ADD: added optimizer workflow

optimizers.py

@@ -0,0 +1,69 @@
+import os
+from typing import Dict, Any
+import strategies
+from optimizer import run_optimized_backtest
+
+
+class Optimizer:
+    """Base optimizer wrapper tying a strategy to an optimization grid."""
+    name = 'BaseOptimizer'
+    strategy_cls = None
+    optimize_kwargs: Dict[str, Any] = None
+
+    @classmethod
+    def run(cls, symbol):
+        # pass optimizer name as report_tag so output files are unique per optimizer
+        return run_optimized_backtest(symbol, strategy_cls=cls.strategy_cls, optimize_kwargs=cls.optimize_kwargs, report_tag=cls.name)
+
+
+class RsiOptimizer(Optimizer):
+    name = 'RsiOptimizer'
+    strategy_cls = strategies.RsiStrategy
+    def constraint(p):
+        return p.rsi_upper > p.rsi_lower
+
+    optimize_kwargs = dict(
+        rsi_period=range(7, 30, 2),
+        rsi_lower=range(20, 40, 5),
+        rsi_upper=range(60, 80, 5),
+        maximize='Return [%]',
+        constraint=constraint,
+        atr_period=range(10, 20, 2),
+        stop_loss_atr_multiplier=[2.0, 2.5, 3.0, 3.5, 4.0],
+    )
+
+
+class CrossEmaOptimizer(Optimizer):
+    name = 'CrossEmaOptimizer'
+    strategy_cls = strategies.CrossEmaStrategy
+
+    def constraint(p):
+        # ensure long EMA period is greater than short EMA period
+        return p.long_ema > p.short_ema
+
+    optimize_kwargs = dict(
+        short_ema=range(5, 20, 3),
+        long_ema=range(20, 60, 5),
+        maximize='Return [%]',
+        constraint=constraint,
+        atr_period=range(10, 20, 2),
+        stop_loss_atr_multiplier=[2.0, 2.5, 3.0, 3.5],
+        finalize_trades=False,  # run final backtest with best params
+    )
+
+
+def get_optimizer_by_name(name: str):
+    mapping = {cls.name: cls for cls in (RsiOptimizer, CrossEmaOptimizer)}
+    # allow passing class name as well
+    if name in mapping:
+        return mapping[name]
+    # try attribute lookup in module
+    cls = getattr(__import__('optimizers'), name, None)
+    return cls
+
+
+def run_optimizer(symbol: str, optimizer_name: str):
+    cls = get_optimizer_by_name(optimizer_name)
+    if cls is None:
+        raise ValueError(f"Optimizer '{optimizer_name}' not found")
+    return cls.run(symbol)