web-scraping - 获取所有 ChatGPT session 的提示和对话列表

标签 web-scraping openai-api

我有数百个 ChatGPT 对话,我想将它们全部提取到一个 JSON 结构中。

  • 对话名称

  • 对话的提示和响应列表。

我在 API Reference 中找不到任何内容所以想知道我刚刚错过了或者是否有一种方法可以使用抓取来做到这一点。

{
  "title": "Escape smoke, move stuff.",
  "prompts": [
    {
      "prompt": "Looking for suggestions! Due to the smoke season, I need to leave town for a few weeks but still want to maintain my current lifestyle. I want to bring my large monitor, chair, and other items with me. How far do I need to travel and which province(s) do you recommend for a good getaway? Also, any recommendations for transportation options like Grab, car, bus, or train?",
      "response": "I would recommend traveling to a province with cleaner air, such as British Columbia or Alberta. As for transportation, it really depends on your preference and budget. If you prefer comfort and convenience, renting a car might be the best option. If you're on a budget, taking a bus or train could be more affordable. However, if you want to save time, taking a Grab or driving your own car would be the most efficient. Just make sure to pack your items securely for the trip!"
    },
    {
      "prompt": "Imagine you are in the midst of a journey to escape the smoke season in your town. You have brought along your large monitor, chair, and other items to maintain your current lifestyle. As you travel, you wonder how far you need to go and which province(s) would be the best for a relaxing getaway. Consider different transportation options like Grab, car, bus, or train. Write about your experience and share your thoughts on the ideal destination for your escape.",
      "response": "As I left my town behind, I felt a sense of relief knowing that I would soon be breathing in cleaner air. I had packed my large monitor, chair, and other essentials to ensure that my escape would be as comfortable as possible. As I journeyed further, I wondered how far I needed to go to find a relaxing destination. I had heard that British Columbia and Alberta were good options, so I decided to head in that direction. Along the way, I considered different transportation options. While renting a car seemed like the most convenient option, I opted for a bus to save on costs. Finally, after a few hours of traveling, I arrived at my destination feeling rejuvenated and ready to enjoy my temporary escape from the smoke."
    }
  ]
}

最佳答案

OpenAI 条款和条件禁止网页抓取。相关:

Is it against ToS to scrape a conversation with ChatGPT?

来自 comment

Have a look at OpenAI’s Terms of Use. #2, section c, part iv:

You may not … use any method to extract data from the Services, including web scraping, web harvesting, or web data extraction methods, other than as permitted through the API;

我的解释是,除非您有权访问 ChatGPT API ,否则抓取您的 ChatGPT 对话将违反服务条款。 .

无论如何,在不影响 OpenAI 基础结构的情况下,使用 Chrome 开发者工具进行 ChatGPT 对话可能并不难

  1. 转到 https://chat.openai.com/chat并登录
  2. 打开 Chrome 开发者工具
  3. 打开“应用程序”标签
  4. 扩展框架 > XHR 和获取

在这里您将看到一个文件列表。从这里您可以获得链接甚至复制文件内容。

据我所知,聊天完成端点(作为 ChatGPT API 销售)不提供对 ChatGPT (https://chat.openai.com/chat) 用户数据的访问。它只是允许访问为 ChatGPT 网络应用提供支持的大型语言模型。

引用

关于web-scraping - 获取所有 ChatGPT session 的提示和对话列表,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/75664531/

相关文章:

python - 使用 Python 进行网页抓取

javascript - 使用 R 从 Javascript 检索文本(html 节点)

python - 无法获取json中存储的某些项目的对应值

android - 如何在没有人窃取 token 的情况下使用我的移动应用程序中的 API

javascript - 通过 VBA 从 IE 页面的下拉列表中选择一个值

python - 迭代多个response.get API的项目列表

node.js - OpenAI 嵌入 API : How to change the embedding output dimension?

lua - 当 key 正确时,OpenAI API 返回 "unauthorized"错误

openai-api - 如何解决 "Models maximum content length exceeded"错误

node.js - OPENAI API 完成不返回文本