Update app.py

app.py

@@ -102,13 +102,10 @@ def filter_semantically_similar_texts_by_embedding(df, mode, embedding_field='em
 
 def search_kpi(kpi_query, kpi_count, mode):
     if mode == "BGE":
-        print("BGE 검색 시작")
         results = kpi_pool_ori.similarity_search_with_relevance_scores(kpi_query, k=50)
     elif mode == "SBERT-snunlp":
-        print("SBERT-snunlp 검색 시작")
         results = kpi_pool.similarity_search_with_relevance_scores(kpi_query, k=50)
     else:
-        print("SBERT-jhgan 검색 시작")
         results = kpi_pool2.similarity_search_with_relevance_scores(kpi_query, k=50)
 
 
@@ -135,13 +132,10 @@ def search_kpi(kpi_query, kpi_count, mode):
 
 def search_kpi_one(kpi_query, kpi_count, mode):
     if mode == "BGE":
-        print("BGE 검색 시작")
         results = kpi_pool_ori.similarity_search_with_relevance_scores(kpi_query, k=50)
     elif mode == "SBERT-snunlp":
-        print("SBERT-snunlp 검색 시작")
         results = kpi_pool.similarity_search_with_relevance_scores(kpi_query, k=50)
     else:
-        print("SBERT-jhgan 검색 시작")
         results = kpi_pool2.similarity_search_with_relevance_scores(kpi_query, k=50)
 
     # 메타데이터 + 점수 추출
@@ -232,11 +226,10 @@ def generate_excel(df1, df2, df3, kpi_list1, kpi_list2, kpi_list3, kpi_query):
         if kpi_list:
             indices = [int(i) - 1 for i in kpi_list]  # -1 보정
             filtered = df.iloc[indices].copy()
-            filtered["출처"] = model_name
             return filtered
         else:
             # 선택된 KPI 없을 때: 빈 DataFrame 반환
-            return pd.DataFrame(columns=list(df.columns) + ["출처"])
+            return pd.DataFrame(columns=list(df.columns))
             
     # 인덱스(-1 보정)로 DataFrame 필터링
     #filtered_df = df.iloc[[int(i) - 1 for i in kpi_list]] if kpi_list else pd.DataFrame(columns=df.columns)