button added

app.py

@@ -115,100 +115,99 @@ if submit_button:
 
 
 # React to user input
-if submit_button:
-    if prompt := st.chat_input(initial_input):
-        with st.spinner("Loading, please be patient with us ... 🙏"):
-            # Display user message in chat message container
-            st.chat_message("user").markdown(prompt)
-            # Add user message to chat history
-            st.session_state.messages.append({"role": "user", "content": prompt})
-
-            question = prompt
-            begin_t = time.time()
-            results = collection.query(query_texts=question, n_results=5)
-            end_t = time.time()
-            st.success(f"Query answser. | Time: {end_t - begin_t} sec")
-            idx = results["ids"][0]
-            idx = [int(i) for i in idx]
-            ref = pd.DataFrame(
-                {
-                    "idx": idx,
-                    "questions": [dataset["train"]["questions"][i] for i in idx],
-                    "answers": [dataset["train"]["answers"][i] for i in idx],
-                    "distances": results["distances"][0],
-                }
+if prompt := st.chat_input(initial_input):
+    with st.spinner("Loading, please be patient with us ... 🙏"):
+        # Display user message in chat message container
+        st.chat_message("user").markdown(prompt)
+        # Add user message to chat history
+        st.session_state.messages.append({"role": "user", "content": prompt})
+
+        question = prompt
+        begin_t = time.time()
+        results = collection.query(query_texts=question, n_results=5)
+        end_t = time.time()
+        st.success(f"Query answser. | Time: {end_t - begin_t} sec")
+        idx = results["ids"][0]
+        idx = [int(i) for i in idx]
+        ref = pd.DataFrame(
+            {
+                "idx": idx,
+                "questions": [dataset["train"]["questions"][i] for i in idx],
+                "answers": [dataset["train"]["answers"][i] for i in idx],
+                "distances": results["distances"][0],
+            }
+        )
+        # special_threshold = st.sidebar.slider('How old are you?', 0, 0.6, 0.1) # 0.3
+        filtered_ref = ref[ref["distances"] < special_threshold]
+        if filtered_ref.shape[0] > 0:
+            st.success("There are highly relevant information in our database.")
+            ref_from_db_search = filtered_ref["answers"].str.cat(sep=" ")
+            final_ref = filtered_ref
+        else:
+            st.warning(
+                "The database may not have relevant information to help your question so please be aware of hallucinations."
             )
-            # special_threshold = st.sidebar.slider('How old are you?', 0, 0.6, 0.1) # 0.3
-            filtered_ref = ref[ref["distances"] < special_threshold]
-            if filtered_ref.shape[0] > 0:
-                st.success("There are highly relevant information in our database.")
-                ref_from_db_search = filtered_ref["answers"].str.cat(sep=" ")
-                final_ref = filtered_ref
-            else:
-                st.warning(
-                    "The database may not have relevant information to help your question so please be aware of hallucinations."
-                )
-                ref_from_db_search = ref["answers"].str.cat(sep=" ")
-                final_ref = ref
-
-            if option == "YSA":
-                try:
-                    begin_t = time.time()
-                    llm_response = llama2_7b_ysa(question)
-                    end_t = time.time()
-                    st.success(f"Running LLM. | Time: {end_t - begin_t} sec")
-                except:
-                    st.warning("Sorry, the inference endpoint is temporarily down. 😔")
-                    llm_response = "NA."
-            else:
-                st.warning(
-                    "Apologies! We are in the progress of fine-tune the model, so it's currently unavailable. ⚙️"
-                )
-                llm_response = "NA"
-
-            finetuned_llm_guess = ["from_llm", question, llm_response, 0]
-            final_ref.loc[-1] = finetuned_llm_guess
-            final_ref = final_ref.reset_index()
-
-            # add ai judge as additional rating
-            if run_ai_judge == "Yes":
-                independent_ai_judge_score = []
-                begin_t = time.time()
-                for i in range(final_ref.shape[0]):
-                    this_content = final_ref["answers"][i]
-                    if len(this_content) > 3:
-                        arr1 = openai_text_embedding(question)
-                        arr2 = openai_text_embedding(this_content)
-                        # this_score = calculate_sts_openai_score(question, this_content)
-                        this_score = quantized_influence(arr1, arr2)
-                    else:
-                        this_score = 0
-                    independent_ai_judge_score.append(this_score)
-
-                final_ref["ai_judge"] = independent_ai_judge_score
+            ref_from_db_search = ref["answers"].str.cat(sep=" ")
+            final_ref = ref
 
+        if option == "YSA":
+            try:
+                begin_t = time.time()
+                llm_response = llama2_7b_ysa(question)
                 end_t = time.time()
-                st.success(f"Using AI Judge. | Time: {end_t - begin_t} sec")
-
-            engineered_prompt = f"""
-                Based on the context: {ref_from_db_search}
-
-                answer the user question: {question}
+                st.success(f"Running LLM. | Time: {end_t - begin_t} sec")
+            except:
+                st.warning("Sorry, the inference endpoint is temporarily down. 😔")
+                llm_response = "NA."
+        else:
+            st.warning(
+                "Apologies! We are in the progress of fine-tune the model, so it's currently unavailable. ⚙️"
+            )
+            llm_response = "NA"
 
-                Answer the question directly (don't say "based on the context, ...")
-            """
+        finetuned_llm_guess = ["from_llm", question, llm_response, 0]
+        final_ref.loc[-1] = finetuned_llm_guess
+        final_ref = final_ref.reset_index()
 
+        # add ai judge as additional rating
+        if run_ai_judge == "Yes":
+            independent_ai_judge_score = []
             begin_t = time.time()
-            answer = call_chatgpt(engineered_prompt)
+            for i in range(final_ref.shape[0]):
+                this_content = final_ref["answers"][i]
+                if len(this_content) > 3:
+                    arr1 = openai_text_embedding(question)
+                    arr2 = openai_text_embedding(this_content)
+                    # this_score = calculate_sts_openai_score(question, this_content)
+                    this_score = quantized_influence(arr1, arr2)
+                else:
+                    this_score = 0
+                independent_ai_judge_score.append(this_score)
+
+            final_ref["ai_judge"] = independent_ai_judge_score
+
             end_t = time.time()
-            st.success(f"Final API Call. | Time: {end_t - begin_t} sec")
-            response = answer
-
-        # Display assistant response in chat message container
-        with st.chat_message("assistant"):
-            with st.spinner("Wait for it..."):
-                st.markdown(response)
-                with st.expander("See reference:"):
-                    st.table(final_ref)
-        # Add assistant response to chat history
-        st.session_state.messages.append({"role": "assistant", "content": response})
+            st.success(f"Using AI Judge. | Time: {end_t - begin_t} sec")
+
+        engineered_prompt = f"""
+            Based on the context: {ref_from_db_search}
+
+            answer the user question: {question}
+
+            Answer the question directly (don't say "based on the context, ...")
+        """
+
+        begin_t = time.time()
+        answer = call_chatgpt(engineered_prompt)
+        end_t = time.time()
+        st.success(f"Final API Call. | Time: {end_t - begin_t} sec")
+        response = answer
+
+    # Display assistant response in chat message container
+    with st.chat_message("assistant"):
+        with st.spinner("Wait for it..."):
+            st.markdown(response)
+            with st.expander("See reference:"):
+                st.table(final_ref)
+    # Add assistant response to chat history
+    st.session_state.messages.append({"role": "assistant", "content": response})