Pandoc's current HTML figure handling. #8

argent0 · 2021-05-31T15:16:13Z

argent0
May 31, 2021

Hi, I've been studying how does pandoc currently handles figures in the HTML
format.

The relevant function is pFigure quoted next:

Lines 537 to 555 in f198356

    
           pFigure :: PandocMonad m => TagParser m Blocks 
        
           pFigure = try $ do 
        
             TagOpen _ _ <- pSatisfy (matchTagOpen "figure" []) 
        
             skipMany pBlank 
        
             let pImg  = (\x -> (Just x, Nothing)) <$> 
        
                          (pInTag TagsOmittable "p" pImage <* skipMany pBlank) 
        
                 pCapt = (\x -> (Nothing, Just x)) <$> do 
        
                           bs <- pInTags "figcaption" block 
        
                           return $ blocksToInlines' $ B.toList bs 
        
                 pSkip = (Nothing, Nothing) <$ pSatisfy (not . matchTagClose "figure") 
        
             res <- many (pImg <|> pCapt <|> pSkip) 
        
             let mbimg = msum $ map fst res 
        
             let mbcap = msum $ map snd res 
        
             TagClose _ <- pSatisfy (matchTagClose "figure") 
        
             let caption = fromMaybe mempty mbcap 
        
             case B.toList <$> mbimg of 
        
                  Just [Image attr _ (url, tit)] -> 
        
                    return $ B.para $ B.imageWith attr url ("fig:" <> tit) caption 
        
                  _ -> mzero

This is where(line 554) the "fig:" prefix is added to the title.

I also ran some experiments with the following HTML input:


<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <title></title>
</head>
<body>

    <h1> Text Float </h1>

    <p> Some paragraph </p>

    <figure>
        <p> Some text </p>
        <ul>
            <li>This is an item of a list</li>
        </ul>
        <figcaption>
            This is the caption of the first figure.
        </figcaption>
    </figure>

    <h1> Image Float </h1>

    <figure>
        <p> Also some paragraph here </p>
        <ul>
            <li>This is an item of a list</li>
        </ul>
        <img src="../media/rId25.jpg" alt="the moon">
        <figcaption>
            This is the caption of the second figure.
        </figcaption>
    </figure>

    <h1> Contrived Image Float </h1>

    <figure>

        <p> Also some paragraph here </p>

        <ul>
            <li>This is an item of a list</li>
        </ul>

        <img src="../media/rId25.jpg" alt="the moon">
        <img src="../media/rId25.jpg" alt="the moon">

        <figcaption>
            This is the caption of the second figure.
        </figcaption>
        <figcaption>
            This is the caption too of the second figure.
        </figcaption>

    </figure>

</body>
</html>

which renders (using Firefox) to:

Current `pandoc` handling

$ stack run pandoc -- --version
pandoc 2.14
...

Native representation

$ stack run pandoc -- figure.html -t native

[Header 1 ("text-float",[],[]) [Str "Text",Space,Str "Float"]
,Para [Str "Some",Space,Str "paragraph."]
,Para [Str "Some",Space,Str "text"]
,BulletList
 [[Plain [Str "This",Space,Str "is",Space,Str "an",Space,Str "item",Space,Str "of",Space,Str "a",Space,Str "list"]]]
,Para [Str "This",Space,Str "is",Space,Str "the",Space,Str "caption",Space,Str "of",Space,Str "the",Space,Str "first",Space,Str "figure."]

,Header 1 ("image-float",[],[]) [Str "Image",Space,Str "Float"]
,Para [Image ("",[],[]) [Str "This",Space,Str "is",Space,Str "the",Space,Str "caption",Space,Str "of",Space,Str "the",Space,Str "second",Space,Str "figure."] ("../media/rId25.jpg","fig:")]

,Header 1 ("contrived-image-float",[],[]) [Str "Contrived",Space,Str "Image",Space,Str "Float"]
,Para [Image ("",[],[]) [Str "This",Space,Str "is",Space,Str "the",Space,Str "caption",Space,Str "of",Space,Str "the",Space,Str "second",Space,Str "figure."] ("../media/rId25.jpg","fig:")]]

The relationship among the elements in the first figure is lost. Perhaps that
information may be preserved using the Div constructor or the newly proposed
Figure.
One image and one caption are also lost in the third heading. Is this the
intended behavior given lines 548-549? @tarleb
The paragraphs in the last two figures are lost

HTML roundtrip

For reference.

$ stack run pandoc -- figure.html -t html

<h1 id="text-float">Text Float</h1>
<p>Some paragraph.</p>
<p>Some text</p>
<ul>
<li>This is an item of a list</li>
</ul>
<p>This is the caption of the first figure.</p>

<h1 id="image-float">Image Float</h1>
<figure>
<img src="../media/rId25.jpg" alt="This is the caption of the second figure." /><figcaption aria-hidden="true">This is the caption of the second figure.</figcaption>
</figure>

<h1 id="contrived-image-float">Contrived Image Float</h1>
<figure>
<img src="../media/rId25.jpg" alt="This is the caption of the second figure." /><figcaption aria-hidden="true">This is the caption of the second figure.</figcaption>
</figure>

Conclusion

Introducing the new constructor would allow pandc to preserve more information
about figures. Although there are, less specific, alternatives to achieve the
same using existing constructors.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Pandoc's current HTML figure handling. #8

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Replies: 0 comments

Select a reply

Pandoc's current HTML figure handling. #8

argent0 May 31, 2021

Current pandoc handling

Native representation

HTML roundtrip

Conclusion

Replies: 0 comments

argent0
May 31, 2021

Current `pandoc` handling