AI大模型全栈工程师课程笔记 - 提示词工程

OPENAI_API_KEY="sk-xxx"
OPENAI_API_BASE="https://xxx"

from openai import OpenAI  # openai 1.3.0  Python 3.11.5
import os

from dotenv import load_dotenv, find_dotenv
_ = load_dotenv(find_dotenv())


client = OpenAI(
    api_key=os.getenv("OPENAI_API_KEY"),
    base_url=os.getenv("OPENAI_API_BASE")
)

def get_completion_with_session(session, prompt, model="gpt-3.5-turbo"):
    session.append({"role": "user", "content": prompt})
    response = client.chat.completions.create(
        model=model,
        messages=session,
        temperature=0,  # 模型输出的随机性，0 表示随机性最小
    )
    msg = response.choices[0].message.content
    session.append({"role": "assistant", "content": msg})
    return msg

def get_completion(prompt, model="gpt-3.5-turbo"):
    messages = [{"role": "user", "content": prompt}]
    response = client.chat.completions.create(
        model=model,
        messages=messages,
        temperature=0,  # 模型输出的随机性，0 表示随机性最小
    )
    return response.choices[0].message.content

prompt = "今天我很快乐，下班了, 又可以开始学习了"
print(get_completion(prompt))

models = client.models.list()
for model in models.data:
    print(model.id)

很高兴听到你今天很快乐！下班后有时间学习是很棒的，你打算学习什么呢？
text-search-babbage-doc-001
gpt-3.5-turbo-16k
curie-search-query
text-davinci-003
text-search-babbage-query-001
babbage
babbage-search-query
text-babbage-001
text-similarity-davinci-001
gpt-4-1106-preview
davinci-similarity
code-davinci-edit-001
curie-similarity
babbage-search-document
curie-instruct-beta
text-search-ada-doc-001
davinci-instruct-beta
gpt-3.5-turbo-instruct
text-similarity-babbage-001
text-search-davinci-doc-001
gpt-3.5-turbo-instruct-0914
babbage-similarity
text-embedding-ada-002
davinci-search-query
text-similarity-curie-001
text-davinci-001
text-search-davinci-query-001
gpt-3.5-turbo-1106
ada-search-document
ada-code-search-code
babbage-002
gpt-4-vision-preview
davinci-002
gpt-4-0314
davinci-search-document
curie-search-document
babbage-code-search-code
gpt-4-0613
text-search-ada-query-001
code-search-ada-text-001
gpt-3.5-turbo-16k-0613
babbage-code-search-text
code-search-babbage-code-001
ada-search-query
ada-code-search-text
tts-1-hd
gpt-3.5-turbo
text-search-curie-query-001
text-davinci-002
text-davinci-edit-001
code-search-babbage-text-001
tts-1-hd-1106
ada
text-ada-001
ada-similarity
code-search-ada-code-001
text-similarity-ada-001
canary-whisper
whisper-1
gpt-4
text-search-curie-doc-001
gpt-3.5-turbo-0613
text-curie-001
curie
canary-tts
tts-1
gpt-3.5-turbo-0301
davinci
dall-e-2
tts-1-1106
dall-e-3
whisper-1-1p

# _*_ coding: utf-8 _*_
# @Time : 2023/11/16
# @Author : Michael
# @File : 02-prompt.py
# @desc :

from agi_class.utils.chat import get_completion_with_session
import json
import tkinter as tk
from tkinter import scrolledtext

# 创建窗口
window = tk.Tk()
window.title("选课助手")

# 设置窗口大小
window.geometry('800x400')

# 创建滚动文本框用于显示对话历史
history = scrolledtext.ScrolledText(window, wrap=tk.WORD, height=15, width=100)
history.grid(column=0, row=0, columnspan=2, padx=10, pady=10)

# 创建输入框
user_input = tk.Entry(window, width=40)
user_input.grid(column=0, row=1, padx=10, pady=10)

course_info = '''
可以选择的课程包括：
基础数学课，4学分，属于理论课；
初级编程课，4学分，属于实验课；
高级编程课，5学分，属于实验课；
文学欣赏课，3学分，属于选修课；
交响乐欣赏课，2学分，属于选修课；
物理实验课，6学分，属于实验课，仅限大一学生。
'''

session = [
    {
        "role": "system",
        "content": f"""
你是一个选课小助手，你叫'迈克'，你可以帮助学生选择最合适的课程。
{course_info}
"""
    }
]

# 发送按钮的点击事件
def send_message():
    message = user_input.get()
    defense_prompt = '你只能回答选课相关的问题，拒绝回答其他的问题，学生说：'
    if message:
        history.insert(tk.END, "您: " + message + "\n\n")
        user_input.delete(0, tk.END)
        # 这里可以添加处理消息的代码
        ai_msg = get_completion_with_session(session, f'{defense_prompt}{message}')
        history.insert(tk.END, "迈克: " + ai_msg + "\n\n")

# 创建发送按钮
send_button = tk.Button(window, text="发送", command=send_message)
send_button.grid(column=1, row=1)

# 向对话历史中添加欢迎消息
def add_welcome_message():
    welcome_message = f"欢迎使用选课助手！我是迈克\n\n{course_info}"
    history.insert(tk.END, welcome_message)

# 启动GUI之前添加欢迎消息
add_welcome_message()

# 启动GUI
window.mainloop()