Better tests for load_profile

src/simstadt/results/load_profile.py

@@ -21,18 +21,37 @@ class LoadProfileResults(SimStadtResults):
 
     def _parse_results(self) -> pd.DataFrame:
         csv_decimal = detect_decimal(self.csv_path, 6, "Area")
+
+        header = pd.read_csv(
+            self.csv_path, nrows=9, sep=";", header=None, index_col=0, decimal=csv_decimal
+        )
+        header.index = header.index.str.lstrip("# ")
+
+        def _header_sum(label: str) -> float:
+            return pd.to_numeric(header.loc[label]).fillna(0).sum()
+
         df = pd.read_csv(self.csv_path, skiprows=range(1, 12), sep=";", decimal=csv_decimal)
-        # First column is the timestamp; remaining columns are per-building loads
-        return df.drop(columns=[df.columns[0]])
+        df = df.drop(columns=[df.columns[0]])
+
+        df.attrs["number_of_households"] = int(_header_sum("Number of Households"))
+        df.attrs["number_of_people"] = int(_header_sum("Number of Occupants"))
+        df.attrs["total_heated_area"] = float(_header_sum("Heated Area"))
+
+        return df
 
     @property
     def kpis(self) -> list[KPI]:
-        # TODO: Add sum, average. Add timestep too?
-        # TODO: Add people? Add total heated area?
         df = self.dataframe
         total = df.sum().sum()
-        return [
+        total_heated_area = df.attrs["total_heated_area"]
+        result = [
             KPI("Number of buildings", df.shape[1], precision=0),
+            KPI("Number of households", df.attrs["number_of_households"], precision=0),
+            KPI("Number of inhabitants", df.attrs["number_of_people"], precision=0),
+            KPI("Total heated area", total_heated_area, "m²", precision=0),
             KPI("Total load", total, "kWh / a", precision=1),
             KPI("Average load", total / 8760, "kWh / h", precision=1),
         ]
+        if total_heated_area > 0:
+            result.append(KPI("Specific load", total / total_heated_area, "kWh / (m² · a)", precision=1))
+        return result

tests/test_results.py

@@ -485,10 +485,18 @@ def test_loadprofile_kpis(loadprofile_results):
     kpis = kpis_by_name(loadprofile_results)
     assert kpis["Number of buildings"].value == 3
     assert kpis["Number of buildings"].unit is None
+    assert kpis["Number of households"].value == 8
+    assert kpis["Number of households"].unit is None
+    assert kpis["Number of inhabitants"].value == 13
+    assert kpis["Number of inhabitants"].unit is None
+    assert kpis["Total heated area"].value == pytest.approx(1198.1, abs=0.1)
+    assert kpis["Total heated area"].unit == "m²"
     assert kpis["Total load"].value == pytest.approx(17004.0, abs=1.0)
     assert kpis["Total load"].unit == "kWh / a"
     assert kpis["Average load"].value == pytest.approx(1.9, abs=0.1)
     assert kpis["Average load"].unit == "kWh / h"
+    assert kpis["Specific load"].value == pytest.approx(14.2, abs=0.2)
+    assert kpis["Specific load"].unit == "kWh / (m² · a)"
 
 
 def test_loadprofile_no_diagrams(loadprofile_results):
@@ -514,7 +522,7 @@ def test_loadprofile_serialization(loadprofile_results):
     assert isinstance(reconstructed, LoadProfileResults)
     assert reconstructed.description == loadprofile_results.description
     assert len(reconstructed.output_files) == len(loadprofile_results.output_files)
-    assert len(reconstructed.kpis) == 3
+    assert len(reconstructed.kpis) == 7
     pd.testing.assert_frame_equal(
         reconstructed.dataframe, loadprofile_results.dataframe
     )

tests/test_runner.py

@@ -42,7 +42,7 @@ def test_citygml_with_explicit_project_path(mock_simstadt, tmp_path):
     assert results.workflow_path.parent == proj
 
     assert len(results.dataframe) > 0
-    assert len(results.kpis) == 3
+    assert len(results.kpis) == 7
     assert results.kpi("Average load").unit == "kWh / h"