1. Update for mapping multilingual share class names.

2. Optimize getting currency logic

main.py

@@ -1203,10 +1203,10 @@ if __name__ == "__main__":
                         "501380497",
                         "514636959", 
                         "508981020"]
-    special_doc_id_list = ["514636993"]
+    special_doc_id_list = ["514636953"]
     output_mapping_child_folder = r"/data/emea_ar/output/mapping_data/docs/"
     output_mapping_total_folder = r"/data/emea_ar/output/mapping_data/total/"
-    re_run_extract_data = True
+    re_run_extract_data = False
     re_run_mapping_data = True
     force_save_total_data = False
     calculate_metrics = False

utils/biz_utils.py

@@ -97,9 +97,11 @@ def get_most_similar_name(text: str,
         
         for i in range(len(copy_name_list)):
             copy_name = copy_name_list[i]
-            share_part = get_share_part_list([copy_name])[0]
+            if matching_type == "share":
-            if '-' in share_part:
+                copy_name, _ = replace_share_name_for_multilingual(copy_name, None)
-                copy_name = copy_name.replace('-', ' ')
+                share_part = get_share_part_list([copy_name])[0]
+                if '-' in share_part:
+                    copy_name = copy_name.replace('-', ' ')
             copy_name = replace_abbrevation(copy_name)
             copy_name_list[i] = copy_name
 
@@ -128,6 +130,9 @@ def get_most_similar_name(text: str,
             share_name = remove_special_characters(share_name)
             share_name = replace_abbrevation(share_name)
             
+            text, share_name = replace_share_name_for_multilingual(text, share_name)
+            
+        
         text_splits = text.split()
         if len(text_splits) == 1:
             text = split_words_without_space(text)
@@ -332,6 +337,23 @@ def get_most_similar_name(text: str,
         return None, 0.0
 
 
+def replace_share_name_for_multilingual(text: str, share_name: str):
+    if text is None or len(text.strip()) == 0:
+        return text, share_name
+    
+    multilingual_share_list = ["Catégorie de parts", "Classe di quote", 
+                               "Kategorie Anteile", "Kategorie anteile",
+                               "Clase de participaciones", "Aandelenklasse", 
+                               "aandelenklasse", "Anteilklasse", "anteilklasse"]
+    for multilingual_share in multilingual_share_list:
+        if multilingual_share in text:
+            text = text.replace(multilingual_share, "Class")
+            if share_name is not None and len(share_name.strip()) > 0:
+                share_name = share_name.replace(multilingual_share, "Class")
+            break
+    return text, share_name
+
+
 def compare_both_short_name(text_short_name_list: list, compare_short_name_list: list):
     copy_text_short_name_list = deepcopy(text_short_name_list)
     copy_compare_short_name_list = deepcopy(compare_short_name_list)
@@ -448,13 +470,21 @@ def get_currency_from_text(text: str):
     text = text.strip()
     text_split = text.split()
     count = 0
+    currency_list = []
     for split in text_split[::-1]:
         if count == 4:
             break
         if split.upper() in total_currency_list:
-            return split
+            currency_list.append(split.upper())
         count += 1
-    return None
+    if len(currency_list) > 1:
+        # remove the first currency from currency list
+        currency_list.pop(0)
+        return currency_list[0]
+    elif len(currency_list) == 1:
+        return currency_list[0]
+    else:
+        return None
 
 
 def update_for_currency(text: str, share_name: str, compare_list: list):