Skip to content

Latest commit

 

History

History
155 lines (107 loc) · 7.63 KB

README.md

File metadata and controls

155 lines (107 loc) · 7.63 KB
Raw

word2grattex

word2grattex is an R packaged deisgned to make Word to LaTeX conversion faster and more accurate. It is relatively flexible but is most effective when used within the wider Grattan report production ecosystem.

Importantly: because of unpredictable (and usually accidental) Word formatting choices, this tool is certainly more of an art than it is a science. For example, there may be loose brackets or braces in your post-conversion document that you'll need to tidy yourself (sorry!).

There may also be incorrect conversions of references when using bib2grattex (via word2grattex(bibReplace = TRUE), the default), because of 'Not Quite Twins' see here.

That being said: it will do most of the work for you, and the problems it can create are -- in 25 full conversions and counting -- fewer than the problems it solves.

It requires:

  • R and pandoc (which usually ships with R, otherwise download here)
  • A Word document in the Grattan report template.
    • Eg Mapping Australian higher education 2018.docx

It can do the most if it also has:

  • The .bib file associated with the report.
    • Eg mapping2018.bib.
  • In-text citations in Word, formatted as The Style of Quiet Achievers or Harvard Reference Format 1.
    • Eg Norton & Cherastidtham (2018);
    • or (Norton & Cherastidtham 2018).
  • The .pdf chart deck in order of appearance in the report.
    • Eg mapping2018.pdf
    • Note that if this is missing, the figure environments will default to calling chartdeck.pdf, which can easily be manually fixed using find-replace.
  • Linked or manual cross-references (not broken).
    • Eg See figure 14 in section 3.4.

Pre-function set up: before you use it

Before you use the function, follow the steps described under Starting a new report on the Grattex homepage.

Once you have created a repository for your new report, add your .docx, .bib and .pdf files. Your repository should look something like this, using the Mapping Australian higher education 2018 example:

./he-mapping-2018/
    - atlas
    - bib
    - doc
    - logos
    - tests
    - travis
    - Report.tex
    - mapping2018.docx
    - mapping2018.bib
    - mapping2018.pdf
    etc

Note that you can run the function before you set up a report repository. Just point the word2grattex function to any folder that contains your .docx (and .bib, .pdf) files.

Install and call: how to use it

The package is run through R and can be installed using remotes::github_install:

# Install devtools if you haven't already (remove the comment):
# install.packages(remotes)

remotes::install_github("wfmackey/word2grattex")

library(word2grattex)

The package contains two large functions: word2grattex and bib2grattex. Note that, by default, word2grattex runs bib2grattex. To run all features described in the what it does section below, point word2grattex to a folder that contains a .docx document and a .bib file (and, preferably, a .pdf file containing charts. If this is absent, figure environments will be built using the default chartpack.pdf).

For our Mapping Australian higher education 2018 example, we can then run:

word2grattex(path = "Dropbox (Grattan Institute)/Apps/Overleaf/mapping2018")

Note: if you do not have a bib file, set bibReplace = F:

word2grattex(path = "Dropbox (Grattan Institute)/Apps/Overleaf/mapping2018",
             bibReplace = F)

Also, there is a yet-to-be-addressed issue with Tables that pops up. If you receive an error about tables, please set buildTables = F:

word2grattex(path = "Dropbox (Grattan Institute)/Apps/Overleaf/mapping2018",
             bibReplace = F, buildTables = F)

This will produce a .tex file.

If there are any issues, please get in touch with Will.

Process and output: what it does

The function takes a Word document in the Grattan template and:

  • Converts it to LaTeX using pandoc.
  • Adds the current LaTeX preamble of the Grattan report template.
  • Cleans up after the pandoc conversion:
    • removes \protect, \hyperlink, \hypertarget, \texorpdfstring;
    • removes empty headings.
  • Adds graphics where an image was found in the Word document:
    • builds Figure environment;
    • converts Word metadata to LaTeX \caption, \unit, \noteswithsource (or \notes/\source only, or \noteswithsources, etc);
    • creates \insertgraphics for a standard Grattan chart size and inserts the nth page of the name of your PDF chart deck, or defaults to the nth page of chartdeck.pdf if a PDF is not provided (this can be quickly fixed afterwards with find/replace);
    • labels charts with fig:figure-caption.
  • Builds Table environments.
    • It doesn't build your tables sorry :(.
    • But, look, it has a proper crack at it. It will make the rows (kind of) but will use a \longtable format by default. i.e. it does a bit. The table will be commented out for compiling convenience, but if it's an easy table it might work (kind of) out-of-the-box.
    • See the kableExtra package in R, or Excel2LaTeX Excel plugin for assistance.
  • Applies appropriate labels to chapters, sections, subsections, etc.:
    • chap:, sec:, subsec:, etc.
  • Replaces cross-references with appropriate labels.
    • "See Section 2.2" → See \Cref{subsec:section-name}.
  • Replaces figure-references with appropriate labels.
    • "Figure 19 shows" → \Cref{fig:figure-caption} shows.
  • Replaces in-text citations with appropriate cite commands.
    • Uses bib2grattex function.
    • .bib keys are automatically generated in the format AuthorYearTitle (default is to cap the title to 20 characters).
      • Handles (all-but-one) citation complications:
      • 'Norton et al. (2018a).' → \footcite{Norton2018droppingoutthecostsa}
      • 'See discussion in Terrill (2018), p. 10.' → \footnote{See discussion in \textcite[][10]{Terrill2018unfreezingdiscountra}.}.
      • 'Daley and Wood (2016), chapter 1; Smith (1776e), chapter 3.' → \footcites[][chapter~3]{Daley2018hotproperty}[][chapter~3]{Smith1776thewealthofnat}.}.

Note: as our Style of Quiet Achievers fails to distinguish between 'not-quite twins', the bib2grattex conversion can't tell the difference. A solution to this problem is being considered. For now, manual identification of not-quite twins is required. Mainly: check your Daley et al. references.

Finalisation: what you do now

When word2grattex is finished, it will produce a .tex file that can be built out of the box. Some things need to be done manually:

  • Check for errors.
  • Add Tables.
  • Add Box environments.
    • This feature can't be added because there is no way to tell when a box ends.
  • Add Overview and Recommendations. Update Acknowledgements, ISBN, report number, FrontPage.
  • Optimise figure placement.
    • Use FigurePlacementScore to help.
  • Proofread lots.

Then, run through checkGrattan/Travis and release.

A continually improved tool

This is a work-in-progress. You will notice errors or think something could be improved (these ideas usually come when you're doing something repetively after the conversion and think "heck I wish this could be done automatically").

If you do, please get in touch. This can be done by raising an issue on the word2grattex Github page (my preference -- it helps keep everything in one place!), or by emailing william.mackey@grattaninstitute.edu.au.