From 73dd1ae3e4f3a09ba581d58aaa8d9e862f710029 Mon Sep 17 00:00:00 2001
From: Rich <24254625+richklee@users.noreply.github.com>
Date: Sun, 9 Jun 2024 17:54:09 +0100
Subject: [PATCH] refactor for correct type hints
---
alphavec/backtest.py | 59 ++++++++++++++++++++++--------------------
example.ipynb | 42 +++++++++++++++---------------
tests/test_backtest.py | 1 -
3 files changed, 52 insertions(+), 50 deletions(-)
diff --git a/alphavec/backtest.py b/alphavec/backtest.py
index fa69b60..4d1e075 100644
--- a/alphavec/backtest.py
+++ b/alphavec/backtest.py
@@ -7,7 +7,7 @@
DEFAULT_RISK_FREE_RATE = 0.02
-def zero_commission(weights: pd.DataFrame, prices: pd.DataFrame) -> float:
+def zero_commission(weights: pd.DataFrame, prices: pd.DataFrame) -> pd.DataFrame:
"""Zero trading commission.
Args:
@@ -17,10 +17,12 @@ def zero_commission(weights: pd.DataFrame, prices: pd.DataFrame) -> float:
Returns:
Always returns 0.
"""
- return 0
+ return pd.DataFrame(0, index=weights.index, columns=weights.columns)
-def flat_commission(weights: pd.DataFrame, prices: pd.DataFrame, fee: float) -> float:
+def flat_commission(
+ weights: pd.DataFrame, prices: pd.DataFrame, fee: float
+) -> pd.DataFrame:
"""Flat commission applies a fixed fee per trade.
Args:
@@ -37,7 +39,9 @@ def flat_commission(weights: pd.DataFrame, prices: pd.DataFrame, fee: float) ->
return commissions
-def pct_commission(weights: pd.DataFrame, prices: pd.DataFrame, fee: float) -> float:
+def pct_commission(
+ weights: pd.DataFrame, prices: pd.DataFrame, fee: float
+) -> pd.DataFrame:
"""Percentage commission applies a percentage fee per trade.
Args:
@@ -60,7 +64,9 @@ def backtest(
freq_day: int = 1,
trading_days_year: int = DEFAULT_TRADING_DAYS_YEAR,
shift_periods: int = 1,
- commission_func: Callable[[pd.DataFrame, pd.DataFrame], float] = zero_commission,
+ commission_func: Callable[
+ [pd.DataFrame, pd.DataFrame], pd.DataFrame
+ ] = zero_commission,
ann_borrow_rate: float = 0,
spread_pct: float = 0,
ann_risk_free_rate: float = DEFAULT_RISK_FREE_RATE,
@@ -69,7 +75,7 @@ def backtest(
]:
"""Backtest a trading strategy.
- Strategy is simulated using the given weights, returns, and cost parameters.
+ Strategy is simulated using the given weights, prices, and cost parameters.
Zero costs are calculated by default: no commission, no borrowing, no spread.
To prevent look-ahead bias the returns will be shifted 1 interval by default relative to the weights during backtest.
@@ -89,7 +95,7 @@ def backtest(
Index should be a DatetimeIndex.
Shape must match returns.
prices:
- Prices of the assets at each interval used to calculate returns ans costs.
+ Prices of the assets at each interval used to calculate returns and costs.
Each column should be the mark prices for a specific asset, with the column name being the asset name.
Column names should match weights.
Index should be a DatetimeIndex.
@@ -129,16 +135,16 @@ def backtest(
asset_cum = (1 + asset_rets).cumprod() - 1
asset_perf = pd.concat(
[
- asset_rets.apply(
- _ann_sharpe, periods=freq_year, risk_free_rate=ann_risk_free_rate
+ _ann_sharpe(
+ asset_rets, periods=freq_year, risk_free_rate=ann_risk_free_rate
),
- asset_rets.apply(_ann_vol, periods=freq_year),
- asset_rets.apply(_cagr, periods=freq_year),
- asset_rets.apply(_max_drawdown),
+ _ann_vol(asset_rets, periods=freq_year),
+ _cagr(asset_rets, periods=freq_year),
+ _max_drawdown(asset_rets),
],
keys=["annual_sharpe", "annual_volatility", "cagr", "max_drawdown"],
axis=1,
- )
+ ) # type: ignore
# Backtest a cost-aware strategy as defined by the given weights:
# 1. Calc costs
@@ -173,12 +179,12 @@ def backtest(
# Evaluate the strategy asset-wise performance
strat_perf = pd.concat(
[
- strat_rets.apply(
- _ann_sharpe, periods=freq_year, risk_free_rate=ann_risk_free_rate
+ _ann_sharpe(
+ strat_rets, periods=freq_year, risk_free_rate=ann_risk_free_rate
),
- strat_rets.apply(_ann_vol, periods=freq_year),
- strat_rets.apply(_cagr, periods=freq_year),
- strat_rets.apply(_max_drawdown),
+ _ann_vol(strat_rets, periods=freq_year),
+ _cagr(strat_rets, periods=freq_year),
+ _max_drawdown(strat_rets),
strat_ann_turnover,
_trade_count(weights) / strat_total_days,
],
@@ -191,7 +197,7 @@ def backtest(
"trades_per_day",
],
axis=1,
- )
+ ) # type: ignore
# Evaluate the strategy portfolio performance
port_rets = strat_rets.sum(axis=1)
@@ -213,8 +219,7 @@ def backtest(
index=["portfolio"],
)
- # Combine the asset and strategy performance metrics
- # into a single dataframe for comparison
+ # Combine the asset and strategy performance metrics into a single dataframe for comparison
perf = pd.concat(
[asset_perf, strat_perf],
keys=["asset", "strategy"],
@@ -264,7 +269,7 @@ def _ann_sharpe(
rets: pd.DataFrame | pd.Series,
risk_free_rate: float = DEFAULT_RISK_FREE_RATE,
periods: int = DEFAULT_TRADING_DAYS_YEAR,
-) -> float:
+) -> pd.DataFrame | pd.Series:
ann_rfr = (1 + risk_free_rate) ** (1 / periods) - 1
mu = rets.mean()
sigma = rets.std()
@@ -290,11 +295,9 @@ def _cagr(
) -> pd.DataFrame | pd.Series:
cumprod = (1 + rets).cumprod().dropna()
if len(cumprod) == 0:
- return 0
+ return rets * 0
final = cumprod.iloc[-1]
- if final <= 0:
- return 0
n = len(cumprod) / periods
cagr = final ** (1 / n) - 1
@@ -327,8 +330,8 @@ def _turnover(
diff = weights.fillna(0).diff()
# Capital is fixed (uncompounded) for each interval so we can calculate the trade volume
# Sum the volume of the buy and sell trades
- buy_volume = (diff.where(diff > 0, 0).abs() * capital).sum()
- sell_volume = (diff.where(diff < 0, 0).abs() * capital).sum()
+ buy_volume = (diff.where(lambda x: x.gt(0), 0).abs() * capital).sum()
+ sell_volume = (diff.where(lambda x: x.lt(0), 0).abs() * capital).sum()
# Trade volume is the minimum of the buy and sell volumes
# Wrap in Series in case of scalar volume sum (when weights is a Series)
trade_volume = pd.concat(
@@ -341,7 +344,7 @@ def _turnover(
return turnover
-def _trade_count(weights: pd.DataFrame | pd.Series) -> pd.DataFrame | pd.Series:
+def _trade_count(weights: pd.DataFrame | pd.Series) -> int | pd.Series:
diff = weights.fillna(0).diff().abs() != 0
tx = diff.astype(int)
return tx.sum()
diff --git a/example.ipynb b/example.ipynb
index d20ee79..32748b6 100644
--- a/example.ipynb
+++ b/example.ipynb
@@ -703,11 +703,11 @@
" | BTCUSDT | \n",
" 0.784522 | \n",
" 0.676503 | \n",
- " 0.371469 | \n",
+ " 0.697523 | \n",
" 0.766293 | \n",
" 2.082824 | \n",
" 0.512587 | \n",
- " 1.616165 | \n",
+ " 4.004693 | \n",
" 0.632456 | \n",
" 2.944701 | \n",
" 1.000637 | \n",
@@ -716,11 +716,11 @@
" DOGEUSDT | \n",
" 1.021940 | \n",
" 2.489353 | \n",
- " 1.826904 | \n",
+ " 3.422070 | \n",
" 0.923328 | \n",
" 0.877637 | \n",
" 1.058797 | \n",
- " 0.699507 | \n",
+ " 1.134931 | \n",
" 0.659074 | \n",
" 2.331747 | \n",
" 1.000745 | \n",
@@ -729,11 +729,11 @@
" ETHUSDT | \n",
" 1.029580 | \n",
" 0.877444 | \n",
- " 0.695604 | \n",
+ " 1.421963 | \n",
" 0.793027 | \n",
" 0.688909 | \n",
" 0.385551 | \n",
- " 0.233350 | \n",
+ " 0.420768 | \n",
" 0.497637 | \n",
" 4.970033 | \n",
" 1.000637 | \n",
@@ -742,11 +742,11 @@
" MATICUSDT | \n",
" 1.409684 | \n",
" 1.405686 | \n",
- " 1.879349 | \n",
+ " 3.913546 | \n",
" 0.879694 | \n",
" 0.398154 | \n",
" 0.615083 | \n",
- " 0.081175 | \n",
+ " 0.124623 | \n",
" 0.585446 | \n",
" 5.556876 | \n",
" 1.000708 | \n",
@@ -768,11 +768,11 @@
" XRPUSDT | \n",
" 0.673812 | \n",
" 1.135899 | \n",
- " 0.190893 | \n",
+ " 0.339995 | \n",
" 0.832385 | \n",
" -0.063805 | \n",
" 0.408337 | \n",
- " -0.086424 | \n",
+ " -0.140457 | \n",
" 0.659143 | \n",
" 8.762681 | \n",
" 1.000637 | \n",
@@ -785,22 +785,22 @@
" asset \\\n",
" annual_sharpe annual_volatility cagr max_drawdown \n",
"symbol \n",
- "BTCUSDT 0.784522 0.676503 0.371469 0.766293 \n",
- "DOGEUSDT 1.021940 2.489353 1.826904 0.923328 \n",
- "ETHUSDT 1.029580 0.877444 0.695604 0.793027 \n",
- "MATICUSDT 1.409684 1.405686 1.879349 0.879694 \n",
+ "BTCUSDT 0.784522 0.676503 0.697523 0.766293 \n",
+ "DOGEUSDT 1.021940 2.489353 3.422070 0.923328 \n",
+ "ETHUSDT 1.029580 0.877444 1.421963 0.793027 \n",
+ "MATICUSDT 1.409684 1.405686 3.913546 0.879694 \n",
"SOLUSDT 0.784384 1.216761 0.250883 0.962695 \n",
- "XRPUSDT 0.673812 1.135899 0.190893 0.832385 \n",
+ "XRPUSDT 0.673812 1.135899 0.339995 0.832385 \n",
"\n",
" strategy \\\n",
" annual_sharpe annual_volatility cagr max_drawdown, \n",
"symbol \n",
- "BTCUSDT 2.082824 0.512587 1.616165 0.632456 \n",
- "DOGEUSDT 0.877637 1.058797 0.699507 0.659074 \n",
- "ETHUSDT 0.688909 0.385551 0.233350 0.497637 \n",
- "MATICUSDT 0.398154 0.615083 0.081175 0.585446 \n",
+ "BTCUSDT 2.082824 0.512587 4.004693 0.632456 \n",
+ "DOGEUSDT 0.877637 1.058797 1.134931 0.659074 \n",
+ "ETHUSDT 0.688909 0.385551 0.420768 0.497637 \n",
+ "MATICUSDT 0.398154 0.615083 0.124623 0.585446 \n",
"SOLUSDT 0.533112 0.477024 0.174253 0.459495 \n",
- "XRPUSDT -0.063805 0.408337 -0.086424 0.659143 \n",
+ "XRPUSDT -0.063805 0.408337 -0.140457 0.659143 \n",
"\n",
" \n",
" annual_turnover trades_per_day \n",
@@ -953,7 +953,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
- "version": "3.11.7"
+ "version": "3.11.9"
}
},
"nbformat": 4,
diff --git a/tests/test_backtest.py b/tests/test_backtest.py
index 872c09f..6ea966f 100644
--- a/tests/test_backtest.py
+++ b/tests/test_backtest.py
@@ -1,7 +1,6 @@
import sys
import os
from pathlib import PurePath
-from functools import partial
import logging
import numpy as np