Create withlogging.py

withlogging.py

@@ -0,0 +1,55 @@
+import pandas as pd
+
+def date_dict(df_list, pif_key):
+    encounter_dates = []
+    for df in df_list:
+        df_date = df.loc[df['pif_key'].astype(str) == str(pif_key), 'encounter_date'].values
+        if len(df_date) > 0:
+            encounter_dates.extend(df_date)
+    if encounter_dates:
+        latest_date = max(encounter_dates)
+        return str(latest_date)
+    else:
+        return ''
+
+def report_date_check(dict_list, df_list, pif_keys_with_report_date, pif_keys_without_report_date):
+    col_names = {col['attribute_name'] for col in dict_list}
+    pif_key = next((col['attribute_prediction'] for col in dict_list if col['attribute_name'] == 'pif_key'), None)
+    if pif_key is not None:
+        if 'report_date' in col_names:
+            pif_keys_with_report_date.add(pif_key)
+        else:
+            pif_keys_without_report_date.add(pif_key)
+            date_insert_dict = date_dict(df_list, pif_key)
+            if date_insert_dict:
+                dict_list.append(date_insert_dict)
+    return dict_list, pif_keys_with_report_date, pif_keys_without_report_date
+
+def json_report_date_insertion(json_data, df_list):
+    pif_keys = set()
+    pif_keys_with_report_date = set()
+    pif_keys_without_report_date = set()
+
+    for biomarker_detail in json_data['patient_level']['biomarkers']['details']:
+        for attribute in biomarker_detail['attribute']:
+            attribute_details = attribute['attribute_details']
+            attribute_details, pif_keys_with_report_date, pif_keys_without_report_date = report_date_check(
+                attribute_details, df_list, pif_keys_with_report_date, pif_keys_without_report_date)
+            pif_keys.add(attribute_details['pif_key'])
+    
+    # Logging
+    print("1. No of unique pif_key present in a json:", len(pif_keys))
+    print("2. Pif_keys with report_date in json:", pif_keys_with_report_date)
+    print("3. Pif_keys without report_date in json:", pif_keys_without_report_date)
+
+    return json_data
+
+# Usage
+# Load dataframes df2022, df2023, df2024
+# df2022 = pd.read_csv('df2022.csv')
+# df2023 = pd.read_csv('df2023.csv')
+# df2024 = pd.read_csv('df2024.csv')
+
+# json_data = {}  # Load JSON data
+
+# json_report_date_insertion(json_data, [df2022, df2023, df2024])