24 / 08 / 07
1.0.0版本——终端回复、无UI、一次性全部输出、无Markdown渲染
from openai import OpenAI client = OpenAI( api_key="sk-azd0QGmMODK0lkbjHBY7", base_url="https://api.chatanywhere.com.cn/v1/" ) completion = client.chat.completions.create( model="gpt-4o-mini", messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "讲一个10个字的对联"}, ] ) print(completion)
1.0.1版本——有UI界面、一次性全部输出、无Markdown渲染
import tkinter as tk from tkinter import scrolledtext from openai import OpenAI import json # 定义API请求的函数 def get_response(): user_input = user_entry.get() if user_input.strip() == "": return user_entry.delete(0, tk.END) chat_text.configure(state='normal') chat_text.insert(tk.END, f"你：\n{user_input}\n") # 使用OpenAI库来发送请求 client = OpenAI(api_key="sk-azd0QkbjHBY7", base_url="https://api.chatanywhere.com.cn/v1/") response = client.chat.completions.create( model="gpt-4o-mini", messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": user_input} ] ) # 提取AI回复的内容 ai_reply = response.choices[0].message.content chat_text.insert(tk.END, f"GPT：\n{ai_reply}\n") chat_text.configure(state='disabled') chat_text.yview(tk.END) # 保存详细的返回数据以供debug使用 global debug_info debug_info = json.dumps(response.to_dict(), indent=4, ensure_ascii=False) # 显示详细的返回数据 def show_debug(): debug_window = tk.Toplevel(root) debug_window.title("Debug Information") debug_text = scrolledtext.ScrolledText(debug_window, wrap=tk.WORD, width=60, height=20) debug_text.pack(padx=10, pady=10, fill=tk.BOTH, expand=True) debug_text.insert(tk.END, debug_info) # 创建主窗口 root = tk.Tk() root.title("AI Chat") # 创建显示聊天内容的文本框 chat_text = scrolledtext.ScrolledText(root, wrap=tk.WORD, width=60, height=20) chat_text.pack(padx=10, pady=10, fill=tk.BOTH, expand=True) chat_text.configure(state='disabled') # 创建输入框和发送按钮 user_entry = tk.Entry(root, width=50) user_entry.pack(side=tk.LEFT, padx=10, pady=10, expand=True, fill=tk.X) send_button = tk.Button(root, text="发送", command=get_response) send_button.pack(side=tk.LEFT, pady=5) # 创建debug按钮 debug_button = tk.Button(root, text="Debug", command=show_debug) debug_button.pack(side=tk.LEFT, pady=5) # 运行主循环 root.mainloop()
1.0.2版本——有UI界面、流式输出、无Markdown渲染
import tkinter as tk from tkinter import scrolledtext, messagebox from openai import OpenAI import json class ChatApp: def __init__(self, root): self.root = root self.root.title("AI Chat") # 使窗口自适应大小 self.root.grid_columnconfigure(0, weight=1) self.root.grid_rowconfigure(0, weight=1) # 创建一个滚动文本框用于显示AI回复 self.text_area = scrolledtext.ScrolledText(self.root, wrap=tk.WORD, state='disabled') self.text_area.grid(column=0, row=0, padx=10, pady=10, sticky="nsew", columnspan=3) # 创建一个输入框 self.entry = tk.Entry(self.root) self.entry.grid(column=0, row=1, padx=10, pady=10, sticky="ew") # 创建一个发送按钮 self.send_button = tk.Button(self.root, text="发送", command=self.send_message) self.send_button.grid(column=1, row=1, padx=10, pady=10, sticky="ew") # 创建一个debug按钮 self.debug_button = tk.Button(self.root, text="debug", command=self.show_debug_info) self.debug_button.grid(column=2, row=1, padx=10, pady=10, sticky="ew") # 设置列的自适应宽度 self.root.grid_columnconfigure(0, weight=1) self.root.grid_columnconfigure(1, weight=0) self.root.grid_columnconfigure(2, weight=0) # 初始化OpenAI API self.client = OpenAI( api_key="sk-azd0jHBY7", base_url="https://api.chatanywhere.com.cn/v1/" ) self.debug_info = "" def send_message(self): user_message = self.entry.get() if user_message: self.text_area.config(state='normal') self.text_area.insert(tk.END, "你: " + "\n" + user_message + "\n") self.text_area.config(state='disabled') self.entry.delete(0, tk.END) self.get_ai_reply(user_message) def get_ai_reply(self, user_message): # 调用OpenAI API获取流式回复 response = self.client.chat.completions.create( model="gpt-4o-mini", messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": user_message}, ], stream=True # 启用流式输出 ) self.text_area.config(state='normal') self.text_area.insert(tk.END, "GPT: " + "\n") # 添加前缀 debug_info_list = [] for chunk in response: debug_info_list.append(chunk) content = chunk.choices[0].delta.content if content: # 检查内容是否为None self.text_area.insert(tk.END, content) self.text_area.update_idletasks() self.text_area.insert(tk.END, "\n") self.text_area.config(state='disabled') # 保存详细的返回数据以供debug使用 self.debug_info = json.dumps([chunk.to_dict() for chunk in debug_info_list], indent=4, ensure_ascii=False) def show_debug_info(self): # 这里显示API响应的详细信息 debug_window = tk.Toplevel(self.root) debug_window.title("Debug Information") debug_text = scrolledtext.ScrolledText(debug_window, wrap=tk.WORD, width=60, height=20) debug_text.pack(padx=10, pady=10, fill=tk.BOTH, expand=True) debug_text.insert(tk.END, self.debug_info) if __name__ == "__main__": root = tk.Tk() app = ChatApp(root) root.mainloop()
1.0.3——有UI界面、流式/非流式输出（可选择）、独立上下文对话、可清空对话记录、无Markdown渲染
import tkinter as tk from tkinter import scrolledtext, messagebox from openai import OpenAI import json class ChatApp: def __init__(self, root): self.root = root self.root.title("AI Chat") # 使窗口自适应大小 self.root.grid_columnconfigure(0, weight=1) self.root.grid_rowconfigure(0, weight=1) # 创建“New Chat”按钮 self.new_chat_button = tk.Button(self.root, text="New Chat", command=self.new_chat) self.new_chat_button.grid(column=0, row=0, padx=10, pady=5, sticky="w") # 创建一个滚动文本框用于显示AI回复 self.text_area = scrolledtext.ScrolledText(self.root, wrap=tk.WORD, state='disabled') self.text_area.grid(column=0, row=1, padx=10, pady=10, sticky="nsew", columnspan=3) # 创建一个输入框 self.entry = tk.Entry(self.root) self.entry.grid(column=0, row=4, padx=10, pady=10, sticky="ew") # 创建一个发送按钮 self.send_button = tk.Button(self.root, text="发送", command=self.send_message) self.send_button.grid(column=1, row=4, padx=10, pady=10, sticky="ew") # 创建一个debug按钮 self.debug_button = tk.Button(self.root, text="Debug", command=self.show_debug_info) self.debug_button.grid(column=2, row=4, padx=10, pady=10, sticky="ew") # 设置列的自适应宽度 self.root.grid_columnconfigure(0, weight=1) self.root.grid_columnconfigure(1, weight=0) self.root.grid_columnconfigure(2, weight=0) # 创建流式输出开关 self.stream_var = tk.BooleanVar(value=True) self.stream_switch = tk.Checkbutton(self.root, text="开启流式输出", variable=self.stream_var) self.stream_switch.grid(column=0, row=3, padx=10, pady=5, sticky="w", columnspan=3) # 初始化OpenAI API self.client = OpenAI( api_key="sk-azd0QGlkbjHBY7", base_url="https://api.chatanywhere.com.cn/v1/" ) self.debug_info = "" def new_chat(self): # 清空展示框的所有内容 self.text_area.config(state='normal') self.text_area.delete(1.0, tk.END) self.text_area.config(state='disabled') def send_message(self): user_message = self.entry.get() if user_message: self.text_area.config(state='normal') self.text_area.insert(tk.END, "你: " + "\n" + user_message + "\n") self.text_area.config(state='disabled') self.entry.delete(0, tk.END) if self.stream_var.get(): self.get_ai_reply_stream(user_message) else: self.get_ai_reply(user_message) def get_ai_reply_stream(self, user_message): # 调用OpenAI API获取流式回复 response = self.client.chat.completions.create( model="gpt-4o-mini", messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": user_message}, ], stream=True # 启用流式输出 ) self.text_area.config(state='normal') self.text_area.insert(tk.END, "GPT: " + "\n") # 添加前缀 debug_info_list = [] for chunk in response: debug_info_list.append(chunk) content = chunk.choices[0].delta.content if content: # 检查内容是否为None self.text_area.insert(tk.END, content) self.text_area.update_idletasks() self.text_area.insert(tk.END, "\n") self.text_area.config(state='disabled') # 保存详细的返回数据以供debug使用 self.debug_info = json.dumps([chunk.to_dict() for chunk in debug_info_list], indent=4, ensure_ascii=False) def get_ai_reply(self, user_message): # 调用OpenAI API获取普通回复 response = self.client.chat.completions.create( model="gpt-4o-mini", messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": user_message}, ] ) ai_reply = response.choices[0].message.content self.text_area.config(state='normal') self.text_area.insert(tk.END, "GPT: " + "\n" + ai_reply + "\n") self.text_area.config(state='disabled') # 保存详细的返回数据以供debug使用 self.debug_info = json.dumps(response.to_dict(), indent=4, ensure_ascii=False) def show_debug_info(self): # 这里显示API响应的详细信息 debug_window = tk.Toplevel(self.root) debug_window.title("Debug Information") debug_text = scrolledtext.ScrolledText(debug_window, wrap=tk.WORD, width=60, height=20) debug_text.pack(padx=10, pady=10, fill=tk.BOTH, expand=True) debug_text.insert(tk.END, self.debug_info) if __name__ == "__main__": root = tk.Tk() app = ChatApp(root) root.mainloop()
1.0.4——有UI界面、流式/非流式输出（可选择）、连续上下文对话、可清空对话记录、无Markdown渲染
import tkinter as tk from tkinter import scrolledtext, messagebox from openai import OpenAI import json class ChatApp: def __init__(self, root): self.root = root self.root.title("AI Chat") # 使窗口自适应大小 self.root.grid_columnconfigure(0, weight=1) self.root.grid_rowconfigure(0, weight=1) # 创建“New Chat”按钮 self.new_chat_button = tk.Button(self.root, text="New Chat", command=self.new_chat) self.new_chat_button.grid(column=0, row=0, padx=10, pady=5, sticky="w") # 创建一个滚动文本框用于显示AI回复 self.text_area = scrolledtext.ScrolledText(self.root, wrap=tk.WORD, state='disabled') self.text_area.grid(column=0, row=1, padx=10, pady=10, sticky="nsew", columnspan=3) # 创建一个输入框 self.entry = tk.Entry(self.root) self.entry.grid(column=0, row=4, padx=10, pady=10, sticky="ew") # 创建一个发送按钮 self.send_button = tk.Button(self.root, text="发送", command=self.send_message) self.send_button.grid(column=1, row=4, padx=10, pady=10, sticky="ew") # 创建一个debug按钮 self.debug_button = tk.Button(self.root, text="Debug", command=self.show_debug_info) self.debug_button.grid(column=2, row=4, padx=10, pady=10, sticky="ew") # 设置列的自适应宽度 self.root.grid_columnconfigure(0, weight=1) self.root.grid_columnconfigure(1, weight=0) self.root.grid_columnconfigure(2, weight=0) # 创建流式输出开关 self.stream_var = tk.BooleanVar(value=True) self.stream_switch = tk.Checkbutton(self.root, text="开启流式输出", variable=self.stream_var) self.stream_switch.grid(column=0, row=3, padx=10, pady=5, sticky="w", columnspan=3) # 初始化OpenAI API self.client = OpenAI( api_key="sk-azd0jHBY7", base_url="https://api.chatanywhere.com.cn/v1/" ) self.debug_info = "" self.messages = [{"role": "system", "content": "You are a helpful assistant."}] def new_chat(self): # 清空展示框的所有内容 self.text_area.config(state='normal') self.text_area.delete(1.0, tk.END) self.text_area.config(state='disabled') # 重置消息历史记录 self.messages = [{"role": "system", "content": "You are a helpful assistant."}] def send_message(self): user_message = self.entry.get() if user_message: self.messages.append({"role": "user", "content": user_message}) self.text_area.config(state='normal') self.text_area.insert(tk.END, "你: " + "\n" + user_message + "\n") self.text_area.config(state='disabled') self.entry.delete(0, tk.END) if self.stream_var.get(): self.get_ai_reply_stream() else: self.get_ai_reply() def get_ai_reply_stream(self): # 调用OpenAI API获取流式回复 response = self.client.chat.completions.create( model="gpt-4o-mini", messages=self.messages, stream=True # 启用流式输出 ) self.text_area.config(state='normal') self.text_area.insert(tk.END, "GPT: " + "\n") # 添加前缀 debug_info_list = [] ai_reply = "" for chunk in response: debug_info_list.append(chunk) content = chunk.choices[0].delta.content if content: # 检查内容是否为None self.text_area.insert(tk.END, content) ai_reply += content self.text_area.update_idletasks() self.text_area.insert(tk.END, "\n") self.text_area.config(state='disabled') # 保存详细的返回数据以供debug使用 self.debug_info = json.dumps([chunk.to_dict() for chunk in debug_info_list], indent=4, ensure_ascii=False) # 保存AI回复到消息历史记录 self.messages.append({"role": "assistant", "content": ai_reply}) def get_ai_reply(self): # 调用OpenAI API获取普通回复 response = self.client.chat.completions.create( model="gpt-4o-mini", messages=self.messages ) ai_reply = response.choices[0].message.content self.text_area.config(state='normal') self.text_area.insert(tk.END, "GPT: " + "\n" + ai_reply + "\n") self.text_area.config(state='disabled') # 保存详细的返回数据以供debug使用 self.debug_info = json.dumps(response.to_dict(), indent=4, ensure_ascii=False) # 保存AI回复到消息历史记录 self.messages.append({"role": "assistant", "content": ai_reply}) def show_debug_info(self): # 这里显示API响应的详细信息 debug_window = tk.Toplevel(self.root) debug_window.title("Debug Information") debug_text = scrolledtext.ScrolledText(debug_window, wrap=tk.WORD, width=60, height=20) debug_text.pack(padx=10, pady=10, fill=tk.BOTH, expand=True) debug_text.insert(tk.END, self.debug_info) if __name__ == "__main__": root = tk.Tk() app = ChatApp(root) root.mainloop()
1.0.5——有UI界面、流式/非流式输出（可选择）、独立/连续上下文对话（可选择）、可清空对话记录、无Markdown渲染
import tkinter as tk from tkinter import scrolledtext, messagebox from openai import OpenAI import json class ChatApp: def __init__(self, root): self.root = root self.root.title("AI Chat") # 使窗口自适应大小 self.root.grid_columnconfigure(0, weight=1) self.root.grid_rowconfigure(0, weight=1) # 创建“New Chat”按钮 self.new_chat_button = tk.Button(self.root, text="New Chat", command=self.new_chat) self.new_chat_button.grid(column=0, row=0, padx=10, pady=5, sticky="w") # 创建一个滚动文本框用于显示AI回复 self.text_area = scrolledtext.ScrolledText(self.root, wrap=tk.WORD, state='disabled') self.text_area.grid(column=0, row=1, padx=10, pady=10, sticky="nsew", columnspan=4) # 创建一个输入框 self.entry = tk.Entry(self.root) self.entry.grid(column=0, row=4, padx=10, pady=10, sticky="ew", columnspan=4) # 创建一个发送按钮 self.send_button = tk.Button(self.root, text="发送", command=self.send_message) self.send_button.grid(column=0, row=5, padx=10, pady=10, sticky="ew") # 创建一个debug按钮 self.debug_button = tk.Button(self.root, text="Debug", command=self.show_debug_info) self.debug_button.grid(column=1, row=5, padx=10, pady=10, sticky="ew") # 设置列的自适应宽度 self.root.grid_columnconfigure(0, weight=1) self.root.grid_columnconfigure(1, weight=0) self.root.grid_columnconfigure(2, weight=0) self.root.grid_columnconfigure(3, weight=0) # 创建流式输出开关 self.stream_var = tk.BooleanVar(value=True) self.stream_switch = tk.Checkbutton(self.root, text="开启流式输出", variable=self.stream_var) self.stream_switch.grid(column=0, row=3, padx=10, pady=5, sticky="w", columnspan=1) # 创建上下文模式选择下拉菜单 self.context_mode_var = tk.StringVar(value="上下文连续对话") self.context_mode_menu = tk.OptionMenu(self.root, self.context_mode_var, "独立对话", "上下文连续对话") self.context_mode_menu.grid(column=1, row=3, padx=10, pady=5, sticky="w", columnspan=3) # 初始化OpenAI API self.client = OpenAI( api_key="sk-azd0HBY7", base_url="https://api.chatanywhere.com.cn/v1/" ) self.debug_info = "" self.messages = [{"role": "system", "content": "You are a helpful assistant."}] def new_chat(self): # 清空展示框的所有内容 self.text_area.config(state='normal') self.text_area.delete(1.0, tk.END) self.text_area.config(state='disabled') # 重置消息历史记录 self.messages = [{"role": "system", "content": "You are a helpful assistant."}] def send_message(self): user_message = self.entry.get() if user_message: self.text_area.config(state='normal') self.text_area.insert(tk.END, "你: " + "\n" + user_message + "\n") self.text_area.config(state='disabled') self.entry.delete(0, tk.END) if self.context_mode_var.get() == "独立对话": self.messages = [{"role": "system", "content": "You are a helpful assistant."}] self.messages.append({"role": "user", "content": user_message}) if self.stream_var.get(): self.get_ai_reply_stream() else: self.get_ai_reply() def get_ai_reply_stream(self): # 调用OpenAI API获取流式回复 response = self.client.chat.completions.create( model="gpt-4o-mini", messages=self.messages, stream=True # 启用流式输出 ) self.text_area.config(state='normal') self.text_area.insert(tk.END, "GPT: " + "\n") # 添加前缀 debug_info_list = [] ai_reply = "" for chunk in response: debug_info_list.append(chunk) content = chunk.choices[0].delta.content if content: # 检查内容是否为None self.text_area.insert(tk.END, content) ai_reply += content self.text_area.update_idletasks() self.text_area.insert(tk.END, "\n") self.text_area.config(state='disabled') # 保存详细的返回数据以供debug使用 self.debug_info = json.dumps([chunk.to_dict() for chunk in debug_info_list], indent=4, ensure_ascii=False) # 保存AI回复到消息历史记录 self.messages.append({"role": "assistant", "content": ai_reply}) def get_ai_reply(self): # 调用OpenAI API获取普通回复 response = self.client.chat.completions.create( model="gpt-4o-mini", messages=self.messages ) ai_reply = response.choices[0].message.content self.text_area.config(state='normal') self.text_area.insert(tk.END, "GPT: " + "\n" + ai_reply + "\n") self.text_area.config(state='disabled') # 保存详细的返回数据以供debug使用 self.debug_info = json.dumps(response.to_dict(), indent=4, ensure_ascii=False) # 保存AI回复到消息历史记录 self.messages.append({"role": "assistant", "content": ai_reply}) def show_debug_info(self): # 这里显示API响应的详细信息 debug_window = tk.Toplevel(self.root) debug_window.title("Debug Information") debug_text = scrolledtext.ScrolledText(debug_window, wrap=tk.WORD, width=60, height=20) debug_text.pack(padx=10, pady=10, fill=tk.BOTH, expand=True) debug_text.insert(tk.END, self.debug_info) if __name__ == "__main__": root = tk.Tk() app = ChatApp(root) root.mainloop()
1.0.6——有UI界面、流式/非流式输出（可选择）、独立/连续上下文对话（可选择）、可清空对话记录、窗口效果显示更好、无Markdown渲染
import tkinter as tk from tkinter import scrolledtext from PIL import Image, ImageTk from openai import OpenAI import json import threading class ChatApp: def __init__(self, root): self.root = root self.root.title("Chat Application") self.root.geometry("800x600") # 配置根窗口的网格权重 self.root.grid_columnconfigure(0, weight=1) self.root.grid_rowconfigure(0, weight=1) # 创建并配置消息框架 self.message_frame = tk.Frame(self.root) self.message_frame.grid(column=0, row=0, padx=10, pady=10, sticky="nsew") self.message_frame.grid_columnconfigure(0, weight=1) self.message_frame.grid_rowconfigure(0, weight=1) # 创建并配置画布和滚动条 self.canvas = tk.Canvas(self.message_frame) self.scrollbar = tk.Scrollbar(self.message_frame, orient="vertical", command=self.canvas.yview) self.scrollable_frame = tk.Frame(self.canvas) self.scrollable_frame.bind( "<Configure>", lambda e: self.canvas.configure( scrollregion=self.canvas.bbox("all") ) ) self.canvas.create_window((0, 0), window=self.scrollable_frame, anchor="nw") self.canvas.configure(yscrollcommand=self.scrollbar.set) self.canvas.pack(side="left", fill="both", expand=True) self.scrollbar.pack(side="right", fill="y") # 绑定鼠标滚轮事件 self.canvas.bind_all("<MouseWheel>", self._on_mousewheel) # 输入框和发送按钮框架 input_frame = tk.Frame(self.root) input_frame.grid(column=0, row=1, padx=10, pady=5, sticky="ew") input_frame.grid_columnconfigure(0, weight=1) # 输入框 self.entry = tk.Entry(input_frame, width=50) self.entry.grid(column=0, row=0, padx=(0, 5), pady=5, sticky="ew") # 发送按钮 self.send_button = tk.Button(input_frame, text="发送", command=self.send_message) self.send_button.grid(column=1, row=0, padx=5, pady=5, sticky="e") # 控制按钮框架 control_frame = tk.Frame(self.root) control_frame.grid(column=0, row=2, padx=10, pady=5, sticky="ew") # 新对话按钮 self.new_chat_button = tk.Button(control_frame, text="新对话", command=self.new_chat) self.new_chat_button.grid(column=0, row=0, padx=(0, 5), pady=5, sticky="w") # Debug按钮 self.debug_button = tk.Button(control_frame, text="Debug", command=self.show_debug_info) self.debug_button.grid(column=1, row=0, padx=5, pady=5, sticky="w") # 流式输出复选框 self.stream_var = tk.BooleanVar(value=True) self.stream_checkbox = tk.Checkbutton(control_frame, text="流式输出", variable=self.stream_var) self.stream_checkbox.grid(column=2, row=0, padx=5, pady=5, sticky="w") # 模式选择下拉菜单 self.context_mode_var = tk.StringVar(value="上下文连续对话") self.context_mode_menu = tk.OptionMenu(control_frame, self.context_mode_var, "独立对话", "上下文连续对话") self.context_mode_menu.grid(column=3, row=0, padx=5, pady=5, sticky="w") # 加载用户和助手头像 self.user_img = self.load_and_resize_image("images/user_icon.png", (30, 30)) self.assistant_img = self.load_and_resize_image("images/chatgpt_icon.png", (30, 30)) # 初始化OpenAI API self.client = OpenAI( api_key="sk-azd0HBY7", base_url="https://api.chatanywhere.com.cn/v1/" ) self.debug_info = "" self.messages = [{"role": "system", "content": "You are a helpful assistant."}] def load_and_resize_image(self, path, size): img = Image.open(path) img = img.resize(size, Image.LANCZOS) return ImageTk.PhotoImage(img) def new_chat(self): for widget in self.scrollable_frame.winfo_children(): widget.destroy() self.messages = [{"role": "system", "content": "You are a helpful assistant."}] # 重新配置画布的滚动区域为初始状态 self.canvas.configure(scrollregion=(0, 0, 0, 0)) def send_message(self): user_message = self.entry.get() if user_message: self.add_message_to_frame("user", user_message) self.entry.delete(0, tk.END) if self.context_mode_var.get() == "独立对话": self.messages = [{"role": "system", "content": "You are a helpful assistant."}] self.messages.append({"role": "user", "content": user_message}) if self.stream_var.get(): threading.Thread(target=self.get_ai_reply_stream, daemon=True).start() else: threading.Thread(target=self.get_ai_reply, daemon=True).start() def add_message_to_frame(self, role, message): frame = tk.Frame(self.scrollable_frame) frame.pack(fill=tk.X, padx=5, pady=5) frame.grid_columnconfigure(1, weight=1) if role == "user": img_label = tk.Label(frame, image=self.user_img) else: img_label = tk.Label(frame, image=self.assistant_img) img_label.grid(row=0, column=0, sticky="nw", padx=(0, 5)) text_widget = tk.Label(frame, text=message, wraplength=600, justify="left", anchor="w") text_widget.grid(row=0, column=1, sticky="nsew") edit_button = tk.Label(frame, text="编辑", fg="gray", cursor="hand2") edit_button.grid(row=0, column=2, sticky="ne", padx=(5, 0)) edit_button.bind("<Button-1>", lambda e: self.toggle_edit_mode(frame, text_widget, edit_button, role, message)) # 修改为传递原始message # 更新画布滚动区域并调整滚动位置 self.adjust_scroll_region() def toggle_edit_mode(self, frame, widget, edit_button, role, original_message): # 增加original_message参数 if edit_button["text"] == "编辑": # 切换到编辑模式 text_widget = tk.Text(frame, wrap=tk.WORD) text_widget.insert(tk.END, widget["text"]) text_widget.grid(row=0, column=1, sticky="nsew") widget.grid_remove() edit_button.configure(text="完成") # 保存对 text_widget 的引用 frame.text_widget = text_widget # 自动调整Text高度以适应内容 max_height = 20 # 设置最大行数为20行，可以根据需要调整 num_lines = int(text_widget.index('end-1c').split('.')[0]) # num_lines = text_widget.count("1.0", "end-1c", "displaylines")[0] text_widget.configure(height=num_lines) else: # 保存编辑并切换回显示模式 new_text = frame.text_widget.get("1.0", tk.END).strip() widget.configure(text=new_text) frame.text_widget.grid_remove() widget.grid() edit_button.configure(text="编辑") # 删除 text_widget 的引用 del frame.text_widget # 更新消息历史，替换原始消息内容 for msg in self.messages: if msg["role"] == role and msg["content"] == original_message: msg["content"] = new_text break self.adjust_scroll_region() def adjust_scroll_region(self): self.scrollable_frame.update_idletasks() bbox = self.canvas.bbox("all") # 获取画布的高度 canvas_height = self.canvas.winfo_height() # 如果内容高度小于画布高度，将内容固定在顶部 if (bbox[3] - bbox[1]) < canvas_height: self.canvas.configure(scrollregion=(bbox[0], bbox[1], bbox[2], canvas_height)) self.canvas.yview_moveto(0) else: self.canvas.configure(scrollregion=bbox) self.canvas.yview_moveto(1) def get_ai_reply_stream(self): response = self.client.chat.completions.create( model="gpt-4o-mini", messages=self.messages, stream=True ) debug_info_list = [] ai_reply = "" temp_frame = tk.Frame(self.scrollable_frame) temp_frame.pack(fill=tk.X, padx=5, pady=5) temp_frame.grid_columnconfigure(1, weight=1) img_label = tk.Label(temp_frame, image=self.assistant_img) img_label.grid(row=0, column=0, sticky="nw", padx=(0, 5)) text_widget = tk.Label(temp_frame, wraplength=600, justify="left", anchor="w") text_widget.grid(row=0, column=1, sticky="nsew") edit_button = tk.Label(temp_frame, text="编辑", fg="gray", cursor="hand2") edit_button.grid(row=0, column=2, sticky="ne", padx=(5, 0)) for chunk in response: debug_info_list.append(chunk) content = chunk.choices[0].delta.content if content: ai_reply += content text_widget.config(text=ai_reply) self.adjust_scroll_region() self.root.update() edit_button.bind("<Button-1>", lambda e: self.toggle_edit_mode(temp_frame, text_widget, edit_button, "assistant", ai_reply)) self.debug_info = json.dumps([chunk.to_dict() for chunk in debug_info_list], indent=4, ensure_ascii=False) self.messages.append({"role": "assistant", "content": ai_reply}) def get_ai_reply(self): response = self.client.chat.completions.create( model="gpt-4o-mini", messages=self.messages ) ai_reply = response.choices[0].message.content self.root.after(0, self.add_message_to_frame, "assistant", ai_reply) self.debug_info = json.dumps(response.to_dict(), indent=4, ensure_ascii=False) self.messages.append({"role": "assistant", "content": ai_reply}) def show_debug_info(self): debug_window = tk.Toplevel(self.root) debug_window.title("Debug Information") debug_text = scrolledtext.ScrolledText(debug_window, wrap=tk.WORD, width=60, height=20) debug_text.pack(padx=10, pady=10, fill=tk.BOTH, expand=True) debug_text.insert(tk.END, self.debug_info) def _on_mousewheel(self, event): # 获取当前的滚动位置 current_position = self.canvas.yview()[0] # 如果当前位置已经在顶部，并且尝试向上滚动，则不执行滚动 if current_position <= 0 and event.delta > 0: return self.canvas.yview_scroll(int(-1 * (event.delta / 120)), "units") if __name__ == "__main__": root = tk.Tk() app = ChatApp(root) root.mainloop()