import numpy as np
import pandas as pd
from intelligence.recommendations import recommend_settings
from intelligence.regime import RegimeDetector
def test_recommend_settings():
# 1. High Drawdown -> Suggest max_alloc_pct reduction
summary = {
"metrics": {"max_drawdown_max": 0.40, "max_drawdown_p95": 0.30},
"strategy_params_detected": {"max_alloc_pct": 0.10}
}
res = recommend_settings(summary)
updates = res["recommended_params"]
assert "max_alloc_pct" in updates
assert updates["max_alloc_pct"] == 0.02
# 2. Negative returns -> Suggest min_signal_strength
summary = {
"metrics": {"return_p05": -0.15},
"strategy_params_detected": {}
}
res = recommend_settings(summary)
# Checks guidance even if not detected
recs = res["recommendations"]
assert any(r["param"] == "min_signal_strength" for r in recs)
def test_regime_detector():
det = RegimeDetector()
# Not enough data
df_small = pd.DataFrame({"close": range(10)})
assert "error" in det.detect(df_small)
# Trending Up
# 100 periods
close = np.linspace(100, 200, 100) # Strong uptrend
df = pd.DataFrame({
"high": close + 5,
"low": close - 5,
"close": close,
})
res = det.detect(df)
# ADX should be high for straight line? Actually straight line has 0 volatility but direction is up.
# ADX calculation is complex.
# Let's ensure it runs and returns keys.
assert "regime" in res
assert "adx" in res
assert "atr_pct" in res
