This template is based on Andrew Heiss' Global-Pandoc-Files and Portable Pandoc Magic to convert Markdown-based documents into Word (docx through odt), HTML, and PDF (through xelatex).
- π tex_out: LaTeX support files
- π pandoc: (1) portable pandoc filters, (2) templates, (3) necessary fonts and (4) ad hoc scripts
- π scripts: R scripts of the analysis - they are called in the preamble of the
rmd-paper.Rmd
. - π sections: include all other (external)
.md
files necessary for the paper (e.g./sections/introduction.md
orsections/conclusion.md
).
- Added
pandoc-crossref.exe
in the project folder. pandoc/templates/xelatex.tex
andpandoc/templates/xelatex-manuscript.tex
replace fontIncosoloataGo
withIncosolata
.- Github repo includes
tex_out/rmd-paper.pdf
, but not any other supporting output files from.tex
.
You'll need to install these things:
- pandoc: Install either with
brew install pandoc
or by downloading it from pandoc.org. - make: The workhorse behind all the conversion is
make
, which uses thisMakefile
to generate different pandoc incantations. On macOS, open Terminal and runxcode-select --install
to install a handful of developer tools, includingmake
. On Windows, follow these instructions. - R: Needed to convert R Markdown files to Markdown (if you're using R Markdown). Also needed to get word count when you run
make count
. Download and install from r-project.org. Ensure you have the following packages installed: tidyverse, knitr, rvest, and stringi (runinstall.packages(c("tidyverse", "knitr", "rvest", "stringi"))
) - Python 3: Install either with
brew install python
or by downloading it from python.org. - TeX: If you want to do anything with PDFs, install LaTeX. It's easiest to just install the massive MacTeX distribution on macOS (or some Windows distribution if you use Windows).
- pandoc-include: Filter for inserting external Markdown files with syntax like
!include path/to/file.md
. Install withpip install pandoc-include
. - pandoc-citeproc: Filter for dealing with bibliographies. Install with
brew install pandoc-citeproc
. It comes with pandoc if you install it from pandoc.org. - pandoc-crossref: Filter for creating "Figure 1" and "Table 3" cross references. Install with
brew install pandoc-crossref
. Alternatively, export the.exe
in the same folder of your project's directory. - bibtool: Script for parsing and dealing with BibTeX files. Used for extracting cited references into a standalone
.bib
files when you runmake bib
. Install withbrew install bib-tool
. - gawk: The version of awk that comes with macOS by default doesn't work correctly with the script that inserts git commit information in the footer of PDFs. Install a more recent one with
brew install gawk
. - LibreOffice: Open source clone of Microsoft Office. Used for converting
.odt
files to.docx
when you runmake docx
. Install by downloading their installer. - Fonts: There are a bunch of fonts included in the
pandoc/fonts/
folder. Install these as needed - ideally for all users. If not installed for all users, you may need to repeat this step in the future.
STEP 1: Install the Installation items
If done STEP 1 before, no need to repeat. Note that if you haven't installed the pandoc/fonts
for all users, you may experience issues in compiling. In that case, you may reinstall the fonts.
-
Create a Markdown file (or R Markdown file) in some directory. Place
Makefile
and thepandoc
folder in the same directory. If you're using a bibliography, include a BibTeX file in the same directory. The directory should look like this:. βββ Makefile βββ manuscript.Rmd βββ references.bib βββ sections βββ introduction.md βββ conclusions.md βββ data βββ derived βββ manual βββ raw βββ scripts βββ pandoc βββ bin βββ csl βββ fonts βββ templates
-
Open
Makefile
and change theSRC
andBIB_FILE
variables to match your (R) Markdown file and BibTeX file names. -
Change any of the other modifiable variables in
Makefile
; e.g.ENDFLOAT
orBLINDED
. -
Use your
scripts/
to perform your empirical analysis in R (or any other software required). -
Save your data frames in
data/derived/
(as.csv
or.rds
) or store R objects as.RData
. For the latter, see here. -
Write stuff in your (R) Markdown file. I recommend you make changes only to your
rmd-paper.Rmd
. -
To convert from Markdown to something else, open a terminal window to your main directory and type
make html
ormake docx
, etc. Here are all the different things you can include aftermake
:make md
: Convert R Markdown to regular Markdownmake html
: Create HTML filemake tex
: Create nice PDF through xelatex inTEX_DIR
foldermake mstex
: Create manuscripty PDF through xelatex inTEX_DIR
foldermake odt
: Create ODT filemake docx
: Create Word file (through LibreOffice)make ms
: Create manuscripty ODT filemake msdocx
: Create manuscripty Word file (through LibreOffice)make bib
: Extract bibliography references to a standalone.bib
filemake count
: Count the words in the manuscriptmake clean
: Remove all output filesmake all
: Creates all the files above
Through the magic of
make
, you can combine any of these, likemake html docx tex
ormake html msdocx mstex
, etc.
make tex
may need several times to run until it produces a
- That's it! Write more, run
make SOMETHING
again, write more, runmake SOMETHING
again, and so on until you have a beautiful final document.
There are two complete minimal examples included in this repository: md-paper.md
(regular Markdown) and rmd-paper.Rmd
(R Markdown). Change the SRC
variable to match one of their names, run make SOMETHING
, and see what happens.
rmd-paper.md
is generated from rmd-paper.Rmd
. Only edit the .Rmd
file, not the .md
file.
You can include other Markdown files (like tables generated from R, for instance) using the following syntax:
!include path/to/file.md
If you set BLINDED = TRUE
in the Makefile, a Python script named accecare.py
will run before compiling the document. Look at the CSV file in pandoc/bin/replacements.csv
to see how to blind specific words and phrases. See the documentation for accecare.py
here.
If you set VC_ENABLE = TRUE
in the Makefile, the current git commit will be included in the footer of your PDF only when running make tex
(make mstex
doesn't do this). Make sure you have a git-repo
entry in your YAML front matter so that it can create a link to that commit at GitHub, GitLab, etc.
You can move all the figures and tables to the end of the document by setting ENDFLOAT = TRUE
in the Makefile. Some journals have this horribly backwards requirement Β―\_(γ)_/Β―. This only happens when running make mstex
; if you need all the figures and tables at the end of a Word file, you'll have to do it manually.
Word and HTML can choke on PDF images, so those targets use a helper script (pandoc/bin/replace_pdfs.py
) to replace all references to PDFs with PNGs andβif neededβconvert existing PDFs to PNG using sips
. However, there are times when it's better to not convert to PNG on the fly, like when using high resolution PNGs exported from R with ggsave()
and Cairo. To disable on-the-fly conversion and supply your own PNGs, use PNG_CONVERT = --no-convert
. The script will still replace references to PDFs with PNGs, but will not convert the PDFs.
You can embed plots in documents automatically. BUT it does not play well with pandoc-crossref
.
For pandoc-crossref
to work, you have to use this syntax:
Here is some text that refers to @fig:myfig.
![Caption for figure](output/myfig.pdf){#fig:myfig}
There's unfortunately no way to get that {#fig:myfig}
into the correct place in a knitted R Markdown document. A solution is to not use R/knitr to include figures. Instead you may create the figures elsewhereβeither in a different R script, or in a chunk in the documentβand then save them to disk as PDF or PNG (or both). Then, use standard Markdown + pandoc-crossref syntax (![](){}
) to include them:
```{r create-figure, echo=FALSE, warning=FALSE, error=FALSE}
library(ggplot)
plot1 <- ggplot(...)
ggsave("output/myfig.pdf", plot1,
width = 5, height = 3, device = cairo_pdf)
ggsave("output/myfig.png", plot1,
width = 5, height = 3, type = "cairo", dpi = 300)
```
Here is some text that refers to @fig:myfig.
![Caption for figure](output/myfig.pdf){#fig:myfig}
You don't need to do this with tables, though. If you use pandoc.table()
from the pander library, you can include the correct pandoc-crossref syntax in the table caption:
```{r example-table, echo=FALSE, warning=FALSE, message=FALSE, results="asis"}
library(tibble)
library(magrittr)
library(pander)
tribble(
~Heading, ~`Other heading`,
2, 3,
5, 7,
9, 1
) %>%
pandoc.table(caption = "This is a table {#tbl:mytable}")
```