest 持仓量
turnover 成交额
settle 结算价
pre_settle 前结算价
variety 合约类别
或 None(给定交易日没有交易数据)
"""
day = cons.convert_date(
date) if date is not None else datetime.date.today()
if day.strftime("%Y%m%d") not in calendar:
warnings.warn("%s非交易日" % day.strftime("%Y%m%d"))
return None
try:
json_data = json.loads(
requests_link(
cons.SHFE_DAILY_URL % (day.strftime("%Y%m%d")),
headers=cons.shfe_headers,
).text)
except requests.HTTPError as reason:
if reason.response != 404:
print(cons.SHFE_DAILY_URL % (day.strftime("%Y%m%d")), reason)
return
if len(json_data["o_curinstrument"]) == 0:
return
df = pd.DataFrame([
row for row in json_data["o_curinstrument"] if
row["DELIVERYMONTH"] not in ["小计", "合计"] and row["DELIVERYMONTH"] != ""
])
df["variety"] = df.PRODUCTID.str.slice(0, -6).str.upper()
df["symbol"] = df["variety"] + df["DELIVERYMONTH"]
df["date"] = day.strftime("%Y%m%d")
v_wap_df = get_shfe_v_wap(day)
if v_wap_df is not None:
df = pd.merge(
df,
v_wap_df[v_wap_df.time_range == "9:00-15:00"],
on=["date", "symbol"],
how="left",
)
df["turnover"] = df.v_wap * df.VOLUME
else:
df["VOLUME"] = df["VOLUME"].apply(lambda x: 0 if x == "" else x)
df["turnover"] = df["VOLUME"] * df["SETTLEMENTPRICE"]
df.rename(columns=cons.SHFE_COLUMNS, inplace=True)
return df[cons.OUTPUT_COLUMNS]
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
def get_czce_daily(date=None):
"""
郑州商品交易所-日频率-量价数据
:param date: 日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象，默认为当前交易日
:type date: str or datetime.date
:return: 郑州商品交易所-日频率-量价数据
:rtype: pandas.DataFrame or None
郑商所每日期货交易数据:
symbol 合约代码
date 日期
open 开盘价
high 最高价
low 最低价
close 收盘价
volume 成交量
open_interest 持仓量
turnover 成交额
settle 结算价
pre_settle 前结算价
variety 合约类别
或
郑商所每日期权交易数据
symbol 合约代码
date 日期
open 开盘价
high 最高价
low 最低价
close 收盘价
pre_settle 前结算价
settle 结算价
delta 对冲值
volume 成交量
open_interest 持仓量
oi_change 持仓变化
turnover 成交额
implied_volatility 隐含波动率
exercise_volume 行权量
variety 合约类别
None(类型错误或给定日期没有交易数据)
"""
day = cons.convert_date(
date) if date is not None else datetime.date.today()
if day.strftime("%Y%m%d") not in calendar:
warnings.warn(f"{day.strftime('%Y%m%d')}非交易日")
return None
if day > datetime.date(2010, 8, 24):
if day > datetime.date(2015, 9, 19):
u = cons.CZCE_DAILY_URL_3
url = u % (day.strftime("%Y"), day.strftime("%Y%m%d"))
elif day < datetime.date(2015, 9, 19):
u = cons.CZCE_DAILY_URL_2
url = u % (day.strftime("%Y"), day.strftime("%Y%m%d"))
listed_columns = cons.CZCE_COLUMNS
output_columns = cons.OUTPUT_COLUMNS
try:
r = requests.get(url)
html = r.text
except requests.exceptions.HTTPError as reason:
if reason.response.status_code != 404:
print(
cons.CZCE_DAILY_URL_3 %
(day.strftime("%Y"), day.strftime("%Y%m%d")),
reason,
)
return
if html.find("您的访问出错了") >= 0 or html.find("无期权每日行情交易记录") >= 0:
return
html = [
i.replace(" ", "").split("|") for i in html.split("\n")[:-4]
if i[0][0] != u"小"
]
if day > datetime.date(2015, 9, 19):
if html[1][0] not in ["品种月份", u"品种代码"]:
return
dict_data = list()
day_const = int(day.strftime("%Y%m%d"))
for row in html[2:]:
m = cons.FUTURES_SYMBOL_PATTERN.match(row[0])
if not m:
continue
row_dict = {
"date": day_const,
"symbol": row[0],
"variety": m.group(1)
}
for i, field in enumerate(listed_columns):
if row[i + 1] == "\r":
row_dict[field] = 0.0
elif field in [
"volume",
"open_interest",
"oi_chg",
"exercise_volume",
]:
row[i + 1] = row[i + 1].replace(",", "")
row_dict[field] = int(row[i + 1])
else:
row[i + 1] = row[i + 1].replace(",", "")
row_dict[field] = float(row[i + 1])
dict_data.append(row_dict)
return pd.DataFrame(dict_data)[output_columns]
elif day < datetime.date(2015, 9, 19):
dict_data = list()
day_const = int(day.strftime("%Y%m%d"))
for row in html[1:]:
row = row[0].split(",")
m = cons.FUTURES_SYMBOL_PATTERN.match(row[0])
if not m:
continue
row_dict = {
"date": day_const,
"symbol": row[0],
"variety": m.group(1)
}
for i, field in enumerate(listed_columns):
if row[i + 1] == "\r":
row_dict[field] = 0.0
elif field in [
"volume",
"open_interest",
"oi_chg",
"exercise_volume",
]:
row_dict[field] = int(float(row[i + 1]))
else:
row_dict[field] = float(row[i + 1])
dict_data.append(row_dict)
return pd.DataFrame(dict_data)[output_columns]
if day <= datetime.date(2010, 8, 24):
u = cons.CZCE_DAILY_URL_1
url = u % day.strftime("%Y%m%d")
listed_columns = cons.CZCE_COLUMNS_2
output_columns = cons.OUTPUT_COLUMNS
df = pd.read_html(url)[1].dropna(how="any")
dict_data = list()
day_const = int(day.strftime("%Y%m%d"))
for row in df.to_dict(orient="records")[1:]:
m = cons.FUTURES_SYMBOL_PATTERN.match(row[0])
if not m:
continue
row_dict = {
"date": day_const,
"symbol": row[0],
"variety": m.group(1)
}
for i, field in enumerate(listed_columns):
if row[i + 1] == "\r":
row_dict[field] = 0.0
elif field in [
"volume", "open_interest", "oi_chg", "exercise_volume"
]:
row_dict[field] = int(row[i + 1])
else:
row_dict[field] = float(row[i + 1])
dict_data.append(row_dict)
return pd.DataFrame(dict_data)[output_columns]
def get_roll_yield_bar(type_method="symbol",
var="RB",
date=None,
start_day=None,
end_day=None,
plot=False):
"""
获取展期收益率
:param type_method: 'symbol'：获取某天某品种所有交割月合约的收盘价, 'var'：获取某天所有品种两个主力合约的展期收益率（展期收益率横截面）, ‘date’：获取某品种每天的两个主力合约的展期收益率（展期收益率时间序列）
:param var: 合约品种如RB、AL等
:param date: 某一天日期 format： YYYYMMDD
:param start_day: 开始日期 format：YYYYMMDD
:param end_day: 结束数据 format：YYYYMMDD
:param plot: True or False作图
:return: pd.DataFrame
展期收益率数据(DataFrame):
ry 展期收益率
index 日期或品种
"""
date = cons.convert_date(
date) if date is not None else datetime.date.today()
start_day = (cons.convert_date(start_day)
if start_day is not None else datetime.date.today())
end_day = (cons.convert_date(end_day)
if end_day is not None else cons.convert_date(
cons.get_latest_data_date(datetime.datetime.now())))
if type_method == "symbol":
df = get_futures_daily(start_day=date,
end_day=date,
market=symbol_market(var))
df = df[df["variety"] == var]
if plot:
_plot_bar_2(df[["symbol", "close"]])
return df
if type_method == "var":
df = pd.DataFrame()
for market in ["dce", "cffex", "shfe", "czce"]:
df = df.append(
get_futures_daily(start_day=date, end_day=date, market=market))
var_list = list(set(df["variety"]))
df_l = pd.DataFrame()
for var in var_list:
ry = get_roll_yield(date, var, df=df)
if ry:
df_l = df_l.append(
pd.DataFrame([ry],
index=[var],
columns=["roll_yield", "near_by",
"deferred"]))
df_l["date"] = date
df_l = df_l.sort_values("roll_yield")
if plot:
_plot_bar(df_l["roll_yield"])
return df_l
if type_method == "date":
df_l = pd.DataFrame()
while start_day <= end_day:
try:
ry = get_roll_yield(start_day, var)
if ry:
df_l = df_l.append(
pd.DataFrame(
[ry],
index=[start_day],
columns=["roll_yield", "near_by", "deferred"],
))
except:
pass
start_day += datetime.timedelta(days=1)
if plot:
_plot(df_l["roll_yield"])
return df_l
def get_receipt(start_day: str = None,
end_day: str = None,
vars_list: List = cons.contract_symbols):
"""
获取大宗商品注册仓单数量
:param start_day: 开始日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
:type start_day: str
:param end_day: 结束数据 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
:type end_day: str
:param vars_list: 合约品种如RB、AL等列表 为空时为所有商品
:type vars_list: str
:return: 展期收益率数据
:rtype: pandas.DataFrame
"""
start_day = cons.convert_date(
start_day) if start_day is not None else datetime.date.today()
end_day = cons.convert_date(
end_day) if end_day is not None else cons.convert_date(
cons.get_latest_data_date(datetime.datetime.now()))
records = pd.DataFrame()
while start_day <= end_day:
if start_day.strftime('%Y%m%d') not in calendar:
warnings.warn(f"{start_day.strftime('%Y%m%d')}非交易日")
else:
print(start_day)
for market, market_vars in cons.market_exchange_symbols.items():
if market == 'dce':
if start_day >= datetime.date(2009, 4, 7):
f = get_dce_receipt
else:
print('20090407起，dce每交易日更新仓单数据')
f = None
elif market == 'shfe':
if datetime.date(2008, 10,
6) <= start_day <= datetime.date(
2014, 5, 16):
f = get_shfe_receipt_1
elif start_day > datetime.date(2014, 5, 16):
f = get_shfe_receipt_2
else:
f = None
print('20081006起，shfe每交易日更新仓单数据')
elif market == 'czce':
if datetime.date(2008, 3, 3) <= start_day <= datetime.date(
2010, 8, 24):
f = get_czce_receipt_1
elif datetime.date(2010, 8,
24) < start_day <= datetime.date(
2015, 11, 11):
f = get_czce_receipt_2
elif start_day > datetime.date(2015, 11, 11):
f = get_czce_receipt_3
else:
f = None
print('20080303起，czce每交易日更新仓单数据')
get_vars = [var for var in vars_list if var in market_vars]
if market != 'cffex' and get_vars != []:
if f is not None:
records = records.append(f(start_day, get_vars))
start_day += datetime.timedelta(days=1)
records.reset_index(drop=True, inplace=True)
if "MA" in records["var"].to_list():
replace_index = records[records["var"] == "MA"]["receipt"].astype(
str).str.split("0", expand=True)[0].index
records.loc[replace_index, "receipt"] = records[
records["var"] == "MA"]["receipt"].astype(str).str.split(
"0", expand=True)[0]
return records
def get_cffex_daily(date=None):
"""
中国金融期货交易所日交易数据
:param date: 日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象; 为空时为当天
:return: pandas.DataFrame
中国金融期货交易所日:
symbol 合约代码
date 日期
open 开盘价
high 最高价
low 最低价
close 收盘价
volume 成交量
open_interest 持仓量
turnover 成交额
settle 结算价
pre_settle 前结算价
variety 合约类别
或 None(给定日期没有交易数据)
"""
day = cons.convert_date(
date) if date is not None else datetime.date.today()
if day.strftime("%Y%m%d") not in calendar:
warnings.warn("%s非交易日" % day.strftime("%Y%m%d"))
return None
try:
html = requests_link(
cons.CFFEX_DAILY_URL.format(day.strftime("%Y%m"),
day.strftime("%d"),
day.strftime("%Y%m%d")),
encoding="gbk",
headers=cons.headers,
).text
except requests.exceptions.HTTPError as reason:
if reason.response != 404:
print(
cons.CFFEX_DAILY_URL %
(day.strftime("%Y%m"), day.strftime("%d"),
day.strftime("%Y%m%d")),
reason,
)
return
if html.find("网页错误") >= 0:
return
html = [
i.replace(" ", "").split(",") for i in html.split("\n")[:-2]
if i[0][0] != "小"
]
if html[0][0] != "合约代码":
return
dict_data = list()
day_const = day.strftime("%Y%m%d")
for row in html[1:]:
m = cons.FUTURES_SYMBOL_PATTERN.match(row[0])
if not m:
continue
row_dict = {"date": day_const, "symbol": row[0], "variety": m.group(1)}
for i, field in enumerate(cons.CFFEX_COLUMNS):
if row[i + 1] == "":
row_dict[field] = 0.0
elif field in ["volume", "open_interest", "oi_chg"]:
row_dict[field] = int(row[i + 1])
else:
try:
row_dict[field] = float(row[i + 1])
except:
pass
row_dict["pre_settle"] = row_dict["close"] - row_dict["change1"]
dict_data.append(row_dict)
return pd.DataFrame(dict_data)[cons.OUTPUT_COLUMNS]
def get_dce_daily(date=None, symbol_type="futures", retries=0):
"""
获取大连商品交易所日交易数据
Parameters
------
date: 日期 format：YYYY-MM-DD 或 YYYYMMDD 或 datetime.date对象 为空时为当天
symbol_type: 数据类型, 为'futures'期货 或 'option'期权二者之一
retries: int, 当前重试次数，达到3次则获取数据失败
Return
-------
DataFrame
大商所日交易数据(DataFrame):
symbol 合约代码
date 日期
open 开盘价
high 最高价
low 最低价
close 收盘价
volume 成交量
open_interest 持仓量
turnover 成交额
settle 结算价
pre_settle 前结算价
variety 合约类别
或
DataFrame
郑商所每日期权交易数据
symbol 合约代码
date 日期
open 开盘价
high 最高价
low 最低价
close 收盘价
pre_settle 前结算价
settle 结算价
delta 对冲值
volume 成交量
open_interest 持仓量
oi_change 持仓变化
turnover 成交额
implied_volatility 隐含波动率
exercise_volume 行权量
variety 合约类别
或 None(给定日期没有交易数据)
"""
day = cons.convert_date(
date) if date is not None else datetime.date.today()
if day.strftime("%Y%m%d") not in calendar:
warnings.warn("%s非交易日" % day.strftime("%Y%m%d"))
return None
if retries > 3:
print("maximum retires for DCE market data: ", day.strftime("%Y%m%d"))
return
if symbol_type == "futures":
url = (cons.DCE_DAILY_URL + "?" +
urllib.parse.urlencode({
"currDate": day.strftime("%Y%m%d"),
"year": day.strftime("%Y"),
"month": str(int(day.strftime("%m")) - 1),
"day": day.strftime("%d"),
}))
listed_columns = cons.DCE_COLUMNS
output_columns = cons.OUTPUT_COLUMNS
elif symbol_type == "option":
url = (cons.DCE_DAILY_URL + "?" +
urllib.parse.urlencode({
"currDate": day.strftime("%Y%m%d"),
"year": day.strftime("%Y"),
"month": str(int(day.strftime("%m")) - 1),
"day": day.strftime("%d"),
"dayQuotes.trade_type": "1",
}))
listed_columns = cons.DCE_OPTION_COLUMNS
output_columns = cons.OPTION_OUTPUT_COLUMNS
else:
print("invalid symbol_type :" + symbol_type +
', should be one of "futures" or "option"')
return
try:
response = requests_link(url, method="post", headers=cons.headers).text
except requests.exceptions.ContentDecodingError as reason:
return get_dce_daily(day, retries=retries + 1)
except requests.exceptions.HTTPError as reason:
if reason.response == 504:
return get_dce_daily(day, retries=retries + 1)
elif reason.response != 404:
print(cons.DCE_DAILY_URL, reason)
return
if "错误：您所请求的网址（URL）无法获取" in response:
return get_dce_daily(day, retries=retries + 1)
elif "暂无数据" in response:
return
data = BeautifulSoup(response, "html.parser").find_all("tr")
if len(data) == 0:
return
dict_data = list()
implied_data = list()
for i_data in data[1:]:
if "小计" in i_data.text or "总计" in i_data.text:
continue
x = i_data.find_all("td")
if symbol_type == "futures":
row_dict = {"variety": cons.DCE_MAP[x[0].text.strip()]}
row_dict["symbol"] = row_dict["variety"] + x[1].text.strip()
for i, field in enumerate(listed_columns):
field_content = x[i + 2].text.strip()
if "-" in field_content:
row_dict[field] = 0
elif field in ["volume", "open_interest"]:
row_dict[field] = int(field_content.replace(",", ""))
else:
row_dict[field] = float(field_content.replace(",", ""))
dict_data.append(row_dict)
elif len(x) == 16:
m = cons.FUTURES_SYMBOL_PATTERN.match(x[1].text.strip())
if not m:
continue
row_dict = {
"symbol": x[1].text.strip(),
"variety": m.group(1).upper(),
"contract_id": m.group(0),
}
for i, field in enumerate(listed_columns):
field_content = x[i + 2].text.strip()
if "-" in field_content:
row_dict[field] = 0
elif field in ["volume", "open_interest"]:
row_dict[field] = int(field_content.replace(",", ""))
else:
row_dict[field] = float(field_content.replace(",", ""))
dict_data.append(row_dict)
elif len(x) == 2:
implied_data.append({
"contract_id": x[0].text.strip(),
"implied_volatility": float(x[1].text.strip()),
})
df = pd.DataFrame(dict_data)
df["date"] = day.strftime("%Y%m%d")
if symbol_type == "futures":
return df[output_columns]
else:
return pd.merge(
df,
pd.DataFrame(implied_data),
on="contract_id",
how="left",
indicator=False,
)[output_columns]
