{
  "title": "Consistent Tags",
  "video_id": "HIH4TEUHL4g",
  "published": "2024-03-06T15:35:05+0000",
  "complexity": "Easy",
  "old_url": "https://workflows-of-ai.com/workflow/consistent-tags",
  "description": "Use the AI Interpolator OpenAI taxonomy rule with find similar tags to tag Drupal articles from text and image without duplicate synonym terms.",
  "subtitle": "Generate tags from an article's title, text and image, and reuse existing terms instead of creating new synonyms.",
  "summary": "An OpenAI taxonomy rule on the article's tags field reads the title, body and image, using tokens and GPT-4 Vision. The find similar tags option checks new tags against existing terms, so the article gets the existing \"Mark Zuckerberg aka the Zuck\" term instead of a new one. A red tie that only appears in the photo also ends up as a tag.",
  "categories": [
    "Drupal AI"
  ],
  "intro_html": "<p>The problem we try to solve today is that we want to:</p><p>1. AI generate free tags from an article text, title and image.<br>2. The image should be used as context.<br>3. The tags should be consistent so we don't regenerate synonyms and create multiple terms meaning the same thing.</p><p>The problem is this - one article might be tagged \"Jesus\", another \"Jesus Christ\" and a third \"The Messiah\". The problem is that they are the same tag, but will now be 3 different tags connected to 1 article, instead of 1 tag connected to 3 articles.</p><p>This we will solve in this video using the AI Interpolator OpenAI module.</p>",
  "intro_added": false,
  "chapters": [
    "0:00 Tags from text and image",
    "0:51 The synonym problem",
    "2:41 The article tags field",
    "3:35 The OpenAI taxonomy rule",
    "5:02 The find similar tags option",
    "5:33 Tagging a Zuckerberg article"
  ],
  "corrections": [
    [
      "druple",
      "Drupal"
    ],
    [
      "inflator open I",
      "Interpolator OpenAI"
    ],
    [
      "insolator",
      "Interpolator"
    ],
    [
      "open AI",
      "OpenAI"
    ],
    [
      "GPT for vision",
      "GPT-4 Vision"
    ],
    [
      "Bill gate",
      "Bill Gates"
    ],
    [
      "Marshall Arch",
      "martial arts"
    ],
    [
      "aka the suck",
      "aka the Zuck"
    ],
    [
      "aka the sck",
      "aka the Zuck"
    ],
    [
      "suckerberg",
      "Zuckerberg"
    ]
  ],
  "ai_note": "The video and the introduction are Marcus's own, first published on workflows-of-ai.com. AI wrote the subtitle, the summary and the video chapters from the transcript, and corrected misheard names in YouTube's automatic captions."
}