Extract booru notes and put them in text file

[ocrhell]

Is there a way to extract notes (translations) and put them in a similarly named downloaded text file? Specifically gelbooru. When running an instance with this in config file:

        "booru":
        {
            "tags": false,
            "notes": true
        }

Only the image is downloaded and the notes aren't extracted at all. Not even in cmd. Should I be adding anything in gelbooru's block?

Thanks.

[Hrxn]

Have you tried it with a "metadata" post-processor?

https://gdl-org.github.io/docs/configuration.html#postprocessor-configuration https://gdl-org.github.io/docs/configuration.html#postprocessor-options

For example

{
    "extractor":
    {
        "booru":
        {
            "..": "..",

            "postprocessors":[

                {
                    "name" : "metadata",
                    "event": "post",
                    "mode" : "custom",
                    "skip": true,
                    "content-format": "{content|description}\n",
                    "filename": "{id}.txt"
                }
            ]

        }
    }
}

Of course, you need to check the output with -K, if it's actually {content} you want, or {description}, or whatever the name is for translation - given that the site provides something like such translations.

[ocrhell]

Closed it prematurely, sorry. Going through the notes block from gelbooru.py and gelbooru_v02.py, is it possible to filter out height, width, x, y?

notes.append({
                "width" : int(extr(note, 'data-width="', '"')[0]),
                "height": int(extr(note, 'data-height="', '"')[0]),
                "x"     : int(extr(note, 'data-x="', '"')[0]),
                "y"     : int(extr(note, 'data-y="', '"')[0]),
                "body"  : extr(note, 'data-body="', '"')[0],
            })

I've tried # but that doesn't work. Tried multiple variations of notes.x / notes.width etc... with an additional postprocessors instance with delete before and after. Also didn't work. -K gives notes[N]['width'] / notes[N]['height'] etc... and I've tried those too.