Update _app.py

_app.py

@@ -134,12 +134,17 @@ def add_query(to_add, history):
     
 #     return q_a_df, answers_df, summary
 
-def qa_summarise(selected_queries, qa_llm_model, text_field, data_df):
+def qa_summarise(selected_queries, qa_llm_model, text_field, response_lang, data_df):
     
     qa_input_df = data_df[data_df["model_label"] != "none"].reset_index()
     texts = qa_input_df[text_field].to_list()
     
-    summary = generate_answer(qa_llm_model, texts, selected_queries[0], selected_queries, mode="multi_summarize")
+    summary = generate_answer(qa_llm_model, 
+                              texts, 
+                              selected_queries[0], 
+                              selected_queries, 
+                              response_lang, 
+                              mode="multi_summarize")
 
     doc_df = pd.DataFrame()
     doc_df["number"] = [i+1 for i in range(len(texts))]
@@ -228,7 +233,7 @@ with gr.Blocks() as demo:
             outputs=[incorrect, correct, accuracy, data_eval, download_csv])
 
     qa_tab = gr.Tab("Question Answering")
-    
+
     
     with qa_tab:
         # XXX Add some button disabling here, if the classification process is not completed first XXX
@@ -251,6 +256,8 @@ with gr.Blocks() as demo:
                 with gr.Column():
                     batch_size = gr.Slider(50, 500, value=150, step=1, label="Batch size", info="Choose between 50 and 500", interactive=True)
                     topk = gr.Slider(1, 10, value=5, step=1, label="Number of results to retrieve", info="Choose between 1 and 10", interactive=True)
+
+            response_lang = gr.Dropdown(["english", "german", "catalan", "spanish"], label="Response language", value="english", interactive=True)
                     
         selected_queries = gr.CheckboxGroup(label="Select at least one query using the checkboxes", interactive=True)
         queries_state = gr.State()
@@ -269,7 +276,7 @@ with gr.Blocks() as demo:
         #                 inputs=[selected_queries, qa_llm_model, aggregator, batch_size, topk, text_field, data], 
         #                 outputs=[qa_df, answers_df, hsummary])
         qa_button.click(qa_summarise, 
-                        inputs=[selected_queries, qa_llm_model, text_field, data], 
+                        inputs=[selected_queries, qa_llm_model, text_field, response_lang, data], 
                         outputs=[hsummary, qa_df])