[Tex/LaTex] How to convert a Google docs document to LaTeX

conversion

I am working in a team with people who didn't know LaTeX, and while I would have loved to get them into it through shareLaTeX or a similar service, we don't have the time for this. As a result, we are working on Google Docs. But while the WYSIWYG is really helpful for them, I am very frustrated with the limitations of the interface and the tediousness of making it look good¹ and I would like to get it as a LaTeX document, essentially keeping the structure (Keeping paragraphs, turning headings into \sections and so on, maybe even page breaks!) and the basic style (bold, italics…)

So: is there a way to convert a Google docs document to a LaTeX file while retaining as much of its structure as possible?

_{1. And don't get me started on the bibliography. Ugh.}

Best Answer

I have found a way through html export

Download the document as html
Use Pandoc to convert the html to LaTeX

It is not perfect but better than my old Markdown solution (see below) and preserves most of the original document. It even keeps images!

An inferior solution that I'll leave here because the export from/to Markdown could still be useful for future readers:

Use Renato Mangini's gdoc2md script to get a Markdown export of your doc. The install instructions are a bit outdated, but it's nothing to be afraid of.
Use Lunamark or Pandoc to get a LaTeX document from the resulting Markdown. You can even make it standalone (or even export to something else than LaTeX)

Now obviously it is not a very polished process and the resulting LaTeX will need some tweaking to get things right, but at least it keeps the \section structure, most of the basic formatting (not fancy things like colours, though), the lists and all the Markdown I did originally use in the Google doc to mark the places where I wanted to put fancy formatting later (e.g. > for quotes).

Related Solutions

[Tex/LaTex] Embed LaTeX math equations into Microsoft Word

You should be able to cut and paste mathematics from your web browser to Word (or any of the Micorsoft Office suite). Unfortunately at present you have to make a small edit but any text editor will do for that.

Given

x=\frac{-b\pm\sqrt{b^2-4ac}}{2a}

Make a small html file that looks like

<!DOCTYPE html>
<html>
<head>
<script type="text/javascript"
  src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML">
</script>
<title>tex texample</title>
</head>
<body>

$$x=\frac{-b\pm\sqrt{b^2-4ac}}{2a}$$

</body>
</html>

View that in a web browser and select "show MathML as/MathML Code" from the right menu:

Select the MathML text from the popup window.

Normally you can paste MathML in to word but for various reasons you need to give Word a hint in this case, so first paste it into a text editor and add the line

<?xml version="1.0"?>

to the start:

enter image description here

Then cut out the edited text and paste it into Word (any version since 2007).

enter image description here

Note the result is a fully editable Word Math Zone, using scalable fonts, not an image.

I used MathJax in a web browser for the initial TeX to MathML conversion as it is the easiest to set up, there are other alternatives. Also, to make it simple, I described the process in terms of cutting and pasting, which works well for one or two expressions but clearly not if you are converting thousands, however the process can be automated in various ways.

[Tex/LaTex] How to convert office documents (odt, doc, rtf) to clean LaTeX

I know what a typical document looks like in the field of philosophy and thus I understand that the formatting is usually not very complex, but works of philosophy are, ideally speaking, structured and semantically-rich documents. It might be worth it to consider e.g. whether the italics are meant to convey emphasis or distinguish a book title and mark them up accordingly. Also, keeping the outline and hierarchy of the headings intact might be beneficial. Which brings me to Writer2LaTeX.

I personally do not see any advantage in using RTF in this day and age and my work revolves around OpenDocument. When I have to integrate different documents and produce a clean typeset version with LaTeX (which happens more and more often), I usually rely on a combination of Pandoc and Writer2LaTeX. Here are some observations on the latter:

In order to preserve all the different kinds of styles and structural elements, one has no choice but to choose a “print” profile; for what I do, the “clean” and “ultraclean” profiles are inadequate. One reason is that I often have multi-column text sections in my ODT documents (I mean actual sections – div’s or environments if you like – not section-level headings) and these sections are not preserved in Writer2LaTeX’s output when a “clean” profile is selected. Then again, even with the “print” profile what gets produced is a multicols environment and not a custom environment with its own name, which is not as semantic as I’d like.
The results of the conversion are usually better when all manual formatting is removed and replaced with styles, even for character styles, which is why I suggest replacing italics with styles like e.g. “Emphasis” and “Work-title”. That way, every style has its own definition in the generated LaTeX preamble.
Even when trying to be as disciplined as possible, there is always cruft to remove. Writer2LaTeX respects the way that LibreOffice writes its ODF files, which means that many automatic styles are produced, even when a named style was applied. It is as though LibreOffice “instantiates” the applied styles and replaces them with automatically numbered ones (P1, P2 etc. for paragraphs, T1, T2 etc. for text spans). I find this very irritating, especially because this behavior is only present in recent versions of LibreOffice and is not present in OpenOffice 3.4.1.
In most cases, you will want to modify the command and environment definitions in the generated preamble, because they are not at all optimal (obviously, the use of an automated tool cannot replace actual knowledge of LaTeX).

Discipline and patience are required, but it is definitely possible to produce beautiful documents with OpenDocument, LibreOffice, Writer2LaTeX and some careful manual editing of the resulting LaTeX source.

Update: between 2012 and 2014, Writer2LaTeX’s development remained dormant and this was disconcerting, because the latest stable version did not work properly with LibreOffice 4.0 and later. My solution was to use the “standalone” version of Writer2LaTeX, available from the project’s website. This standalone version is operated from the command line, but I actually find it simpler to configure that way.

Now, since the end of 2014, development of Writer2LaTeX has resumed and it is better than ever. Compatibility with XeTeX has improved, making the extension much easier to use in a typical multilingual environment.

Best Answer

Related Solutions

[Tex/LaTex] Embed LaTeX math equations into Microsoft Word

[Tex/LaTex] How to convert office documents (odt, doc, rtf) to clean LaTeX

Related Question