windows class working nicely

epc_data/attributes/WindowAttributes.py

@@ -1,11 +1,20 @@
 from typing import Dict, Union
-from epc_data.attributes.attribute_utils import clean_description, process_part
+from epc_data.attributes.attribute_utils import clean_description
 
 
 class WindowAttributes:
     GLAZING_KEYWORDS = ["glazing", "glazed", "glaze"]
-    GLAZING_COVERAGE = ["fully", "mostly", "partial", "some"]
+    GLAZING_COVERAGE = ["fully", "mostly", "partial", "some", "full", "thoughout"]
-    GLAZING_TYPES = ["double", "triple", "secondary", "multiple", "high performance"]
+    GLAZING_TYPES = ["double", "triple", "secondary", "multiple", "high performance", "single"]
+
+    coverage_map = {
+        "full": "full",
+        "fully": "full",
+        "mostly": "most",
+        "partial": "partial",
+        "some": "partial",
+        "throughout": "full"
+    }
 
     def __init__(self, description: str):
         self.description: str = clean_description(description.lower())
@@ -23,22 +32,34 @@ class WindowAttributes:
 
     def process(self) -> Dict[str, Union[str, bool]]:
         result: Dict[str, Union[str, bool]] = {
-            f'has_{wt.replace(" ", "_")}': False for wt in self.GLAZING_KEYWORDS
+            "has_glazing": False,
+            "glazing_coverage": None,
+            "glazing_type": None,
+            "no_data": self.nodata
         }
-        result.update({f'is_{gc.replace(" ", "_")}': False for gc in self.GLAZING_COVERAGE})
-        result.update({f'is_{gt.replace(" ", "_")}': False for gt in self.GLAZING_TYPES})
-        result["no_data"] = self.nodata
 
         if self.nodata:
             return result
 
-        description = self.description.split(',')
+        # We consolidate GLAZING_KEYWORDS into a single attribute
+        result["has_glazing"] = any(keyword in self.description for keyword in self.GLAZING_KEYWORDS)
 
-        # Process each part separately
+        # For coverage and type, we will only store the first one we find
-        for part in description:
+        for part in self.description.split(','):
             part = part.strip()  # remove leading/trailing white spaces
-            process_part(result, part, self.GLAZING_KEYWORDS, 'has_')
+            if not result["glazing_coverage"]:
-            process_part(result, part, self.GLAZING_COVERAGE, 'is_')
+                for coverage in self.GLAZING_COVERAGE:
-            process_part(result, part, self.GLAZING_TYPES, 'is_')
+                    if coverage in part:
+                        result["glazing_coverage"] = self.coverage_map[coverage]
+                        break
+            if not result["glazing_type"]:
+                for glazing_type in self.GLAZING_TYPES:
+                    if glazing_type in part:
+                        result["glazing_type"] = glazing_type
+                        break
+
+        # If we didn't find any coverage or type, we assume full coverage
+        if not result["glazing_coverage"]:
+            result["glazing_coverage"] = "full"
 
         return result

epc_data/tests/test_data/test_window_attributes_cases.py

@@ -0,0 +1,39 @@
+windows_cases = [
+    {'original_description': '', 'has_glazing': False, 'glazing_coverage': None, 'glazing_type': None, 'no_data': True},
+    {'original_description': 'Full secondary glazing', 'has_glazing': True, 'glazing_coverage': 'full',
+     'glazing_type': 'secondary', 'no_data': False},
+    {'original_description': 'Fully double glazed', 'has_glazing': True, 'glazing_coverage': 'full',
+     'glazing_type': 'double', 'no_data': False},
+    {'original_description': 'Fully triple glazed', 'has_glazing': True, 'glazing_coverage': 'full',
+     'glazing_type': 'triple', 'no_data': False},
+    {'original_description': 'High performance glazing', 'has_glazing': True, 'glazing_coverage': 'full',
+     'glazing_type': 'high performance', 'no_data': False},
+    {'original_description': 'Mostly double glazing', 'has_glazing': True, 'glazing_coverage': 'most',
+     'glazing_type': 'double', 'no_data': False},
+    {'original_description': 'Mostly multiple glazing', 'has_glazing': True, 'glazing_coverage': 'most',
+     'glazing_type': 'multiple', 'no_data': False},
+    {'original_description': 'Mostly secondary glazing', 'has_glazing': True, 'glazing_coverage': 'most',
+     'glazing_type': 'secondary', 'no_data': False},
+    {'original_description': 'Mostly triple glazing', 'has_glazing': True, 'glazing_coverage': 'most',
+     'glazing_type': 'triple', 'no_data': False},
+    {'original_description': 'Multiple glazing throughout', 'has_glazing': True, 'glazing_coverage': 'full',
+     'glazing_type': 'multiple', 'no_data': False},
+    {'original_description': 'Partial double glazing', 'has_glazing': True, 'glazing_coverage': 'partial',
+     'glazing_type': 'double', 'no_data': False},
+    {'original_description': 'Partial multiple glazing', 'has_glazing': True, 'glazing_coverage': 'partial',
+     'glazing_type': 'multiple', 'no_data': False},
+    {'original_description': 'Partial secondary glazing', 'has_glazing': True, 'glazing_coverage': 'partial',
+     'glazing_type': 'secondary', 'no_data': False},
+    {'original_description': 'Partial triple glazing', 'has_glazing': True, 'glazing_coverage': 'partial',
+     'glazing_type': 'triple', 'no_data': False},
+    {'original_description': 'Single glazed', 'has_glazing': True, 'glazing_coverage': 'full', 'glazing_type': 'single',
+     'no_data': False},
+    {'original_description': 'Some double glazing', 'has_glazing': True, 'glazing_coverage': 'partial',
+     'glazing_type': 'double', 'no_data': False},
+    {'original_description': 'Some multiple glazing', 'has_glazing': True, 'glazing_coverage': 'partial',
+     'glazing_type': 'multiple', 'no_data': False},
+    {'original_description': 'Some secondary glazing', 'has_glazing': True, 'glazing_coverage': 'partial',
+     'glazing_type': 'secondary', 'no_data': False},
+    {'original_description': 'Some triple glazing', 'has_glazing': True, 'glazing_coverage': 'partial',
+     'glazing_type': 'triple', 'no_data': False}
+]