Update app.py

app.py

@@ -3,8 +3,46 @@ import requests
 import json
 import os
 from datetime import datetime, timedelta
+from huggingface_hub import InferenceClient
 
 API_KEY = os.getenv("SERPHOUSE_API_KEY")
+hf_client = InferenceClient("CohereForAI/c4ai-command-r-plus-08-2024", token=os.getenv("HF_TOKEN"))
+
+# 국가별 언어 코드 매핑
+COUNTRY_LANGUAGES = {
+    "South Korea": "ko",
+    "Japan": "ja",
+    "China": "zh",
+    "Russia": "ru",
+    "France": "fr",
+    "Germany": "de",
+    "Spain": "es",
+    "Italy": "it",
+    "Netherlands": "nl",
+    "Portugal": "pt",
+    "Thailand": "th",
+    "Vietnam": "vi",
+    "Indonesia": "id",
+    "Malaysia": "ms",
+    "Saudi Arabia": "ar",
+    "United Arab Emirates": "ar",
+    "Egypt": "ar",
+    "Morocco": "ar",
+    "Greece": "el",
+    "Poland": "pl",
+    "Czech Republic": "cs",
+    "Hungary": "hu",
+    "Turkey": "tr",
+    "Romania": "ro",
+    "Bulgaria": "bg",
+    "Croatia": "hr",
+    "Serbia": "sr",
+    "Slovakia": "sk",
+    "Slovenia": "sl",
+    "Estonia": "et",
+    "Latvia": "lv",
+    "Lithuania": "lt"
+}
 
 COUNTRY_LOCATIONS = {
     "United States": "United States",
@@ -78,16 +116,37 @@ COUNTRY_LOCATIONS = {
 
 MAJOR_COUNTRIES = list(COUNTRY_LOCATIONS.keys())
 
+def translate_query(query, country):
+    try:
+        if country in COUNTRY_LANGUAGES:
+            target_lang = COUNTRY_LANGUAGES[country]
+            prompt = f"Translate the following English text to {target_lang} language. Only output the translated text without any explanations or quotes: {query}"
+            
+            translated = hf_client.text_generation(
+                prompt,
+                max_new_tokens=100,
+                temperature=0.3
+            )
+            return translated.strip()
+        return query
+    except Exception as e:
+        print(f"Translation error: {str(e)}")
+        return query
 
 def search_serphouse(query, country, page=1, num_result=10):
     url = "https://api.serphouse.com/serp/live"
     
+    # 검색어 번역
+    translated_query = translate_query(query, country)
+    print(f"Original query: {query}")
+    print(f"Translated query: {translated_query}")
+    
     payload = {
         "data": {
-            "q": query,
+            "q": translated_query,
             "domain": "google.com",
-            "loc": COUNTRY_LOCATIONS.get(country, "United States"),  # loc 사용
-            "lang": "en",
+            "loc": COUNTRY_LOCATIONS.get(country, "United States"),
+            "lang": COUNTRY_LANGUAGES.get(country, "en"),
             "device": "desktop",
             "serp_type": "news",
             "page": "1",
@@ -109,11 +168,11 @@ def search_serphouse(query, country, page=1, num_result=10):
         print("Response status:", response.status_code)
         
         response.raise_for_status()
-        return response.json()
+        return response.json(), translated_query
     except requests.RequestException as e:
-        return {"error": f"Error: {str(e)}"}
+        return {"error": f"Error: {str(e)}"}, query
 
-def format_results_from_raw(results):
+def format_results_from_raw(results, translated_query):
     if isinstance(results, dict) and "error" in results:
         return "Error: " + results["error"], []
 
@@ -131,15 +190,16 @@ def format_results_from_raw(results):
                 "snippet": result.get("snippet", "내용 없음"),
                 "channel": result.get("channel", result.get("source", "알 수 없음")),
                 "time": result.get("time", result.get("date", "알 수 없는 시간")),
-                "image_url": result.get("img", result.get("thumbnail", ""))
+                "image_url": result.get("img", result.get("thumbnail", "")),
+                "translated_query": translated_query
             })
         return "", articles
     except Exception as e:
         return f"결과 처리 중 오류 발생: {str(e)}", []
 
 def serphouse_search(query, country):
-    results = search_serphouse(query, country)
-    return format_results_from_raw(results)
+    results, translated_query = search_serphouse(query, country)
+    return format_results_from_raw(results, translated_query)
 
 css = """
 footer {visibility: hidden;}
@@ -154,7 +214,6 @@ with gr.Blocks(theme="Nymbo/Nymbo_Theme", css=css, title="NewsAI 서비스") as
             country = gr.Dropdown(MAJOR_COUNTRIES, label="국가", value="South Korea")
             search_button = gr.Button("검색", variant="primary")
 
-        # 프로그레스바 추가
         progress = gr.Progress()
         status_message = gr.Markdown(visible=False)
         articles_state = gr.State([])
@@ -200,9 +259,11 @@ with gr.Blocks(theme="Nymbo/Nymbo_Theme", css=css, title="NewsAI 서비스") as
                     image_url = article['image_url']
                     image_update = gr.update(value=image_url, visible=True) if image_url and not image_url.startswith('data:image') else gr.update(value=None, visible=False)
 
+                    translated_info = f"\n**번역된 검색어:** {article['translated_query']}" if article['translated_query'] != query else ""
+                    
                     outputs.extend([
                         gr.update(visible=True),
-                        gr.update(value=f"### [{article['title']}]({article['link']})"),
+                        gr.update(value=f"### [{article['title']}]({article['link']}){translated_info}"),
                         image_update,
                         gr.update(value=f"**요약:** {article['snippet']}"),
                         gr.update(value=f"**출처:** {article['channel']} | **시간:** {article['time']}")