def bcolz_table_path(table_name, bundle='cninfo'):
"""bcolz文件路径"""
root_dir = zipline_path(['bcolz', bundle])
if not os.path.exists(root_dir):
os.makedirs(root_dir)
path_ = os.path.join(root_dir, '{}.bcolz'.format(table_name))
return path_
def get_ff_factors(n):
"""读取3因子或5因子数据"""
assert n in (3, 5), "仅支持3因子或5因子"
file_name = f"ff{n}"
root_dir = zipline_path(['factors'])
result_path_ = os.path.join(root_dir, f'{file_name}.pkl')
return pd.read_pickle(result_path_)
def get_backtest(dir_name=zipline_path(['backtest']), file_name=None):
"""获取最近的回测结果(数据框)"""
if file_name is None:
pref_file, _ = get_latest_backtest_info(dir_name)
else:
assert isinstance(file_name, str)
assert file_name.endswith('.pkl'), '文件名必须带".pkl"扩展'
pref_file = os.path.join(dir_name, file_name)
return pd.read_pickle(pref_file)
def get_latest_backtest_info(dir_name=zipline_path(['backtest'])):
"""最新回测结果文件路径及更新时间"""
assert os.path.isdir(dir_name)
try:
candidates = [os.path.join(dir_name, x) for x in os.listdir(dir_name)]
most_recent = max(candidates, key=os.path.getmtime)
return most_recent, pd.Timestamp(int(os.path.getmtime(most_recent)),
unit='s',
tz='Asia/Shanghai')
except (ValueError, OSError) as e:
if getattr(e, 'errno', errno.ENOENT) != errno.ENOENT:
raise
raise ValueError('在目录{}下，没有发现回测结果'.format(dir_name))
def backtest_result_path():
from zipline.utils import paths as pth
ret = pth.zipline_path(['backtest'])
pth.ensure_directory(ret)
return ret
def show_draw_result(title, results_df, bundle):
import matplotlib
from matplotlib import gridspec
import matplotlib.image as mpimg
import matplotlib.pyplot as plt
from zipline.utils import paths
from datetime import datetime
plt.style.use('ggplot')
red = "#aa4643"
blue = "#4572a7"
black = "#000000"
figsize = (18, 6)
f = plt.figure(title, figsize=figsize)
gs = gridspec.GridSpec(10, 8)
# TODO draw logo
# ax = plt.subplot(gs[:3, -1:])
# ax.axis("off")
# filename = os.path.join(paths.zipline_root(), 'zipline.png')
# img = mpimg.imread(filename)
# imgplot = ax.imshow(img, interpolation="nearest")
# ax.autoscale_view()
# draw risk and portfolio
series = results_df.iloc[-1]
font_size = 12
value_font_size = 11
label_height, value_height = 0.8, 0.6
label_height2, value_height2 = 0.35, 0.15
fig_data = [
(0.00, label_height, value_height, "Total Returns",
"{0:.3%}".format(series.algorithm_period_return), red, black),
(0.15, label_height, value_height, "Annual Returns",
"{0:.3%}".format(series.annualized_algorithm_return), red, black),
(0.00, label_height2, value_height2, "Benchmark Total",
"{0:.3%}".format(series.benchmark_period_return), blue, black),
(0.15, label_height2, value_height2, "Benchmark Annual",
"{0:.3%}".format(series.annualized_benchmark_return), blue, black),
(0.30, label_height, value_height, "Alpha",
"{0:.4}".format(series.alpha), black, black),
(0.40, label_height, value_height, "Beta",
"{0:.4}".format(series.beta), black, black),
(0.55, label_height, value_height, "Sharpe",
"{0:.4}".format(series.sharpe), black, black),
(0.70, label_height, value_height, "Sortino",
"{0:.4}".format(series.sortino), black, black),
(0.85, label_height, value_height, "Information Ratio",
"{0:.4}".format(series.information), black, black),
(0.30, label_height2, value_height2, "Volatility",
"{0:.4}".format(series.algo_volatility), black, black),
(0.40, label_height2, value_height2, "MaxDrawdown",
"{0:.3%}".format(series.max_drawdown), black, black),
# (0.55, label_height2, value_height2, "Tracking Error", "{0:.4}".format(series.tracking_error), black, black),
# (0.70, label_height2, value_height2, "Downside Risk", "{0:.4}".format(series.downside_risk), black, black),
]
ax = plt.subplot(gs[:3, :-1])
ax.axis("off")
for x, y1, y2, label, value, label_color, value_color in fig_data:
ax.text(x, y1, label, color=label_color, fontsize=font_size)
ax.text(x, y2, value, color=value_color, fontsize=value_font_size)
# strategy vs benchmark
ax = plt.subplot(gs[4:, :])
ax.get_xaxis().set_minor_locator(matplotlib.ticker.AutoMinorLocator())
ax.get_yaxis().set_minor_locator(matplotlib.ticker.AutoMinorLocator())
ax.grid(b=True, which='minor', linewidth=.2)
ax.grid(b=True, which='major', linewidth=1)
ax.plot(results_df["benchmark_period_return"],
label="benchmark",
alpha=1,
linewidth=2,
color=blue)
ax.plot(results_df["algorithm_period_return"],
label="algorithm",
alpha=1,
linewidth=2,
color=red)
# manipulate
vals = ax.get_yticks()
ax.set_yticklabels(['{:3.2f}%'.format(x * 100) for x in vals])
leg = plt.legend(loc="upper left")
leg.get_frame().set_alpha(0.5)
plt.show()
now = datetime.now()
paths.ensure_directory(paths.zipline_path(['perf']))
plt.savefig(filename=os.path.join(
paths.zipline_path(['perf']),
os.path.basename(title).split('.')[0] + '_' + bundle + '_' +
now.strftime('%Y%m%dT%H%M%s') + '.png'))
def backtest_result_path():
from zipline.utils import paths as pth
ret = pth.zipline_path(['backtest'])
pth.ensure_directory(ret)
return ret
df_rf = get_treasury_data(None, None)['3month']
df_rm = get_cn_benchmark_returns('000300')
# 目标输出
rm_rf = df_rm - df_rf
result = pd.DataFrame({
'Mkt-RF': rm_rf.values,
'SMB': 0.0,
'HML': 0.0
},
index=rm_rf.index)
result.loc[df_rf.index, 'RF'] = df_rf
# 结果存放路径
file_name = "ff3"
root_dir = zipline_path(['factors'])
if not os.path.exists(root_dir):
os.makedirs(root_dir)
result_path_ = os.path.join(root_dir, f'{file_name}.pkl')
class MarketEquity(CustomFactor):
"""
this factor outputs the market cap of every security on the day.
"""
window_length = 1
inputs = [USEquityPricing.market_cap]
def compute(self, today, assets, out, mcap):
out[:] = mcap[-1]