Providing #+NAME and #+CAPTION for images

[ispringle]

Orgmode has a few keywords you can specify for an image and they are #+NAME, #+CAPTION, and #+ATTR_. Using Uniorg the attr values come through when I specify ATTR_HTML (I believe this is the correct behavior). But I am not sure how to access the name or caption. For reference, I'm using AstroJS and I am overwriting img tags in my Content by specifying a replacement in the components lookup attribute.

When an image is processed the props provided include anything set in the ATTR_HTML and then the src data, so filepath, format, and dimensions. I see that the parser captures these but I'm not sure if I'm doing something wrong or if they're not being passed through for further processing. https://github.com/rasendubi/uniorg/blob/cb0709da1478d92b26391e288d1133cf959e2ff7/packages/uniorg-parse/src/parser.ts#L2162C2-L2162C2

[ispringle]

As best as I can tell, the parser is grabbing this data and putting it into the affiliated object. But then org-rehype ignores this data, extracting the attrs but not the other possible values in there such as NAME and CAPTION:

uniorg/packages/uniorg-rehype/src/org-to-hast.ts

Line 442 in cb0709d

const attrs = isFirstLink ? getAffiliatedAttrs(parent) : {};

[rasendubi]

Hey, you're correct that #+NAME: and #+CAPTION: are ignored by uniorg-rehype.

I think the good behavior is to transform

#+CAPTION: hello
[[image.png]]

into:

<figure>
  <img src="image.png" />
  <figcaption>hello</figcaption>
</figure>

I think that should be a straightforward fix somewhere around here:

uniorg/packages/uniorg-rehype/src/org-to-hast.ts

Lines 444 to 446 in cb0709d

	if (isImageLink(org, this.options)) {
	return h(org, 'img', { ...attrs, src: link });
	}

#+NAME: is not exported in org-export but can be referenced with Internal Links (currently not supported #77).

[ispringle]

Yes, I was thinking it would be best to just pass this data without changing how it renders, since org-export never reaches for the CAPTION or NAME, as you suggest. Instead it makes sense to me at least that uniorg would just ensure this data is passed on so that it can be used for further use if desired.