Extended unit testing for roof class

epc_data/attributes/RoofAttributes.py

@@ -3,16 +3,28 @@ from typing import Dict, Union
 
 
 class RoofAttributes:
+    ROOF_TYPES = ['pitched', 'roof room', 'loft', 'flat', 'thatched', 'at rafters', 'assumed']
+    DWELLING_ABOVE = ["another dwelling above", "other premises above"]
 
     def __init__(self, description: str):
         """
         :param description: Description of the roof.
         """
-        self.description: str = description
+
+        self.description: str = description.lower()
+
+        if not description or not any(
+            rt in self.description for rt in self.ROOF_TYPES + self.DWELLING_ABOVE + ["average thermal transmittance"]
+        ):
+            raise ValueError('Invalid description')
 
     def process(self) -> Dict[str, Union[float, str, bool, None]]:
+
+        if not self.description:
+            raise ValueError("Description is empty")
+
         result: Dict[str, Union[float, str, bool, None]] = {}
-        description = self.description.lower()
+        description = self.description
 
         # thermal transmittance
         match = re.search(r"average thermal transmittance (-?\d+\.\d+)\s(w/m-¦k)", description)
@@ -26,9 +38,8 @@ class RoofAttributes:
             result['thermal_transmittance_unit'] = None
 
         # roof type
-        roof_types = ['pitched', 'roof room', 'loft', 'flat', 'thatched', 'at rafters', 'assumed']
 
-        for rt in roof_types:
+        for rt in self.ROOF_TYPES:
             result[f'is_{rt.replace(" ", "_")}'] = rt in description
             # Remove the roof type from the description
             description = description.replace(rt, "")
@@ -36,7 +47,7 @@ class RoofAttributes:
         result["has_dwelling_above"] = (
             "another dwelling above" in description or "other premises above" in description
         )
-        for dwelling_above in ["another dwelling above", "other premises above"]:
+        for dwelling_above in self.DWELLING_ABOVE:
             description = description.replace(dwelling_above, "")
 
         result["is_valid"] = "invalid" not in description

epc_data/tests/test_epc_clean.py

@@ -0,0 +1,33 @@
+import pytest
+import pickle
+from epc_data.EpcClean import EpcClean
+from pathlib import Path
+from epc_data.tests.test_data.test_roof_attributes_cases import clean_roof_test_cases
+from epc_data.attributes.RoofAttributes import RoofAttributes
+
+# For local testing
+if __file__ == "<input>":
+    input_data_path = Path("./epc_data/tests/test_data/EpcClean_inputs.obj")
+else:
+    current_file_path = Path(__file__)
+    input_data_path = current_file_path.parent / 'test_data' / 'EpcClean_inputs.obj'
+
+
+class TestEpcClean:
+
+    @staticmethod
+    def load_data(path):
+        with open(path, "rb") as file:
+            return pickle.load(file)
+
+    @pytest.fixture(autouse=True)
+    def setup_class(self):
+        self.cleaner = EpcClean(self.load_data(input_data_path))
+
+    def test_clean(self):
+        self.cleaner.clean()
+        assert len(self.cleaner.cleaned["roof-description"]) == len(self.cleaner.unique_vals["roof-description"])
+
+    def test__init_empty_cleaned_obj(self):
+        self.cleaner._init_empty_cleaned_obj()
+        assert all([len(values) == 0 for values in self.cleaner.cleaned.values()])

epc_data/tests/test_roof_attributes.py

@@ -13,24 +13,7 @@ else:
     input_data_path = current_file_path.parent / 'test_data' / 'EpcClean_inputs.obj'
 
 
-class TestEpcClean:
-
-    @staticmethod
-    def load_data(path):
-        with open(path, "rb") as file:
-            return pickle.load(file)
-
-    @pytest.fixture(autouse=True)
-    def setup_class(self):
-        self.cleaner = EpcClean(self.load_data(input_data_path))
-
-    def test_clean(self):
-        self.cleaner.clean()
-        assert len(self.cleaner.cleaned["roof-description"]) == len(self.cleaner.unique_vals["roof-description"])
-
-    def test__init_empty_cleaned_obj(self):
-        self.cleaner._init_empty_cleaned_obj()
-        assert all([len(values) == 0 for values in self.cleaner.cleaned.values()])
+class TestRoofAttributes:
 
     def test_clean_roof(self):
         result = RoofAttributes('Pitched, 270 mm loft insulation').process()
@@ -84,3 +67,28 @@ class TestEpcClean:
 
         for k in expected_output:
             assert result[k] == expected_output[k]
+
+    def test_clean_roof_invalid(self):
+        with pytest.raises(ValueError):
+            RoofAttributes('').process()
+        with pytest.raises(ValueError):
+            RoofAttributes('nonsense string').process()
+
+    def test_clean_roof_edge_cases(self):
+        # Insulation thickness edge case
+        assert RoofAttributes('Pitched, 99999 mm loft insulation').process()['insulation_thickness'] == "99999"
+        assert RoofAttributes('Pitched, 1 mm loft insulation').process()['insulation_thickness'] == "1"
+
+        # Special characters in the description - implement this later
+        # result = RoofAttributes('Pitched, **270** mm loft insulation').process()
+        # assert result['insulation_thickness'] == 270
+
+    def test_clean_roof_type(self):
+        for roof_type in ['pitched', 'roof room', 'loft', 'flat', 'thatched', 'at rafters', 'assumed']:
+            result = RoofAttributes(f'{roof_type}, 200 mm insulation').process()
+            assert result[f'is_{roof_type.replace(" ", "_")}'] is True
+
+    def test_clean_roof_thermal_transmittance(self):
+        result = RoofAttributes('Pitched, 200 mm insulation, average thermal transmittance 0.13 w/m-¦k').process()
+        assert result['thermal_transmittance'] == 0.13
+        assert result['thermal_transmittance_unit'] == 'w/m-¦k'