Update app.py

app.py

@@ -125,25 +125,8 @@ def predict_faiss(model, tokenizer, embedding_model, df, question, index):
     }
     return output
 
-def predict(model, tokenizer, embedding_model, df, question, index):
-    t = time.time()
-    question = question.strip()
-    question_vector = get_embeddings(embedding_model, question)
-    question_vector = prepare_sentences_vector([question_vector])
-    distances,indices = faiss_search(index, question_vector)
 
-    # Answer = model_pipeline(model, tokenizer, df['Question'][indices[0][0]], df['Context'][indices[0][0]])
-    Answer = model_pipeline(model, tokenizer, question, df['Context'][indices[0][0]])
-    _time = time.time() - t
-    output = {
-        "user_question": question,
-        "answer": Answer,
-        "totaltime": round(_time, 3),
-        "distance": round(distances[0][0], 4)
-    }
-    return Answer
-
-def predict_test(model, tokenizer, embedding_model, df, question, index):  # sent_tokenize pythainlp
+def predict(model, tokenizer, embedding_model, df, question, index):  # sent_tokenize pythainlp
     t = time.time()
     question = question.strip()
     question_vector = get_embeddings(embedding_model, question)
@@ -206,12 +189,9 @@ def highlight_text(text, start_index, end_index):
             highlighted_text += "</mark>"
     return highlighted_text
 
-def chat_interface_before(question, history):
-    response = predict(model, tokenizer, embedding_model, df, question, index)
-    return response
 
-def chat_interface_after(question, history):
-    response = predict_test(model, tokenizer, embedding_model, df, question, index)
+def chat_interface(question, history):
+    response = predict(model, tokenizer, embedding_model, df, question, index)
     highlighted_answer = highlight_text(response["answer"], response["highlight_start"], response["highlight_end"])
     return highlighted_answer
 
@@ -223,13 +203,10 @@ examples=[
             'อยากทราบความถี่ในการดึงข้อมูลของ DXT360 บน Twitter',
             # 'ช่องทางติดตามข่าวสารของเรา',
                                 ]
-demo_before = gr.ChatInterface(fn=chat_interface_before, 
-                                examples=examples)
 
-demo_after = gr.ChatInterface(fn=chat_interface_after, 
+interface = gr.ChatInterface(fn=chat_interface, 
                               examples=examples)
 
-interface = gr.TabbedInterface([demo_before, demo_after], ["Before", "After"])
 
 if __name__ == "__main__":
     # Load your model, tokenizer, data, and index here...