Alice — an open-source AI trading bot covering equities, crypto, commodities, forex, and macro

master
py 107 lines 3.33 KB
Raw
1 """Pure helpers that turn strategy output into dashboard JSON payloads."""
2
3 from __future__ import annotations
4
5 from typing import Any
6
7 import pandas as pd
8
9 from open_alice.engine import run_backtest
10 from open_alice.strategies import (
11 BollingerStrategy,
12 MacdStrategy,
13 RsiStrategy,
14 SmaCrossover,
15 Strategy,
16 )
17
18 _OVERLAY_COLUMNS: dict[str, list[str]] = {
19 "sma_crossover": ["SMA_fast", "SMA_slow"],
20 "rsi": ["RSI"],
21 "macd": ["MACD", "MACD_Signal"],
22 "bollinger": ["BB_Upper", "BB_Mid", "BB_Lower"],
23 }
24
25
26 def build_strategy(name: str, fast: int = 20, slow: int = 50) -> Strategy:
27 """Build a strategy instance by name."""
28 if name == "sma_crossover":
29 return SmaCrossover(fast=fast, slow=slow)
30 factories: dict[str, type[Strategy]] = {
31 "rsi": RsiStrategy,
32 "macd": MacdStrategy,
33 "bollinger": BollingerStrategy,
34 }
35 return factories[name]()
36
37
38 def chart_payload(
39 df: pd.DataFrame,
40 strategy_name: str,
41 fast: int = 20,
42 slow: int = 50,
43 max_points: int = 300,
44 ) -> dict[str, Any]:
45 """Run a strategy + backtest and return a JSON-serializable chart payload.
46
47 Args:
48 df: OHLCV DataFrame with a Returns column (see data.add_returns).
49 strategy_name: One of sma_crossover, rsi, macd, bollinger.
50 fast: Fast SMA period (sma_crossover only).
51 slow: Slow SMA period (sma_crossover only).
52 max_points: Downsample to at most this many chart points.
53
54 Returns:
55 Dict with labels, close, overlays, markers, equity and metrics.
56 """
57 strategy = build_strategy(strategy_name, fast=fast, slow=slow)
58 framed = strategy.generate_signals(df)
59 if len(framed) > max_points:
60 step = len(framed) // max_points + 1
61 framed = framed.iloc[::step]
62
63 result = run_backtest(framed)
64
65 labels = [_label(ts) for ts in framed.index]
66 payload: dict[str, Any] = {
67 "labels": labels,
68 "close": [float(v) for v in framed["Close"].tolist()],
69 "overlays": {
70 col: [None if pd.isna(v) else float(v) for v in framed[col].tolist()]
71 for col in _OVERLAY_COLUMNS[strategy_name]
72 if col in framed.columns
73 },
74 "markers": _signal_markers(framed),
75 "equity": {
76 "labels": labels,
77 "values": [float(v) for v in result.equity_curve.tolist()],
78 },
79 "metrics": {
80 "total_return": result.total_return,
81 "sharpe_ratio": result.sharpe_ratio,
82 "max_drawdown": result.max_drawdown,
83 "num_trades": result.num_trades,
84 "win_rate": result.win_rate,
85 },
86 }
87 return payload
88
89
90 def _label(ts: Any) -> str:
91 if hasattr(ts, "isoformat"):
92 return str(ts.isoformat())
93 return str(ts)
94
95
96 def _signal_markers(framed: pd.DataFrame) -> list[dict[str, Any]]:
97 """Extract fresh buy/sell markers from Signal/Position columns."""
98 if "Signal" not in framed.columns:
99 return []
100 sig = framed["Signal"]
101 prev = sig.shift(1).fillna(0)
102 markers: list[dict[str, Any]] = []
103 for ts, row in framed[(sig == 1) & (prev != 1)].iterrows():
104 markers.append({"t": _label(ts), "price": float(row["Close"]), "side": "buy"})
105 for ts, row in framed[(sig == -1) & (prev != -1)].iterrows():
106 markers.append({"t": _label(ts), "price": float(row["Close"]), "side": "sell"})
107 return markers