georgm8/fix emergency care features

avoidable_admissions/data/validate.py

@@ -205,6 +205,10 @@ class Config:
             ],
         ),
         "opertn_count": pa.Column(int, nullable=False, checks=[pa.Check.ge(0)]),
+        "opertn_cat": pa.Column(str, nullable=False, checks=[pa.Check.isin(['Yes', 'No', 'Missing'])]),
+
+        "comorb_count": pa.Column(int, nullable=False, checks=[pa.Check.ge(0)]),
+        "comorb_cat": pa.Column(str, nullable=False, checks=[pa.Check.isin(['Yes', 'No', 'Missing'])]),           
     }
 )
 

avoidable_admissions/features/admitted_care_features.py

@@ -140,9 +140,27 @@ def _procedures(df: pd.DataFrame) -> pd.DataFrame:
         .count(axis=1)
     )
 
+    rules = {
+        'Yes': df['opertn_count'] > 0,
+        'No': df['opertn_count'] <= 0,
+        'Missing': df['opertn_count'].isna()
+    }
+
+    df['opertn_cat'] = np.select(
+        list(rules.values()), list(rules.keys()), default='Missing'
+    )    
+
+    return df
+
+def _comorbidities(df: pd.DataFrame) -> pd.DataFrame:        
+    diag_cols = [f'diag_{i:02d}' for i in range(2, 21)]
+    df['comorb_count'] = df[diag_cols].count(axis=1)
+    df['comorb_cat'] = df['comorb_count'].apply(lambda x: 'Yes' if x > 0 else 'No')
+
     return df
 
 
+
 def build_all(df: pd.DataFrame) -> pd.DataFrame:
 
     df = (
@@ -157,6 +175,7 @@ def build_all(df: pd.DataFrame) -> pd.DataFrame:
         .pipe(_dismeth)
         .pipe(_acsc_code)
         .pipe(_procedures)
+        .pipe(_comorbidities)
     )
 
     return df

avoidable_admissions/features/emergency_care_features.py

@@ -164,7 +164,7 @@ def _cc_code(df: pd.DataFrame) -> pd.DataFrame:
 
     # TODO: This section needs manual review of a good sample size to ensure it works
 
-    cc_mapping = load_ed_cc_mapping()
+    cc_mapping = feature_maps.load_ed_cc_mapping()
     df["edchiefcomplaint_cat"] = replace_values(df.edchiefcomplaint, cc_mapping)
 
     return df