matplot 绘图 最大回撤
修改了一位大佬的blog，找不到出处了，最终效果：
import datetime
import matplotlib.pyplot as plt
import matplotlib.dates as mdate
import numpy as np
import random
import pandas as pd
DAYS = 365
random.seed(2022)
def Init():
""" 初始化设置 """
startdate = datetime.date.today()
xdate = pd.date_range(startdate, periods=DAYS, freq='B')
ycapital = [3000]
for _ in range(DAYS - 1):
ycapital.append(ycapital[-1] + random.uniform(-1, 1.1))
return xdate, ycapital
def cal_maxdrawdown(df_stock, ori_column='price', dd_column='returns'):
""" Calculate the MaxDrawdown """
df_stock = cal_drawdown(df_stock, column=ori_column)
MaxDrawdown = np.max(df_stock.loc[:, 'drawdown'])
end_idx = np.argmax(df_stock.loc[:, 'drawdown'])
start_idx = np.argmax(df_stock[ori_column].values[:end_idx])
return MaxDrawdown, end_idx, start_idx
def max_drawdown(ycapital):
""" Calculate the MaxDrawdown """
# 计算每日的回撤
drawdown = []
tmp_max_capital = ycapital[0]
for c in ycapital:
tmp_max_capital = max(c, tmp_max_capital)
drawdown.append(1 - c / tmp_max_capital)
MaxDrawdown = max(drawdown) # 最大回撤
endidx = np.argmax(drawdown) # 计算最大回撤日期范围
startidx = np.argmax(ycapital[:endidx]) # enddate = xdate[endidx]
# startdate = xdate[startidx]
# 仅仅画图的话，我们只要索引值更加方便
return MaxDrawdown, startidx, endidx
def max_drawdown_duration(ycapital):
duration = []
tmp_max_capital = ycapital[0]
for c in ycapital:
if c >= tmp_max_capital:
duration.append(0)
else:
duration.append(duration[-1] + 1)
# tmp_max_capital = max(c, tmp_max_capital)
MaxDDD = max(duration)
endidx = np.argmax(duration)
startidx = endidx - MaxDDD
return MaxDDD, startidx, endidx
def max_drawdown_restore_time(startidx, endidx, xdate, ycapital):
"""
startidx:表示最大回撤的开始时间在 xdate 中的索引，由 max_drawdown 方法返回
endidx:表示最大回撤的结束时间在 xdate 中的索引，由 max_drawdown 方法返回
"""
maxdd_resore_time = 0
restore_endidx = np.inf
for t in range(endidx, len(xdate)):
if ycapital[t] >= ycapital[startidx]:
restore_endidx = t
break
else:
maxdd_resore_time += 1
restore_endidx = min(restore_endidx, len(xdate) - 1)
return maxdd_resore_time, restore_endidx
def set_spline(ax, label: str = 'lightgray'):
""" 设置边框 """
ax.spines['top'].set_visible(False) # 去掉边框 top
ax.spines['left'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.spines['bottom'].set_color(label) # 设置 x 轴颜色
def set_sizes(font_size: int = 10):
"""
set fonts for plt
https://matplotlib.org/2.0.2/api/pyplot_api.html?highlight=rc#matplotlib.pyplot.rc
:param fig_size: figure size
:param font_tsize: figure title size
:param font_size: font size
:return:
"""
# Set the default text font size, weight='bold'
plt.rc('font', size=font_size)
plt.rc('xtick', labelsize=font_size)
plt.rc('ytick', labelsize=font_size)
plt.rc('axes', labelsize=font_size)
plt.rc('axes', titlesize=font_size)
plt.rc('legend', fontsize=font_size)
def set_lable(label_dict, fontdict, bold=True):
""" 设置 label 标签"""
if bold:
plt.title(label_dict["title"], fontdict=fontdict, weight='bold')
plt.xlabel(label_dict["xlabel"], fontdict=fontdict, weight='bold')
plt.ylabel(label_dict["ylabel"], fontdict=fontdict, weight='bold')
else:
plt.title("random account value", fontdict=fontdict) # or plt.suptitle
plt.xlabel("date(day)", fontdict=fontdict)
plt.ylabel("account value", fontdict=fontdict)
def set_xtick_cut(ax, format: str = '%Y-%m-%d', cut: int = 9):
""" 将xtick切分成多等分 """
ax.xaxis.set_major_formatter(mdate.DateFormatter(format)) # 设置时间标签显示格式
delta = round(len(xdate) / cut) # 分成 (cut+1) 份
plt.xticks([xdate[i * delta] for i in range(cut)] + [xdate[-1]])
def plot(xdate, ycapital, df_stock):
set_sizes(font_size=11)
# plt.style.use('seaborn-whitegrid')
fig, ax = plt.subplots(figsize=(12, 8))
plt.plot(xdate, ycapital, 'red', label='My Strategy', linewidth=2)
MaxDrawdown, startidx, endidx = max_drawdown(ycapital)
# MaxDrawdown, startidx, endidx = cal_maxdrawdown(df_stock, ori_column='price', dd_column='returns')
print("最大回撤为：", MaxDrawdown)
plt.scatter([xdate[startidx], xdate[endidx]], [ycapital[startidx], ycapital[endidx]],
s=100, c='b', marker='s', label='MaxDrawdown')
maxdd_resore_time, restore_endidx = max_drawdown_restore_time(startidx, endidx, xdate, ycapital)
print("最大回撤恢复时间为(天)：", maxdd_resore_time)
plt.scatter([xdate[endidx], xdate[restore_endidx]], [ycapital[endidx], ycapital[restore_endidx]],
s=100, c='cyan', marker='D', label='MaxDrawdown Restore Time')
# 绘制最大回撤持续期标识 marker = 'D'
MaxDDD, startidx, endidx = max_drawdown_duration(ycapital)
print("最大回撤持续期为(天)：", MaxDDD)
plt.scatter([xdate[startidx], xdate[endidx]], [ycapital[startidx], ycapital[endidx]],
s=80, c='g', marker='v', label='MaxDrawdown Duration')
plt.xticks(rotation=15)
# plt.yticks(color='gray') # 设置刻度值颜色
fontdict = {"family": "serif", 'size': 13} # Times New Roman, Arial; 'color': 'gray'
label_dict = {
"title": "Random account value",
"xlabel": "Date(daily)",
"ylabel": "Account value",
}
set_lable(label_dict, fontdict, bold=True)
set_spline(ax) # 去掉边框
set_xtick_cut(ax) # 分成 10 份
plt.tick_params(left='off')
# 设置刻度的朝向，宽，长度
plt.tick_params(which='major', direction='out', width=0.3, length=3) # in, out or inout
plt.grid(axis='y', color='lightgray', linestyle='-', linewidth=0.5)
plt.legend(loc='best', frameon=False, ncol=1)
plt.show()
def cal_drawdown(data, column='price'):
""" Calculate drawdown"""
data['returns'] = np.log(data[column] / data[column].shift(1))
data['cumret'] = data['returns'].cumsum().apply(np.exp)
data['cummax'] = data['cumret'].cummax()
data['drawdown'] = data['cummax'] - data['cumret']
return data
if __name__ == '__main__':
xdate, ycapital = Init()
df_stock = pd.DataFrame({'date': xdate, 'price': ycapital})
df_stock = df_stock.set_index('date', drop=True)
MaxDrawdown, end_idx, start_idx = cal_maxdrawdown(df_stock, ori_column='price', dd_column='returns')
plot(xdate, ycapital, df_stock)
