docx manuscript fails to open on MS word

[cgreene]

We've got a manuscript here where we're not able to open in microsoft word on mac when building with the latest rootstock:
https://github.com/greenelab/nature_news_manuscript/blob/output/manuscript.docx

Has anyone else experienced this before?

[nrosed]

So I was able to get it to work by changing citeproc to pandoc-citeproc in build/pandoc/defaults/common.yaml

Not sure if this is the most correct fix... I also tried just commenting out citeproc, but this did not work. I did this because I thought citeproc was already included in pandoc and the extra filter was not needed.

You can see the commit here: greenelab/nature_news_manuscript@e042065

[nrosed]

Also, this was after updating rootstock and rebuilding my environment (greenelab/nature_news_manuscript#16)

[agitter]

I can't open your linked docx file in Word on Windows either. I could open a recent docx from the COVID-19 review though.

The error from Word points to this part of the XML

<w:t xml:space="preserve">2</w:t></w:r></w:hyperlink><w:r><w:t xml:space="preserve">.</w:t></w:r><w:r><w:t xml:space="preserve"> </w:t></w:r><w:r><w:t xml:space="preserve">We did not consider any names where no prediction could be made or quotes where neither speaker nor gendered pronoun was associated.</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val="BodyText" /></w:pPr><w:bookmarkStart w:id="0" w:name="eq:quote"/><w:r><w:t></

An XML viewer reported

This page contains the following errors:
error on line 1 at column 66619: Opening and ending tag mismatch: t line 1 and p

Based on where the XML syntax error occurs, it points to a problem exporting these equations to docx

$$\textrm{Prop. Male Quotes} = \frac{|\textrm{Male Speaker Quotes}|} {|\textrm{Male or Female Speaker Quotes}|}$${#eq:quote}
$$\textrm{Prop. Male First Authors} = \frac{|\textrm{Male First Authors}|} {|\textrm{Male or Female First Authors}|}$${#eq:first-author}

That's consistent with other errors with equation numbering in docx manubot/rootstock#435. The other issue linked in that thread suggests it may have started in pandoc 2.11.3.

[agitter]

Is this equation numbering problem blocking your resubmission? In manubot/rootstock#435 (comment) I describe initial attempts to use pandoc-crossref for numbering. It has some peculiarities and breaks some of Manubot's interactive HTML elements. However, it can export a DOCX with equation numbers that I can open.

[nrosed]

Thanks @agitter, no we were able to work around it and resubmit. I was able to get it to work somehow in the most recent build, but not sure if it is a transient issue and if it will come back later.

[miltondp]

I ran across the same issue with this manuscript. The final Word file does not open in Word, and I get an error with LibreOffice Writer 7.3 in Ubuntu:

After opening it with LibreOffice Writer, though, the file is only read until it reaches Equation 1 (not showing it) in section "Methods and materials." Also, images are not shown. This used to work with previous versions of manubot/rootstock/conda environment.

As a workaround, I created an old conda environment with pandoc 2.9 and build locally. For that, I used this version of the build/environment.yml file and ran conda env create -n manubot-pandoc29 -f environment.yml. I also applied the patch by @nrosed here to use pandoc-citeproc instead of citeproc:

index bfa5d97..12c0c34 100644
--- a/build/pandoc/defaults/common.yaml
+++ b/build/pandoc/defaults/common.yaml
@@ -6,7 +6,7 @@ filters:
 - pandoc-eqnos
 - pandoc-tablenos
 - pandoc-manubot-cite
-- citeproc
+- pandoc-citeproc
 wrap: preserve
 metadata:
   csl: build/assets/style.csl

[dhimmel]

Noting the underlying issue tomduck/pandoc-eqnos#60. One alternative might be to manually number equations if switching to pandoc-crossref is problematic.