Raise clear error for empty Hazard in ImpactCalc.impact

Computing impact with a Hazard containing zero events previously
crashed deep inside ``np.array_split`` with the obscure message
"number sections must be larger than 0". This adds an early guard
in ``ImpactCalc.impact`` that raises a ``ValueError`` with a clear
explanation when ``self.hazard.size == 0``.

Closes #814

Author: ankurmahiwal118028

CHANGELOG.md

@@ -23,6 +23,7 @@ Code freeze date: YYYY-MM-DD
 
 - Fixed asset count in impact logging message [#1195](https://github.com/CLIMADA-project/climada_python/pull/1195).
 - `Hazard.from_raster_xarray` now returns a sparse matrix instead of a sparse array [#1261](https://github.com/CLIMADA-project/climada_python/pull/1261).
+- `ImpactCalc.impact` now raises a clear `ValueError` when the supplied `Hazard` contains no events, instead of failing later inside `np.array_split` with an obscure message [#814](https://github.com/CLIMADA-project/climada_python/issues/814).
 
 ### Deprecated
 - `Impact.calc_freq_curve()` should not be given the parameter `return_per`. Use the parameter `return_periods` in `Impact.calc_freq_curve().interpolate()` instead.

climada/engine/impact_calc.py

@@ -136,6 +136,15 @@ def impact(
         climada.entity.exposures.assign_centroids : assign centroids to exposures explicitly
         """
         # TODO: consider refactoring, making use of Exposures.hazard_impf
+        # check that the hazard contains at least one event; an empty hazard
+        # otherwise produces an obscure error from ``np.array_split`` deeper in
+        # the calculation (see GH #814).
+        if self.hazard.size == 0:
+            raise ValueError(
+                "Impact calculation not possible. The hazard object contains "
+                "no events. Please provide a Hazard with at least one event."
+            )
+
         # check for compatibility of exposures and hazard type
         if all(
             name not in self.exposures.gdf.columns

climada/engine/test/test_impact_calc.py

@@ -147,6 +147,22 @@ def test_error_handling_mismatch_haz_type(self):
                 "functions found for hazard type TC in impf_set.",
             )
 
+    def test_error_handling_empty_hazard(self):
+        """An empty Hazard must raise a clear ValueError (see GH #814)."""
+        haz_empty = Hazard("TC")
+        exp = Exposures()
+        exp.gdf["impf_TC"] = 1
+        impf = ImpactFunc(
+            haz_type="TC",
+            id=1,
+            intensity=np.array([0, 20]),
+            paa=np.array([0, 1]),
+            mdd=np.array([0, 0.5]),
+        )
+        impfset = ImpactFuncSet([impf])
+        with self.assertRaisesRegex(ValueError, "no events"):
+            ImpactCalc(exp, impfset, haz_empty).impact()
+
     def test_error_handling_mismatch_impf_ids(self):
         """Test error handling in case impf ids in exposures
         does not appear in impf_set"""