working on heating controls

epc_data/app.py

@@ -56,7 +56,7 @@ def handler():
     df = df.reset_index(drop=True)
 
     import numpy as np
-    idx = 5
+    idx = 3
     record = df[df.index == idx].to_dict("records")[0]
     record = {k: v for k, v in record.items() if v not in [None, np.nan, False]}
     from pprint import pprint

epc_data/attributes/MainheatControlAttributes.py

@@ -1,5 +1,5 @@
 from typing import Dict, Union
-from epc_data.attributes.attribute_utils import clean_description
+from epc_data.attributes.attribute_utils import clean_description, remove_punctuation
 
 
 class MainheatControlAttributes:
@@ -46,6 +46,11 @@ class MainheatControlAttributes:
         'use of community heating'
     ]
 
+    MULTIPLE_ROOM_THERMOSTATS_PHRASES = [
+        'at least two room stats',
+        'at least two room thermostats'
+    ]
+
     def __init__(self, description: str):
         self.description: str = clean_description(description.lower())
 
@@ -57,7 +62,8 @@ class MainheatControlAttributes:
                 self.THERMOSTATIC_CONTROL_KEYWORDS,
                 self.CHARGING_SYSTEM_KEYWORDS,
                 self.SWITCH_SYSTEM_KEYWORDS,
-                self.DHW_CONTROL_KEYWORDS
+                self.DHW_CONTROL_KEYWORDS,
+                self.COMMUNITY_HEATING_KEYWORDS
             ]
         ):
             raise ValueError('Invalid description')
@@ -73,13 +79,31 @@ class MainheatControlAttributes:
             "no_control": self._find_keyword(self.NO_CONTROL_SYSTEM_KEYWORDS),
             "dhw_control": self._find_keyword(self.DHW_CONTROL_KEYWORDS),
             "community_heating": self._find_keyword(self.COMMUNITY_HEATING_KEYWORDS),
+            "multiple_room_thermostats": any(
+                phrase in self.description for phrase in self.MULTIPLE_ROOM_THERMOSTATS_PHRASES),
         }
 
         return result
 
     def _find_keyword(self, keywords):
+        description_words = set(self.description.split())
+
+        # Sort keywords by length, longest first. This ensures that 'time and temperature zone control'
+        # will be checked before 'temperature zone control' if both are present in the keywords list
+        keywords.sort(key=len, reverse=True)
+
         for keyword in keywords:
-            if keyword in self.description:
+            keyword_words = set(keyword.split())
+            if keyword_words.issubset(description_words):
+                return keyword
+
+        # If no keyword is found, try again after removing punctuation
+        description_without_punct = remove_punctuation(self.description)
+        description_words_without_punct = set(description_without_punct.split())
+
+        for keyword in keywords:
+            keyword_words = set(keyword.split())
+            if keyword_words.issubset(description_words_without_punct):
                 return keyword
 
         return None

epc_data/attributes/attribute_utils.py

@@ -1,4 +1,5 @@
 import re
+import string
 from typing import Tuple, Union, Dict, List
 
 THERMAL_TRANSMITTENCE_STR = r"average thermal transmittance (-?\d+\.\d+)\s(w/m-¦k)"
@@ -93,3 +94,13 @@ def process_part(result: Dict[str, Union[str, bool]], part: str, attr_list: List
         raise ValueError("No attribute matches found")
 
     return result
+
+
+def remove_punctuation(text: str) -> str:
+    # Create a translation table using the string.punctuation string
+    translation_table = str.maketrans("", "", string.punctuation)
+
+    # Use the translation table to remove punctuation from the text
+    text_without_punctuation = text.translate(translation_table)
+
+    return text_without_punctuation