Remove Styles but Retain Formatting

[dstark]

In my field, publishers often require documents to be submitted with direct formatting only (i.e., no styles other than the default style).

So, authors either need to (1) never use styles or (2), before submission, convert documents using styles so that they retain their formatting but remove the styles.

There's a very cumbersome way to do this with Word and Wordpad. There's seemingly no equivalent for LibreOffice that wouldn't require using the same process of shuttling documents through Word and Wordpad, which then becomes even messier.

Is this something Pandoc can or could do with a flag that would produce a document that was formatted as if it used however many predefined styles (e.g., from whatever provided reference file) but only actually has in use a given the default style(s) (e.g., Normal)?

If so, we'd be deeply grateful for any pointers or possible efforts anyone with the appropriate technical expertise might be able to provide in this direction. And of course, if I can provide any other details that might be helpful in this context, please feel free to ask.

Thanks so much!

[rnwst]

If I understand correctly, on an OOXML level, you want to remove all <w:pStyle w:val="StyleName" /> statements in /word/document.xml and replace them with style contents, such as

<w:rPr>
  <w:color w:val="FF0000" />
  <w:u w:val="single" />
</w:rPr>

or whatever style properties StyleName has, as defined in /word/styles.xml. To streamline your current (cumbersome) approach, you could probably write a script (e.g. in Powershell) to unzip the DOCX, modify the XML to replace all style declarations with thier contents, and then zip things back up again. This approach would be complicated by styles which are based on other styles though, which would require some careful thought to be put into the implementation of such a script. Alternatively, this could probably be done as well with a Word macro (though I wouldn't be able to help with that).

Regarding pandoc, such functionality does not currently exist to my knowledge. It is also a rather specialised use case from what I can see, so I'm not sure if it would be implemented.

I do wonder, however, why publishers wouldn't want styles to be used. Surely they can modify the DOCX to supply their own styles, which would obviously require the document to use styles in the first place? If direct formatting was used, this wouldn't be possible.

[dstark]

this could probably be done as well with a Word macro (though I wouldn't be able to help with that).

Thanks @rnwst. That's the heart of the current approach with some additional maneuvering before and after the conversion.

why publishers wouldn't want styles to be used. Surely they can modify the DOCX to supply their own styles, which would obviously require the document to use styles in the first place? If direct formatting was used, this wouldn't be possible.

I've wondered this too and did ask recently. thinking that perhaps styles might complicate migration into InDesign or something. The not-very-informative response I got was, in part, as follows:

Word documents will be the best way to submit files …. This allows the editorial team … to easily go through the manuscript prior to production and ensure the manuscript is ready to move forward. By keeping specific style and formatting out of the documents, the production team is then easily able to typeset the manuscript and format on their end with ease. We do not use InDesign for typesetting ….

With the current macro process, I can remove styles from a volume in a couple hours. But this part of the process of preparing the manuscript still seems comparatively medieval.

[jgm]

You could try pandoc -t rtf | pandoc -f rtf -t docx.

[dstark]

Thanks so much. -s does restore the missing metadata. HTML, LaTeX, and DocBook all produce more-or-less the same output results (with variations but with various styles still in use). I'll have a look through the other formats that Pandoc handles both as outputs and as inputs and see if I can find a combination that works. Appreciate your thoughts!

[jgm]

Docx reader should be able to handle the footnotes, and these should round-trip through LaTeX at least, so not sure what is happening here?

[dstark]

Ah, my apologies for not being sufficiently precise—HTML, LaTeX, and DocBook all produce more-or-less the same output results as each other. They do retain footnotes (unlike RTF). So that's an upside. But round-tripping through all three of those formats still yields a final DOCX that still uses various styles.

[jgm]

Sorry for putting you on the wrong track.

[dstark]

No worries at all. Appreciate the suggestions and will continue looking for a better solution than I have currently, particularly one that isn't specifically tied to Microsoft Word.

[dstark]

@jgm and @rnwst, as an alternative approach, is this something that could perhaps be done with a custom writer? E.g., have Pandoc ingest a file and write a DOCX output that simply applies the otherwise-handled-by-styles formatting (e.g., makes level 1 headings centered, applies a first-line indent to footnotes).

[jgm]

Probably not; creating a docx is complex and it's not just a matter of generating a text document, which is what our custom writers do.