feat: add experiment smart group base text embedding

[lawvs]

This PR is only for verification purposes, please do not merge it.

I made an attempt to group tabs based on the text embedded from the link preview, but the insufficient information provided in the preview led to unsatisfactory results.

Here are some example data:

{
  "url": "https://github.com/",
  "type": "Books",  // It should be grouped into the Productivity category instead of Game or Books
  "similarity": 0.7600699589640771,
  "preview": {
    "description": "GitHub is where over 100 million developers shape the future of software, together. Contribute to the open source community, manage your Git repositories, review code like a pro, track bugs and fea...",
    "siteName": "GitHub",
    "title": "GitHub: Let’s build from here",
    "url": "https://github.com/"
  }
}

{
  "url": "https://en.wikipedia.org/wiki/",
  "type": "News", // There should be better options
  "similarity": 0.785299761399085,
  "preview": {
    "url": "https://en.wikipedia.org/wiki/Main_Page",
    "title": "Wikipedia, the free encyclopedia"
  }
}

{
  "url": "https://www.xiaohongshu.com/",
  "type": "Travel", // Not bad
  "similarity": 0.785504850187694,
  "preview": {
    "url": "http://www.xiaohongshu.com/explore",
    "description": "2 亿人的生活经验，都在小红书"
  }
}

References

• Embeddings - openAI
• Create embeddings API - openAI