Update app.py

app.py

@@ -8,7 +8,6 @@ import time
 import requests
 import urllib.parse
 import pandas as pd
-import re  # 新增 re 模組用於解析地址
 from langchain_huggingface import HuggingFaceEmbeddings
 from langchain_community.vectorstores import FAISS
 from PIL import Image
@@ -93,86 +92,25 @@ def check_api_key_status(name, key):
     return f"✅ 已設定 ({masked})"
 
 # ==========================================
-# 2. Google Maps 爬蟲與資料處理功能
+# 2. Google Maps 爬蟲功能
 # ==========================================
-def extract_city(address):
-    """
-    通用版地址解析：支援中文(台/臺)、英文地址、與郵遞區號判斷
-    """
-    if not isinstance(address, str) or not address:
-        return "未知地區"
-    
-    # 1. 預處理：轉小寫以利英文比對，將 '臺' 統一轉為 '台' 以利中文比對
-    addr_norm = address.lower().replace("臺", "台")
-    
-    # 定義對照表 (關鍵字 -> 標準縣市名稱)
-    # 順序很重要：縣市層級要比區層級優先，但這裡我們主要抓大區域
-    city_map = {
-        "keelung": "基隆市", "基隆": "基隆市",
-        "taipei": "台北市", "台北": "台北市",
-        "new taipei": "新北市", "新北": "新北市",
-        "taoyuan": "桃園市", "桃園": "桃園市",
-        "hsinchu city": "新竹市", "新竹市": "新竹市",
-        "hsinchu county": "新竹縣", "新竹縣": "新竹縣",
-        "hsinchu": "新竹縣市", # 模糊比對放後面
-        "miaoli": "苗栗縣", "苗栗": "苗栗縣",
-        "taichung": "台中市", "台中": "台中市",
-        "changhua": "彰化縣", "彰化": "彰化縣",
-        "nantou": "南投縣", "南投": "南投縣",
-        "yunlin": "雲林縣", "雲林": "雲林縣",
-        "chiayi city": "嘉義市", "嘉義市": "嘉義市",
-        "chiayi county": "嘉義縣", "嘉義縣": "嘉義縣",
-        "chiayi": "嘉義縣市",
-        "tainan": "台南市", "台南": "台南市",
-        "kaohsiung": "高雄市", "高雄": "高雄市",
-        "pingtung": "屏東縣", "屏東": "屏東縣",
-        "yilan": "宜蘭縣", "宜蘭": "宜蘭縣",
-        "hualien": "花蓮縣", "花蓮": "花蓮縣",
-        "taitung": "台東縣", "台東": "台東縣",
-        "penghu": "澎湖縣", "澎湖": "澎湖縣",
-        "kinmen": "金門縣", "金門": "金門縣",
-        "lienchiang": "連江縣", "連江": "連江縣", "matsu": "連江縣"
-    }
-    
-    # 2. 關鍵字比對
-    for key, std_city in city_map.items():
-        if key in addr_norm:
-            return std_city
-            
-    # 3. 補救措施：正規表達式抓取 "3碼郵遞區號" (台灣地址常出現)
-    # 例如：106台北市大安區 -> 106 可能是關鍵
-    # 這邊簡單實作，若需要更精準可能需要郵遞區號對照表，但通常上述關鍵字已足夠
-    
-    return "其他地區"
-
 def sync_google_maps(url):
     global global_df, global_retriever, rag_initialized
     clean_url = url.strip() if url else ""
-    
-    no_update = gr.update()
-    empty_df = pd.DataFrame() # 空表格
-    
     if not clean_url or "http" not in clean_url.lower():
-        # 注意：這裡多回傳一個 empty_df
-        yield "❌ 請輸入有效的 Google Maps 分享連結。", no_update, empty_df
+        yield "❌ 請輸入有效的 Google Maps 分享連結。"
         return
 
     try:
-        yield "🚀 [1/4] 啟動瀏覽器...", no_update, empty_df
+        yield "🚀 [1/4] 啟動瀏覽器..."
         with sync_playwright() as p:
-            # 加上 user_agent 模擬正常瀏覽器，減少變成英文版的機率 (雖然我們已經做了英文處理)
             browser = p.chromium.launch(headless=True)
-            context = browser.new_context(
-                locale='zh-TW', # 嘗試強制指定繁體中文環境
-                timezone_id='Asia/Taipei'
-            )
-            page = context.new_page()
-            
-            yield "🌐 [2/4] 連線中...", no_update, empty_df
+            page = browser.new_page()
+            yield "🌐 [2/4] 連線中..."
             page.goto(clean_url, wait_until="domcontentloaded", timeout=60000)
-            yield "⏳ [3/4] 等待列表加載 (約 10 秒)...", no_update, empty_df
+            yield "⏳ [3/4] 等待列表加載 (約 10 秒)..."
             time.sleep(10) 
-            yield "📄 [4/4] 解析餐廳資訊...", no_update, empty_df
+            yield "📄 [4/4] 解析餐廳資訊..."
             
             titles = page.locator('div.fontHeadlineSmall').all_inner_texts()
             details = page.locator('div.fontBodyMedium').all_inner_texts()
@@ -181,36 +119,23 @@ def sync_google_maps(url):
             for i, name in enumerate(titles):
                 name = name.strip()
                 if name:
-                    raw_addr = details[i].strip() if i < len(details) else ""
-                    # 這裡使用新的 extract_city
-                    city = extract_city(raw_addr)
-                    
-                    search_query = f"{name} {raw_addr}"
+                    addr = details[i].strip() if i < len(details) else ""
+                    search_query = f"{name} {addr}"
                     restaurant_list.append({
                         "Name": name,
-                        "City": city,      # 顯示這欄很重要
-                        "Address": raw_addr, # 顯示原始地址以供除錯
+                        "Address": addr,
                         "Category": "未分類",
-                        "RAG_Content": f"餐廳：{name}，地區：{city}，資訊：{raw_addr}",
+                        "RAG_Content": f"餐廳：{name}，資訊：{addr}",
                         "URL": f"https://www.google.com/maps/search/?api=1&query={urllib.parse.quote(search_query)}"
                     })
             browser.close()
 
             if not restaurant_list:
-                yield "⚠️ 找不到餐廳。請確認連結格式正確且已公開。", no_update, empty_df
+                yield "⚠️ 找不到餐廳。請確認連結格式正確且已公開。"
                 return
 
             global_df = pd.DataFrame(restaurant_list)
             
-            # 準備下拉選單
-            unique_cities = sorted(global_df['City'].unique().tolist())
-            if "其他地區" in unique_cities:
-                unique_cities.remove("其他地區")
-                unique_cities.append("其他地區")
-            
-            dropdown_choices = ["全區"] + unique_cities
-            
-            # RAG 初始化 (維持不變)
             try:
                 embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
                 vectorstore = FAISS.from_texts(global_df['RAG_Content'].tolist(), embeddings, metadatas=global_df.to_dict('records'))
@@ -219,23 +144,14 @@ def sync_google_maps(url):
                 print(f"RAG 初始化警告: {e}")
 
             rag_initialized = True
-            
-            # 簡化顯示用的 DataFrame (只秀出重點欄位給使用者看)
-            display_df = global_df[['Name', 'City', 'Address']].head(20) # 只顯示前20筆避免太長
-            
-            # 回傳：訊息, 下拉選單更新, 表格資料更新
-            yield f"✅ 同步成功！共 {len(restaurant_list)} 筆。", gr.update(choices=dropdown_choices, value="全區", visible=True), display_df
-            
+            yield f"✅ 同步成功！已載入 {len(restaurant_list)} 間餐廳。"
     except Exception as e:
-        yield f"❌ 系統錯誤：{str(e)}", no_update, empty_df
+        yield f"❌ 系統錯誤：{str(e)}"
 
 # ==========================================
 # 3. 核心功能 (AI Agent)
 # ==========================================
-def get_restaurant_data(mood_score_str, food_choice, location_filter="全區"):
-    """
-    依據心情、食物類別與「地區」篩選餐廳
-    """
+def get_restaurant_data(mood_score_str, food_choice):
     init_static_data() # 確保心情指南已載入
     
     if global_df is None or global_df.empty: 
@@ -256,15 +172,6 @@ def get_restaurant_data(mood_score_str, food_choice, location_filter="全區"):
             mood_reason = mood_info.iloc[0]['原因']
     
     candidates = global_df.copy()
-    
-    # --- 新增：地區篩選邏輯 ---
-    if location_filter and location_filter != "全區":
-        candidates = candidates[candidates['City'] == location_filter]
-        # 如果篩選後沒資料，則退回全部資料並標記 Random (或者回傳空值讓前端顯示警告)
-        if candidates.empty:
-            return None, True, "", f"在【{location_filter}】找不到餐廳，請嘗試切換地區。"
-    # ------------------------
-
     food_keyword = "飯" if food_choice == "吃飯" else "麵" if food_choice == "吃麵" else ""
     
     is_random = False
@@ -276,11 +183,7 @@ def get_restaurant_data(mood_score_str, food_choice, location_filter="全區"):
             is_random = True 
     
     if candidates.empty:
-        # 如果經過兩層篩選都沒東西，就從 global_df 隨便抓 (但盡量遵守地區)
-        backup_pool = global_df[global_df['City'] == location_filter] if location_filter != "全區" else global_df
-        if backup_pool.empty: backup_pool = global_df
-        
-        result = backup_pool.sample(1).iloc[0]; is_random = True
+        result = global_df.sample(1).iloc[0]; is_random = True
     else:
         result = candidates.sample(1).iloc[0]
         
@@ -334,18 +237,14 @@ def generate_image_huggingface(prompt):
         # 回傳具體錯誤訊息
         return None, str(e)
 
-def mood_agent_logic(score_input, food_input, location_input, diary_input, debug_mode):
+def mood_agent_logic(score_input, food_input, diary_input, debug_mode):
     if not rag_initialized:
         yield "⚠️ 請先同步地圖清單！", None, "", gr.update()
         return
     
-    # 傳入 location_input 進行篩選
-    restaurant, is_random, rec_categories, mood_reason = get_restaurant_data(score_input, food_input, location_input)
-    
-    # 錯誤處��：如果 restaurant 是 None，通常是因為該地區沒餐廳
-    if restaurant is None:
-        error_msg = mood_reason # 這裡借用 mood_reason 回傳錯誤訊息
-        yield f"❌ {error_msg}", None, "", gr.update()
+    restaurant, is_random, rec_categories, mood_reason = get_restaurant_data(score_input, food_input)
+    if restaurant is None: 
+        yield "資料庫讀取錯誤", None, "", gr.update()
         return
     
     name = restaurant['Name']; url = restaurant['URL']
@@ -357,11 +256,8 @@ def mood_agent_logic(score_input, food_input, location_input, diary_input, debug
     ai_text, groq_debug_log = generate_content_with_groq(name, rag_info, diary_input, score_input, mood_reason, debug_mode)
     
     prefix = ""
-    if is_random:
-        if food_input != "隨便":
-            prefix = f"> 💡 **溫馨提示**：在 {location_input} 的清單中暫無『{food_input}』，已為您挑選其他類型的美味！\n\n"
-        else:
-             prefix = f"> 💡 **探索模式**：在 {location_input} 為您精選了一家餐廳！\n\n"
+    if is_random and food_input != "隨便":
+        prefix = f"> 💡 **溫馨提示**：清單中暫無『{food_input}』，已從現有名單挑選最適合的店！\n\n"
 
     # 組合完整的除錯資訊，包含 API Key 狀態與錯誤訊息
     full_debug_log = ""
@@ -375,7 +271,6 @@ def mood_agent_logic(score_input, food_input, location_input, diary_input, debug
 - **GROQ_API_KEY**: {groq_status}
 - **HF_TOKEN**: {hf_status}
 - **RAG 狀態**: {"✅ 已初始化" if rag_initialized else "❌ 未初始化"}
-- **目前篩選地區**: {location_input}
 """
         full_debug_log = api_debug_block + "\n" + groq_debug_log + f"\n\n### 🎨 Image Prompt Debug\n{img_prompt}"
 
@@ -474,19 +369,6 @@ with gr.Blocks(title="AI 心情食堂", css=css_) as demo:
                 with gr.Row():
                     map_url = gr.Textbox(label="Google Maps Saved Lists 連結", placeholder="請貼上清單的分享連結...", scale=3)
                     sync_btn = gr.Button("🔄 同步清單", variant="secondary", scale=1)
-                
-                # [NEW] 新增：資料預覽表格
-                # 放在下拉選單上面或下面都可以，這裡放在下面當作檢查用
-                with gr.Accordion("📋 點此查看已抓取的餐廳資料 (除錯用)", open=False):
-                    data_preview = gr.Dataframe(
-                        headers=["Name", "City", "Address"], 
-                        datatype=["str", "str", "str"],
-                        label="餐廳資料預覽 (顯示前 20 筆)"
-                    )
-                
-                # 新增：動態下拉選單
-                location_dropdown = gr.Dropdown(label="📍 選擇餐廳地區 (同步後自動更新)", choices=["請先同步清單"], value="請先同步清單", interactive=True)
-                
                 sync_msg = gr.Markdown("ℹ️ 尚未同步資料庫")
                 gr.Markdown("---")
                 gr.Markdown(f"## 🍱 今天想吃點什麼？")
@@ -506,22 +388,16 @@ with gr.Blocks(title="AI 心情食堂", css=css_) as demo:
                         map_output = gr.HTML(label="地圖導航")
 
     # 事件
-    # 修改：sync_btn 現在會同時更新 sync_msg 和 location_dropdown
-    sync_btn.click(
-        fn=sync_google_maps, 
-        inputs=[map_url], 
-        outputs=[sync_msg, location_dropdown, data_preview] # 這裡加上 data_preview
-    )
-    
+    sync_btn.click(fn=sync_google_maps, inputs=[map_url], outputs=[sync_msg])
     btn_start.click(fn=bridge_start_click, inputs=[st_state], outputs=[result_markdown, cam, btn_stop, btn_start, st_state, btn_go_dining])
     btn_stop.click(fn=bridge_stop_click, inputs=[st_state], outputs=[result_markdown, cam, btn_stop, btn_start, st_state])
     cam.stream(fn=bridge_predict_frame, inputs=[cam, st_state], outputs=[cam, result_markdown, st_state, btn_stop, btn_start, score_input, btn_go_dining])
     btn_go_dining.click(fn=lambda: gr.Tabs(selected=1), inputs=None, outputs=tabs)
     
-    # 修改：submit_btn 加入 location_dropdown 作為輸入
+    # debug_mode
     submit_btn.click(
         fn=mood_agent_logic, 
-        inputs=[score_input, food_input, location_dropdown, diary_input, debug_mode_btn], 
+        inputs=[score_input, food_input, diary_input, debug_mode_btn], 
         outputs=[agent_output, image_output, map_output, debug_output]
     )