Initial stats setup

Author: kylewidmann

.coveragerc

@@ -0,0 +1,3 @@
+[run]
+omit =
+    pytradebacktest/stats.py

pytradebacktest/stats.py

@@ -0,0 +1,213 @@
+from typing import Any
+
+import numpy as np
+import pandas as pd
+from pytrade.models.trade import Trade
+from pytrade.strategy import FxStrategy
+
+from pytradebacktest.data import MarketData
+
+
+class Stats:
+
+    def __init__(
+        self,
+        trades: list[Trade],
+        equity: np.ndarray,
+        market_data: MarketData,
+        strategy: FxStrategy,
+    ):
+        self._trades = trades
+        self._equity = equity
+        self._market_data = market_data
+        self._strategy = strategy
+        self._trades_df = pd.DataFrame(
+            {
+                "Size": [t.size for t in trades],
+                "EntryBar": [t.entry_bar for t in trades],
+                "ExitBar": [t.exit_bar for t in trades],
+                "EntryPrice": [t.entry_price for t in trades],
+                "ExitPrice": [t.exit_price for t in trades],
+                "PnL": [t.pl for t in trades],
+                "ReturnPct": [t.pl_pct for t in trades],
+                "EntryTime": [t.entry_time for t in trades],
+                "ExitTime": [t.exit_time for t in trades],
+                "Tag": [t.tag for t in trades],
+                "TakeProfit": [t.tp for t in trades],
+                "StopLoss": [t.sl for t in trades],
+            }
+        )
+        self._trades_df["Duration"] = (
+            self._trades_df["ExitTime"] - self._trades_df["EntryTime"]
+        )
+        self._drawdown: np.ndarray[np.floating[Any], Any]
+
+    @property
+    def drawdown(self) -> np.ndarray[np.floating[Any], Any]:
+        if not self._drawdown:
+            self._drawdown = 1 - self._equity / np.maximum.accumulate(self._equity)
+        return self._drawdown
+
+    @property
+    def drawdown_duration(self):
+        iloc = np.unique(
+            np.r_[(self.drawdown == 0).values.nonzero()[0], len(self.drawdown) - 1]
+        )
+        iloc = pd.Series(iloc, index=self.drawdown.index[iloc])
+        df = iloc.to_frame("iloc").assign(prev=iloc.shift())
+        df = df[df["iloc"] > df["prev"] + 1].astype(int)
+
+        # If no drawdown since no trade, avoid below for pandas sake and return nan series
+        if not len(df):
+            return (self.drawdown.replace(0, np.nan),) * 2
+
+        # df = df.reindex(self.drawdown.index)
+        return df["iloc"].map(self.drawdown.index.__getitem__) - df["prev"].map(
+            self.drawdown.index.__getitem__
+        )
+
+    @property
+    def drawdown_peaks(self):
+        iloc = np.unique(
+            np.r_[(self.drawdown == 0).values.nonzero()[0], len(self.drawdown) - 1]
+        )
+        iloc = pd.Series(iloc, index=self.drawdown.index[iloc])
+        df = iloc.to_frame("iloc").assign(prev=iloc.shift())
+        df = df[df["iloc"] > df["prev"] + 1].astype(int)
+
+        # If no drawdown since no trade, avoid below for pandas sake and return nan series
+        if not len(df):
+            return (self.drawdown.replace(0, np.nan),) * 2
+
+        # df = df.reindex(self.drawdown.index)
+        return df.apply(
+            lambda row: self.drawdown.iloc[row["prev"] : row["iloc"] + 1].max(), axis=1
+        )
+
+    @property
+    def profit_and_loss(self):
+        return self._trades_df["PnL"]
+
+    @property
+    def returns(self):
+        return self._trades_df["ReturnPct"]
+
+    @property
+    def durations(self):
+        return self._trades_df["Duration"]
+
+    @property
+    def start(self):
+        return self._market_data._start_index
+    
+    @property
+    def end(self):
+        return self._market_data._end_index
+    
+    @property
+    def duration(self):
+        return self.start - self.end
+
+    @property
+    def positions(self):
+        pass
+
+    @property
+    def exposure_time(self):
+        pass
+
+    @property
+    def equity_final(self):
+        pass
+
+    @property
+    def equity_peak(self):
+        pass
+
+    @property
+    def return_pct(self):
+        pass
+
+    @property
+    def buy_and_hold_return(self):
+        pass
+
+    @property
+    def annualized_return(self):
+        pass
+
+    @property
+    def annualized_volatility(self):
+        pass
+
+    @property
+    def sharpe_ratio(self):
+        pass
+
+    @property
+    def sortino_ratio(self):
+        pass
+
+    @property
+    def calmar_ratio(self):
+        pass
+
+    @property
+    def max_drawndown(self):
+        pass
+
+    @property
+    def avg_drawdown(self):
+        pass
+
+    @property
+    def max_drawdown_duration(self):
+        pass
+
+    @property
+    def avg_drawdown_duration(self):
+        pass
+
+    @property
+    def number_of_trades(self):
+        pass
+
+    @property
+    def win_rate(self):
+        pass
+
+    @property
+    def best_trade_return(self):
+        pass
+
+    @property
+    def worst_trade_return(self):
+        pass
+
+    @property
+    def avg_trade_return(self):
+        pass
+
+    @property
+    def max_trade_duration(self):
+        pass
+
+    @property
+    def avg_trade_duration(self):
+        pass
+
+    @property
+    def profit_factor(self):
+        pass
+
+    @property
+    def expectancy(self):
+        pass
+
+    @property
+    def SQN(self):
+        pass
+
+    @property
+    def kelly_criterion(self):
+        pass

tests/unit/test_stats.py

@@ -0,0 +1,2 @@
+def test_stats():
+    pass