IDs can I let ChatGPT generate them?

The CMS Questions
1

I have a lot of pages, with a lot of text inside one field. I now want to break the text up into blocks. Rather than do this manually, which would take hours, I am hoping to get ChatGPT to modify my content .txt files to put the text into blocks. It is all working fine and ChatGPT is generating IDs for the blocks like this:

Block-content:
[
    {
        "content": {
            "block_heading": "Background",
            "block_text": "Some text"
        },
        "id": "d3084133-a89a-4a26-a124-82b1dd226ec4",
        "isHidden": false,
        "type": "block_section"
    },
    {
        "content": {
            "block_heading": "Future",
            "block_text": "Some text"
        },
        "id": "c6a4e6d2-678f-4a5b-b3d6-3510d5df2b4b",
        "isHidden": false,
        "type": "block_section"
    }
]

It seems to work in the Panel and in the frontend – but is this okay or is my plan going to run into trouble? Does Kirby have to generate the IDs?

I tried leaving the "id": "" empty, but the Panel and frontend didn’t like that. I’ve also tried removing the "id" line entirely and that worked.

So I don’t know what is best. Get ChatGPT to generate IDs or remove the line entirely and have blocks with no ID?

2

They have to be unique - which can be a problem with these LLMs as they are not putting any actual logic to their output, just probability. So most times they’ll probably be unique, but no guarantee that they always are.

3

Thanks for the quick response! I’ll take my chances with what is generated!

Can I ask, if there was an element with the same ID on my website, that might created problems? But what if there was an element with the same ID but on a different website - presumably that wouldn’t be a potential problem?

4

If you mean website page, not sure - there could be edge cases. Best to have truly unique IDs.