Update app.py

app.py

@@ -4,13 +4,14 @@ import jieba.analyse
 import aiohttp
 import asyncio
 import ssl
-from aiohttp import ClientSession
 from functools import lru_cache
 
+# SSL context setup
 ssl_context = ssl.create_default_context()
 ssl_context.check_hostname = False
 ssl_context.verify_mode = ssl.CERT_NONE
 
+# Area data (unchanged)
 area_data = {
     '香港': {
         '中西區': ["西環", "堅尼地城", "石塘咀", "西營盤", "上環", "中環", "金鐘", "西半山", "中半山", "半山", "山頂"],
@@ -38,15 +39,14 @@ area_data = {
     }
 }
 
+
 @lru_cache(maxsize=None)
 def load_user_dict_terms():
     user_dict_terms = set()
     dictionaries = [
-        'flag/RVT_AddressCh.txt',
-        'flag/RVT_AddressEn.txt',
-        'flag/RVT_Area.txt',
-        'flag/RVT_BuildingCh.txt',
-        'flag/RVT_BuildingEn.txt'
+        'flag/RVT_AddressCh.txt', 'flag/RVT_AddressEn.txt', 'flag/RVT_Area.txt',
+        'flag/RVT_BuildingCh.txt', 'flag/RVT_BuildingEn.txt', 'flag/ChiVillage.txt',
+        'flag/ChiEstate.txt', 'flag/ChiStreet.txt', 'flag/ChiBuilding.txt'
     ]
     for file_path in dictionaries:
         try:
@@ -58,17 +58,17 @@ def load_user_dict_terms():
             print(f'Error reading file {file_path}: {e}')
     return user_dict_terms
 
+
 def setup_jieba_dictionaries():
     dictionaries = [
-        'flag/RVT_AddressCh.txt',
-        'flag/RVT_AddressEn.txt',
-        'flag/RVT_Area.txt',
-        'flag/RVT_BuildingCh.txt',
-        'flag/RVT_BuildingEn.txt'
+        'flag/RVT_AddressCh.txt', 'flag/RVT_AddressEn.txt', 'flag/RVT_Area.txt',
+        'flag/RVT_BuildingCh.txt', 'flag/RVT_BuildingEn.txt', 'flag/ChiVillage.txt',
+        'flag/ChiEstate.txt', 'flag/ChiStreet.txt', 'flag/ChiBuilding.txt'
     ]
     for file_path in dictionaries:
         jieba.load_userdict(file_path)
 
+
 def process_text(text):
     setup_jieba_dictionaries()
     user_dict_terms = load_user_dict_terms()
@@ -84,26 +84,24 @@ def process_text(text):
 
     return results
 
+
 def reformat_text(text):
     return [line.strip() for line in text.splitlines() if line.strip()]
 
+
 def process_text_only(text, reformat):
     extracted_keywords = process_text(text)
     if reformat:
         extracted_keywords = reformat_text('\n'.join(extracted_keywords))
     return '\n'.join(extracted_keywords)
 
+
 async def lookup_address(query, language='zh-Hant'):
     url = 'https://www.als.gov.hk/lookup'
-    headers = {
-        'Accept': 'application/json',
-        'Accept-Language': language
-    }
-    payload = {
-        'q': query
-    }
+    headers = {'Accept': 'application/json', 'Accept-Language': language}
+    payload = {'q': query}
 
-    async with ClientSession() as session:
+    async with aiohttp.ClientSession() as session:
         try:
             async with session.post(url, headers=headers, data=payload, ssl=ssl_context) as response:
                 if response.status == 200:
@@ -118,20 +116,11 @@ async def lookup_address(query, language='zh-Hant'):
             print(f'General Error: {e}')
             return {'error': f'General Error: {e}'}
 
+
 async def get_address_lookup_results(keywords):
     results = []
-    tasks = []
-
-    for keyword in keywords:
-        keyword = keyword.strip()
-        if not keyword:
-            continue
-
-        if not is_valid_for_lookup(keyword):
-
-            continue
-        else:
-            tasks.append(lookup_address(keyword))
+    tasks = [lookup_address(keyword.strip()) for keyword in keywords if
+             keyword.strip() and is_valid_for_lookup(keyword.strip())]
 
     lookup_results = await asyncio.gather(*tasks)
 
@@ -163,11 +152,16 @@ async def get_address_lookup_results(keywords):
 
             if len(result) > 1:  # Only add if there's more than just the Keyword
                 results.append(result)
-        else:
-            pass
 
     return results
 
+
+def is_valid_for_lookup(keyword):
+    return not any(keyword in districts.keys() or keyword in subdistrict
+                   for districts in area_data.values()
+                   for subdistrict in districts.values())
+
+
 async def gradio_function(text, reformat, perform_lookup):
     extracted_keywords = process_text_only(text, reformat)
     keywords_list = extracted_keywords.splitlines()
@@ -178,18 +172,11 @@ async def gradio_function(text, reformat, perform_lookup):
 
     return extracted_keywords, address_results
 
-def is_valid_for_lookup(keyword):
-    for region, districts in area_data.items():
-        if keyword in districts.keys():
-            return False
-        for subdistricts in districts.values():
-            if keyword in subdistricts:
-                return False
-    return True
 
 def gradio_interface(text, reformat, perform_lookup):
     return asyncio.run(gradio_function(text, reformat, perform_lookup))
 
+
 interface = gr.Interface(
     fn=gradio_interface,
     inputs=[