modified code, add another file for gradio non-stream

testGenius.py

@@ -8,48 +8,54 @@ import gradio as gr
 dashscope.api_key = 'sk-83b8ed0ead0849ae9e63a2ae5bdbde0d'  # Rayman's API key
 
 
-def respond(prompt, chat_history, instruction, model, if_stream):
+def respond_nonStream(prompt, chat_history, instruction, model):
-    """
-    与AI助手进行对话，并返回对话历史。
-
-    参数:
-    - prompt: 用户输入的文本。
-    - chat_history: 之前的聊天历史，列表，每个元素是二元组，包括用户输入和AI响应。
-    - instruction: 系统指令，作为对话的起始信息。
-
-    返回值:
-    - 生成器，每次产生一个二元组，包括空字符串和更新后的聊天历史。
-    """
     # 构建对话消息结构
+    messages = [{'role': 'system', 'content': instruction},
+                {'role': 'user', 'content': prompt}]
+    print(f"Messages: {messages}")
+
+    # 初始化空字符串以聚合响应
+    full_response = ""
+
+    # 调用AI模型生成响应
+    try:
+        response = Generation.call(model=model,
+                                   messages=messages,
+                                   seed=1234,
+                                   result_format='message',
+                                   stream=False,
+                                   incremental_output=False)
+        print(f"Response: {response}")
+
+        if not chat_history or chat_history[-1][0] != prompt:
+            chat_history.append([prompt, ""])
+        print(f"old chat history: {chat_history}")
+        if response.status_code == HTTPStatus.OK:
+            # 获取响应中的消息内容
+            message = response.output.choices[0]['message']['content']
+            print(f"Generated message: {message}")
+
+            # 更新聊天历史记录中的最后一条记录
+            chat_history[-1] = [prompt, message]
+            print(f"Updated chat_history: {chat_history}")
+            return "", chat_history
+        else:
+            print(f"Error: Received response status {response.status_code}")
+            return "Error: Could not generate response", chat_history
+
+    except Exception as e:
+        print(f"Exception occurred: {e}")
+        return f"Exception occurred: {e}", chat_history
+
+
+def respond(prompt, chat_history, instruction, model, if_stream='Stream'):
+    if if_stream == 'Stream':
         messages = [{'role': 'system',
                      'content': instruction},
                     {'role': 'user',
                      'content': prompt}
                     ]
         full_response = ""  # 初始化空字符串以聚合响应
-    # 调用AI模型生成响应
-    # -------非流式输出-------
-    if if_stream == 'Non-Stream':
-        response = Generation.call(model=model,
-                                   messages=messages,
-                                   # 设置随机数种子seed，如果没有设置，则随机数种子默认为1234
-                                   seed=1234,
-                                   # 将输出设置为"message"格式
-                                   result_format='message',
-                                   # 设置输出方式为非流式输出
-                                   stream=False,
-                                   # 设置输出方式为非增量式输出
-                                   incremental_output=False)
-        if not chat_history or chat_history[-1][0] != prompt:
-            chat_history.append((prompt, ""))
-        if response.status_code == HTTPStatus.OK:
-            # 获取响应中的消息内容
-            message = response.output.choices[0]['message']['content']
-            # 将消息内容添加到聊天历史中
-            chat_history.append((prompt, message))
-            # 返回更新后的聊天历史
-            return "", chat_history
-    elif if_stream == 'Stream':
         # -------流式输出-------
         responses = Generation.call(model=model,
                                     messages=messages,
@@ -88,31 +94,45 @@ def respond(prompt, chat_history, instruction, model, if_stream):
                 yield "", chat_history
                 break  # 出现错误时终止循环
 
-
+    elif if_stream == 'Non-Stream':
-def respond_nonStream(prompt, chat_history, instruction, model):
         # 构建对话消息结构
-    messages = [{'role': 'system',
+        messages = [{'role': 'system', 'content': instruction},
-                 'content': instruction},
+                    {'role': 'user', 'content': prompt}]
-                {'role': 'user',
+        print(f"Messages: {messages}")
-                 'content': prompt}
+
-                ]
-    full_response = ""  # 初始化空字符串以聚合响应
         # 调用AI模型生成响应
-    responses = Generation.call(model=model,
+        try:
+            response = Generation.call(model=model,
                                        messages=messages,
-                                # 设置随机数种子seed，如果没有设置，则随机数种子默认为1234
                                        seed=1234,
-                                # 将输出设置为"message"格式
                                        result_format='message',
-                                stream=True,  # 设置输出方式为流式输出
+                                       stream=False,
-                                incremental_output=True,  # 增量式流式输出
+                                       incremental_output=False)
-                                temperature=1.8,
+            print(f"Response: {response}")
-                                top_p=0.9,
+
-                                top_k=999)
+            if not chat_history or chat_history[-1][0] != prompt:
+                chat_history.append([prompt, ""])
+            print(f"old chat history: {chat_history}")
+            if response.status_code == HTTPStatus.OK:
+                # 获取响应中的消息内容
+                message = response.output.choices[0]['message']['content']
+                print(f"Generated message: {message}")
+
+                # 更新聊天历史记录中的最后一条记录
+                chat_history[-1] = [prompt, message]
+                print(f"Updated chat_history: {chat_history}")
+                return "", chat_history
+            else:
+                print(f"Error: Received response status {response.status_code}")
+                return "Error: Could not generate response", chat_history
+
+        except Exception as e:
+            print(f"Exception occurred: {e}")
+            return f"Exception occurred: {e}", chat_history
 
 
 
-llm_model_list = ['qwen-turbo','qwen-plus', 'qwen-max']
+llm_model_list = ['qwen-turbo', 'qwen-plus', 'qwen-max']
 init_llm = llm_model_list[0]
 
 # 创建 Gradio 界面
@@ -122,17 +142,20 @@ with gr.Blocks() as demo:
         # AI TestGenius
         A simple LLM app for generating test cases from function design.
         """)
-    chatbot = gr.Chatbot()
+    history = [["Hello", "Hello, how can I help you?"]]
+    chatbot = gr.Chatbot(history)
     msg = gr.Textbox(label="Prompt")
     with gr.Accordion(label="Advanced options", open=False):
         system = gr.Textbox(label="System prompts", lines=2,
                             value="A conversation between a user and an LLM-based AI assistant. The assistant gives helpful and honest answers.")
+        # 选择模型
         llm = gr.Dropdown(
             llm_model_list,
             label='Choose LLM Model',
             value=init_llm,
             interactive=True
         )
+        # 选择是否流式输出
         if_stream = gr.Dropdown(
             ["Stream", "Non-Stream"],
             label='Choose Streaming',
@@ -145,7 +168,6 @@ with gr.Blocks() as demo:
     btn.click(respond, inputs=[msg, chatbot, system, llm, if_stream], outputs=[msg, chatbot])  # click to submit
     msg.submit(respond, inputs=[msg, chatbot, system, llm, if_stream], outputs=[msg, chatbot])  # Press enter to submit
 
-
 # 运行界面
 if __name__ == "__main__":
     gr.close_all()

testGradio.py

@@ -0,0 +1,59 @@
+import gradio as gr
+import random
+import time
+from http import HTTPStatus
+import dashscope
+from dashscope import Generation
+
+dashscope.api_key = 'sk-83b8ed0ead0849ae9e63a2ae5bdbde0d'  # Rayman's API key
+
+with gr.Blocks() as demo:
+    history = [["Hello","Hello, how can I help you?"]]
+    chatbot = gr.Chatbot(history)
+    msg = gr.Textbox()
+    clear = gr.ClearButton([msg, chatbot])
+
+
+    def respond_nonStream(prompt, chat_history):
+        # 构建对话消息结构
+        messages = [{'role': 'system', 'content': "A conversation between a user and an LLM-based AI assistant. The assistant gives helpful and honest answers."},
+                    {'role': 'user', 'content': prompt}]
+        print(f"Messages: {messages}")
+
+        # 初始化空字符串以聚合响应
+        full_response = ""
+
+        # 调用AI模型生成响应
+        try:
+            response = Generation.call(model='qwen-turbo',
+                                       messages=messages,
+                                       seed=1234,
+                                       result_format='message',
+                                       stream=False,
+                                       incremental_output=False)
+            print(f"Response: {response}")
+
+            if not chat_history or chat_history[-1][0] != prompt:
+                chat_history.append([prompt, ""])
+            print(f"old chat history: {chat_history}")
+            if response.status_code == HTTPStatus.OK:
+                # 获取响应中的消息内容
+                message = response.output.choices[0]['message']['content']
+                print(f"Generated message: {message}")
+
+                # 更新聊天历史记录中的最后一条记录
+                chat_history[-1] = [prompt, message]
+                print(f"Updated chat_history: {chat_history}")
+                return "", chat_history
+            else:
+                print(f"Error: Received response status {response.status_code}")
+                return "Error: Could not generate response", chat_history
+
+        except Exception as e:
+            print(f"Exception occurred: {e}")
+            return f"Exception occurred: {e}", chat_history
+
+    msg.submit(respond_nonStream, inputs=[msg, chatbot], outputs=[msg, chatbot])
+
+if __name__ == "__main__":
+    demo.launch()