def fund_nav_daily(code: str) -> pd.DataFrame:
assert code[0] == 'f'
code = code[1:]
df = ak.fund_em_open_fund_info(code, indicator="单位净值走势")
# print(df)
df2 = ak.fund_em_open_fund_info(code, indicator="累计净值走势")
df = pd.merge(df, df2, on='净值日期', how='outer')
df2 = ak.fund_em_open_fund_info(code, indicator="同类排名百分比")
# print(df2)
df2 = df2.rename({'报告日期': '净值日期'}, axis=1)
df = pd.merge(df, df2, on='净值日期', how='outer')
df = df.rename(
{
'净值日期': '_id',
'单位净值': 'nav',
'日增长率': 'growth_rate',
'累计净值': 'cum_nav',
'同类型排名-每日近3月收益排名百分比': 'rank_3m'
},
axis=1)
if df.dtypes['nav'] != np.dtype('float64'):
df['nav'] = df['nav'].apply(lambda x: None
if x is None else float(x))
if df.dtypes['cum_nav'] != np.dtype('float64'):
df['cum_nav'] = df['cum_nav'].apply(lambda x: None
if x is None else float(x))
df['_id'] = pd.to_datetime(df['_id'])
print(df)
return df
def fund_list(tscode):
ts_name = tscode
df = ak.fund_em_open_fund_info(fund="%s" % ts_name, indicator="单位净值走势")
df.to_sql("%s" % ts_name, con=ak_engine, if_exists="append", index=False)
x = df['净值日期']
y = df['单位净值']
return x, y
def combine_data(self, start, end, mat: pd.DataFrame, batch=30):
df_list = []
for ix, row in list(mat.iterrows()):
sleep(1)
fund_code = row['fund_code']
fund_name = row['fund_name']
try:
# print('fund_code:', fund_code)
fund_em_info_df = ak.fund_em_open_fund_info(fund=fund_code)
except:
logger.warning(f'获取 {fund_code}-{fund_name} 失败')
continue
# print(fund_em_info_df.head())
fund_em_info_df.rename(columns={'净值日期': 'Date', '单位净值': 'price', '日增长率': 'ret'}, inplace=True)
fund_em_info_df.Date = pd.to_datetime(fund_em_info_df.Date)
# print(fund_em_info_df.head())
fund_em_info_df.price = fund_em_info_df.price.astype(float)
fund_em_info_df.set_index('Date', inplace=True)
fund_em_info_df.rename(columns={'price': fund_name}, inplace=True)
df_list.append(fund_em_info_df[[fund_name]])
total_df = pd.concat(df_list, axis=1)
total_df = total_df.loc[start:end, :]
total_df.fillna(method='ffill', inplace=True)
return total_df
def fetch_fund_em_open_fund_info(code='163412', path='./data/bs'):
file = '%s/%s.csv' % (path, code)
if os.path.exists(file):
logger.info('%s already exists' % file)
return
df = ak.fund_em_open_fund_info(code, '累计净值走势')
df.columns = ['date_time', 'accumulated_value']
df.to_csv(file, encoding='utf_8')
def filled_fund_data_df(fund_code):
'''
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3770 entries, 0 to 3769
Data columns (total 2 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 净值日期 3770 non-null object
1 累计净值 3770 non-null object
'''
src_df = ak.fund_em_open_fund_info(fund_code, indicator="累计净值走势")
df_date_index = src_df.set_index('净值日期')
date_rng = pd.date_range(df_date_index.index[0],
df_date_index.index[len(df_date_index) - 1],
freq="D")
return df_date_index.reindex(date_rng, method='ffill')
def max_drawdown(fCode, cPeriod=0, sDate='', eDate=''):
"""
计算历史数据区间某基金的最大回撤
:param fCode: 基金代码
:param cPeriod: 区间长度（天）
:param sDate: 开始日期（yyyy-mm-dd）
:param eDate: 结束日期（yyyy-mm-dd）
:return: 最大回撤（%）
"""
# 排除非法参数： 1) cPeriod与(sDate或eDate)同时出现 2) 有sDate无eDate 3) 有eDate无sDate
if (cPeriod and
(sDate or eDate)) or (sDate and not eDate) or (eDate and not sDate):
raise ValueError("Illegal Argument")
# 用akshare获取：开放式基金-历史数据
fInfo = ak.fund_em_open_fund_info(fund=fCode, indicator="单位净值走势")
# 用cPeriod参数：选出历史数据区间
if cPeriod:
fInfo = fInfo[-cPeriod:].reset_index(drop=True)
# 用sDate和eDate参数：选出历史数据区间
elif sDate and eDate:
sDate, eDate = datetime.strptime(sDate,
"%Y-%m-%d").date(), datetime.strptime(
eDate, "%Y-%m-%d").date()
fInfo = fInfo.loc[(fInfo["净值日期"] >= sDate)
& (fInfo["净值日期"] <= eDate)].reset_index(drop=True)
# 舍弃无用的数据包
fInfo = fInfo["单位净值"]
# 计算历史数据区间的最大回撤
maximumDrawdown = 0
for i in range(len(fInfo) - 1):
for j in range(i + 1, len(fInfo)):
drawdown = (fInfo["单位净值"][i] - fInfo["单位净值"][j]) / fInfo["单位净值"][i]
if drawdown > maximumDrawdown:
maximumDrawdown = drawdown
return "{:.2%}".format(maximumDrawdown)
def test_load_fund_data(self):
df = ak.fund_em_open_fund_info(fund="000217", indicator="单位净值走势")
print(df.to_string())
df_spx['SPX'] = calc(df_spx['SPX'])
df_spx.dropna(inplace=True)
# For looking up other funds
# df_funds = ak.fund_em_fund_name()
# df_funds.query('基金代码=="096001"')
funds_list = ['096001','161125','006075','050025','007722','007721']
# Putting the funds into one dataframe
df_list = []
for fd in funds_list:
df = ak.fund_em_open_fund_info(fund=fd, indicator="单位净值走势")
df = df.drop(['equityReturn','unitMoney'],axis=1)
df.rename(columns={'x':'date','y':fd},inplace=True)
df[fd] = df[fd].astype(float)
df.set_index('date',inplace=True)
df_list.append(df)
df_sum = pd.concat(df_list,axis=1,sort=True)
df_sum.dropna(inplace=True)
for fd in funds_list:
df_sum[fd] = calc(df_sum[fd])
df_sum.index = pd.to_datetime(df_sum.index)
df_spx.index = pd.to_datetime(df_spx.index)
df_rf = ak.rate_interbank(market="上海银行同业拆借市场",
symbol="Shibor人民币",
indicator="3月",
need_page="1")
print('无风险利率:' + str(df_rf['利率(%)'][0]) + '%\n')
rf = np.power(1 + df_rf['利率(%)'][0] / 100, 1 / 360) - 1 # 转为日利率
# 随机抽取code_num只基金获取收益率数据并回归
codes = list(codes)
random.shuffle(codes)
print('基金代码' + '\t' + 'Alpha的T值')
i = 0
while i < code_num and i < len(codes):
code = codes[i]
try:
df_i = ak.fund_em_open_fund_info(
fund=code, indicator="单位净值走势").rename(columns={'x': 'date'})
except:
time.sleep(10)
continue
df_i = df_i[['净值日期', '日增长率']]
if df_i['净值日期'].max().year - df_i['净值日期'].min().year < minYears:
codes.pop(i)
continue
df_i['净值日期'] = df_i['净值日期'].apply(lambda date: time.strftime('%Y-%m-%d'))
df_i['日增长率'] /= 100
df = pd.merge(df_M, df_i, how='inner', left_on='date', right_on='净值日期')
y = df['日增长率'].apply(float).values - rf
x = sm.add_constant(df['updown'].apply(float).values - rf)
df['Ri'] = y
df['Rm'] = df['updown'].apply(float).values - rf
df.to_excel('./原始数据/' + code + '.xlsx')
def beat_benchmark_ratio(fCode, cPeriod=0, sDate='', eDate=''):
# 排除非法参数： 1) cPeriod与(sDate或eDate)同时出现 2) 有sDate无eDate 3) 有eDate无sDate
if (cPeriod and
(sDate or eDate)) or (sDate and not eDate) or (eDate and not sDate):
raise ValueError("Illegal Argument")
# 用akshare获取：开放式基金-历史数据
fInfo = ak.fund_em_open_fund_info(fund=fCode, indicator="单位净值走势")
# 舍弃无用的数据包
fInfo = fInfo[["净值日期", "单位净值"]]
# 计算收益率为多少
rate = [(fInfo["单位净值"][i] - fInfo["单位净值"][i - 1]) / (fInfo["单位净值"][i - 1])
for i in range(1, len(fInfo))]
fInfo = fInfo[1:]
fInfo["rate"] = rate
# 用cPeriod参数：选出历史数据区间
if cPeriod:
fInfo = fInfo[-cPeriod:].reset_index(drop=True)
# 用sDate和eDate参数：选出历史数据区间
elif sDate and eDate:
f_sDate, f_eDate = datetime.strptime(
sDate, "%Y-%m-%d").date(), datetime.strptime(eDate,
"%Y-%m-%d").date()
fInfo = fInfo.loc[(fInfo["净值日期"] >= f_sDate)
& (fInfo["净值日期"] <= f_eDate)].reset_index(drop=True)
# 舍弃无用的数据包
fInfo = fInfo["rate"]
# 用akshare获取：沪深300-历史数据
benchmark = ak.stock_zh_index_daily(symbol="sh000300")
# 把date转化为column而不是作为index，方便后面的历史数据区间的选中
benchmark = benchmark.reset_index().rename({'index': 'date'},
axis='columns')
benchmark["date"] = benchmark["date"].apply(
lambda x: x.to_pydatetime().date())
# 舍弃无用的数据包
benchmark = benchmark[["date", "close"]]
# 计算收益率为多少
rate = [(benchmark["close"][i] - benchmark["close"][i - 1]) /
(benchmark["close"][i - 1]) for i in range(1, len(benchmark))]
benchmark = benchmark[1:]
benchmark["rate"] = rate
# 用cPeriod参数：选出历史数据区间
if cPeriod:
benchmark = benchmark[-cPeriod:].reset_index(drop=True)
# 用sDate和eDate参数：选出历史数据区间
elif sDate and eDate:
b_sDate, b_eDate = datetime.strptime(
sDate, "%Y-%m-%d").date(), datetime.strptime(eDate,
"%Y-%m-%d").date()
benchmark = benchmark.loc[(benchmark["date"] >= b_sDate) & (
benchmark["date"] <= b_eDate)].reset_index(drop=True)
# 舍弃无用的数据包
benchmark = benchmark["rate"]
# 计算区间内收益率击败沪深300的比例
benchmark = [
1 if fInfo[i] > benchmark[i] else 0 for i in range(len(fInfo))
]
benchmark_ratio = Counter(benchmark)[1] / len(benchmark)
return "{:.2%}".format(benchmark_ratio)
"HA-NSDK": "040046",
"JS-XXCZ": "260108",
"XQ-HR": "163406",
"YFD-ZXP": "110011",
"PH-HB": "000409",
"PH-FL": "003547",
"PH-FR": "000345",
"PH-GM": "160644",
"PH-KJ": "008811"
}
arr_fund_name = []
arr_fund_value = []
for fund_name, fund_code in fund_dict.items():
df_fund = ak.fund_em_open_fund_info(fund=fund_code, indicator="单位净值走势")
df_fund = df_fund.set_index("净值日期")
arr_fund_name.append(fund_name)
arr_fund_value.append(df_fund["单位净值"][-1])
print(fund_name)
time.sleep(1)
for stock_name, stock_code in stock_dict.items():
df_stock = ak.stock_zh_a_daily(symbol=stock_code,
start_date=e_date,
end_date=e_date,
adjust="qfq")
arr_fund_name.append(stock_name)
arr_fund_value.append(df_stock.loc[e_date, 'close'])
print(stock_name)
time.sleep(1)
#%%
import akshare as ak
#%%
import akshare as ak
fund_em_fund_name_df = ak.fund_em_fund_name()
print(fund_em_fund_name_df)
#%%
fund_etf_hist_sina_df = ak.fund_etf_hist_sina(
symbol="sz169103")
print(fund_etf_hist_sina_df)
#%%
fund_em_info_df = ak.fund_em_open_fund_info(
fund="163402", indicator="累计收益率走势")
print(fund_em_info_df)
#%%
import akshare as ak
js_news_df = ak.js_news(indicator='最新资讯')
# js_news_df['content']
for i in range(len(js_news_df)):
print(f"{js_news_df.iloc[i, 0]}")
print(f"{js_news_df.iloc[i, 1]}")
print()
#%%
import akshare as ak
import streamlit as st
import akshare as ak
import altair as alt
my_fund = [{'code': '0001630', 'name': '天弘中证计算机主题ETF连接C'}]
res = ak.fund_em_open_fund_info(fund="001630", indicator="单位净值走势")
st.dataframe(res.iloc())
import akshare as ak # 沪深300 6% # fund_codes = '110020' # 上证50 fund_codes = '110003' fund_history = ak.fund_em_open_fund_info(fund_codes) print(fund_history) # 1 * (1 + x) ** 10 = 1.7974 # import math # # print(math.pow(2.4335, 1 / 15) - 1) # print((1 + 0.10) ** 10)
#!/usr/bin/python
import akshare as ak
import pandas as pd
import matplotlib.pyplot as plt
fund_num = '110003'
#data="单位净值走势"
data = "累计收益率走势"
name_df = ak.fund_em_fund_name()
fund_name = fund_num + name_df.loc[name_df['基金代码'] == fund_num]['基金简称'].item()
df = ak.fund_em_open_fund_info(fund=fund_num, indicator=data)
#df['日增长率']=df['日增长率'].astype('float')
print(df)
df.plot(x='净值日期', figsize=(20, 10), grid=True,
title=fund_name).get_figure().savefig('fund.jpg')
#df.plot(x='净值日期',figsize=(20,10),grid=True,title=fund_name,secondary_y=['日增长率']).get_figure().savefig('fund.jpg')
