An image will be wrapped in a `<figure>` when it is the only thing in a
paragraph, so the workaround is to add a non-breaking space after the image.

This filter removes the Figure leaving the Image and an nbspace in a Para
unless the Image has a class `.fig`.

It can be used both to globally insert the needed nbspace into Markdown or
used always to so to speak implement a better syntax where you clearly mark
a lone-image paragraph if you want a figure.

https://gist.github.com/bpj/6664b0b0755e7ebe9007de4d1426d400

Den sön 11 juni 2023 04:26H <agents-FcZObrvlYduBUy7/sJONFg@public.gmane.org> skrev:

> On 06/06/2023 03:14 PM, H wrote:
>
> On 06/01/2023 04:34 AM, 'William Lupton' via pandoc-discuss wrote:
>
> Am I correct in thinking that when you say "filter" here, you really mean
> "writer"?
>
> In filters, you can use the 'traverse' field to force top-down traversal.
> See https://pandoc.org/lua-filters.html#traversal-order
>
> In writers, you can choose to call doc:walk(), which will honour
> 'traverse'. See
> https://pandoc.org/custom-writers.html#example-modified-markdown-writer
> (and example below)
>
> However, if you use pandoc.scaffolding.Writer (as I think you are?) then I
> suspect that you can't control the traversal order (but I'm not sure that
> it makes sense to do so, because, for example, surely you always want the
> contents of a <div> to have been traversed before you add '<div>' and
> '</div>'?).
>
> With the custom writer shown below, and with an input file that consists
> of the single line 'Text' (and using my logging module
> <https://github.com/pandoc-ext/logging>), you get this with the default
> 'typewise' traversal:
>
> (#) inline Str "Text"
> (#) inlines Inlines[1] {[1] Str "Text"}
> (#) block Para {content: Inlines[1] {[1] Str "Text"}}
> (#) blocks Blocks[1] {[1] Para {content: Inlines[1] {[1] Str "Text"}}}
> (#) meta Meta {}
> (#) doc Pandoc {
>   blocks: Blocks[1] {
>     [1] Para {
>       content: Inlines[1] {
>         [1] Str "Text"
>       }
>     }
>   }
>   meta: Meta {}
> }
>
> ...and this with 'topdown' traversal:
>
> (#) doc Pandoc {
>   blocks: Blocks[1] {
>     [1] Para {
>       content: Inlines[1] {
>         [1] Str "Text"
>       }
>     }
>   }
>   meta: Meta {}
> }
> (#) meta Meta {}
> (#) blocks Blocks[1] {[1] Para {content: Inlines[1] {[1] Str "Text"}}}
> (#) block Para {content: Inlines[1] {[1] Str "Text"}}
> (#) inlines Inlines[1] {[1] Str "Text"}
> (#) inline Str "Text"
>
> The writer:
>
> local logging = require 'logging'
>
> local function report(label, elem)
>     logging.temp(label, elem)
> end
>
> -- 'typewise' (default) or 'topdown'
> local traverse = 'topdown'
>
> function Writer (doc, opts)
>     local filter = {
>         traverse = traverse,
>         Inline = function(inline)
>             report('inline', inline)
>         end,
>         Inlines= function(inlines)
>             report('inlines', inlines)
>         end,
>         Block = function(block)
>             report('block', block)
>         end,
>         Blocks = function(blocks)
>             report('blocks', blocks)
>         end,
>         Meta = function(meta)
>             report('meta', meta)
>         end,
>         Pandoc = function(doc)
>             report('doc', doc)
>         end,
>     }
>     return pandoc.write(doc:walk(filter), 'html', opts)
> end
>
> Thank you for your thoughtful response. I realize I have used the term
> 'filter' and 'writer' cavalierly. I have a "slightly" complicated layout I
> want to output from the writer and my thought right now is that I should
> try to add the <div> and </div> in the right places in a filter and then
> let the writer just output the resulting html.
>
> I may have to traverse the AST several times in my filter to accomplish
> the result since multiple <div></div> need to be added at different levels
> to achieve the desired output.
>
> I just downloaded your logging filter to aid in the development and am
> also working on formalizing the rules for adding <div></div>.
>
>
> I spent some more time learning more about writing filters and have made
> some progress.
>
> I think the entire logic for adding <div></div> to achieve the desired
> layout might be doable in the filter and most of it seems to be working.
>
> One thing I found is that when an image is included using the following
> syntax
> ![test](https://www.somedomain.tld/test.jpg)
> pandoc wraps the <image> element in <figure></figure> which I think I
> don't want/need.
>
> How can I get rid of <figure> wrapped around <image> elements? At the very
> least I want got get rid of <figurecaption>some text I do not
> need</figurecaption> inside the <figure></figure> element?
>
> --
> You received this message because you are subscribed to the Google Groups
> "pandoc-discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/pandoc-discuss/992af9a4-285c-a20f-56de-347ea88f021f%40meddatainc.com
> <https://groups.google.com/d/msgid/pandoc-discuss/992af9a4-285c-a20f-56de-347ea88f021f%40meddatainc.com?utm_medium=email&utm_source=footer>
> .
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/CADAJKhDBqvOL469ZXigacDug2dMx%3DOiGWjteOM3fXqT-29BM5A%40mail.gmail.com.