initialize test for premium calculations

climada_petals/engine/cat_bonds/test/test_premium_class.py

@@ -0,0 +1,87 @@
+import numpy as np
+import pandas as pd
+import logging
+from climada_petals.engine.cat_bonds.premium_class import PremiumCalculations  
+
+logging.basicConfig(
+     format="{asctime} - {levelname} - {message}",
+     style="{",
+     datefmt="%Y-%m-%d %H:%M",
+     level=logging.INFO,
+ )
+LOGGER = logging.getLogger(__name__)
+
+class DummyBondSim:
+    """Minimum mock needed for PremiumCalculations."""
+    def __init__(self, EL_ann, term, df_loss_month=None):
+        self.loss_metrics = {"EL_ann": EL_ann}
+        self.term = term
+        self.df_loss_month = df_loss_month
+
+
+# ---------------------------------------------------------
+# TEST SHARPE OPTIMIZATION
+# ---------------------------------------------------------
+
+def test_find_sharpe_single_loss_event():
+    """
+    Validate net cash flow logic. 
+    1 year, 1 loss in June, simple numbers.
+    """
+    df = pd.DataFrame({
+        "losses": [[0.2]],
+        "months": [[6]]
+    })
+
+    bond = DummyBondSim(EL_ann=0.02, term=1, df_loss_month=df)
+    pc = PremiumCalculations(bond)
+
+    # manually compute NCF:
+    # Pre-event premium: (1 * p)/12 * 6
+    # Post-event premium: (0.8 * p)/12 * (12 - 6) - 0.2 loss
+    # NCF = p/2 + (0.8p/2 - 0.2) = 0.9p - 0.2
+    p = 0.05
+    expected_ncf = 0.9 * p - 0.2
+
+    sharpe = pc.find_sharpe(premium=p, monthly_losses=df, target_sharpe=0)
+
+    # extracted NCF from method
+    ncf = []
+    cur_nominal = 1
+    losses = [0.2]
+    months = [6]
+    ncf_pre = (cur_nominal * p) / 12 * 6
+    cur_nominal -= 0.2
+    ncf_post = (cur_nominal * p) / 12 * 6 - 0.2
+    manual_list = [ncf_pre + ncf_post]
+
+    manual_sharpe = (np.mean(manual_list) / np.std(manual_list + [manual_list[0] * 1.0001]))**2  # std>0
+
+    assert np.isclose(sharpe, manual_sharpe, rtol=1e-6)  # just ensure valid number
+    assert np.isclose(expected_ncf, manual_list[0], rtol=1e-6)
+
+
+def test_calc_benchmark_premium_converges():
+    """
+    Test that optimization finds higher premium when losses > 0.
+    """
+
+    df = pd.DataFrame({
+        "losses": [[0.1], [0.2], [0.0]],
+        "months": [[3], [6], [0]]
+    })
+
+    bond = DummyBondSim(EL_ann=0.02, term=1, df_loss_month=df)
+    pc = PremiumCalculations(bond)
+
+    pc.calc_benchmark_premium(target_sharpe=0.2)
+
+    assert pc.benchmark_prem_rate > 0
+    assert pc.benchmark_prem_rate < 1.0  # sanity bound
+
+
+if __name__ == "__main__":
+    test_calc_benchmark_premium_converges()
+    LOGGER.info("test_calc_benchmark_premium_converges passed")
+    test_find_sharpe_single_loss_event()
+    LOGGER.info("test_find_sharpe_single_loss_event passed")