First write a markdown file using pandoc markdown. The package uses pandoc to parse the markdown and runs any code blocks tagged as being R-code. The result a new markdown file that can be further processed using pandoc and compiled to, for example, PDF or HTML.
To process your markdown file you can use the mdweave
function:
mdweave("mydocument.md", "mydocument_woven.md")
The resulting document mydocument_woven.md
can then be processes
further using pandoc.
Below is a ready to run example using an example document in the package:
library(simplermarkdown)
example1 <- system.file("examples/example1.md", package = "simplermarkdown")
mdweave(example1, "example1_woven.md")
system("pandoc example1_woven.md -o example1.pdf")
The package also includes the functions mdweave_to_pdf
,
mdweave_to_html
and mdweave_to_tex
, that combine these last two
function calls. Therefore, the example above could also have been
written as:
library(simplermarkdown)
example1 <- system.file("examples/example1.md", package = "simplermarkdown")
mdweave_to_pdf(example1, "example1.pdf")
Although it is possible to pass additional arguments to pandoc
through
the mdweave_to_...
functions, it is probably just as easy to call
pandoc
directly.
An example of a basic code block tagged as being R-code is shown below.
The id/label if the block is codeblock1
.
```{#codeblock1 .R}
a <- 1+1
a
```
By default the code in the code block is run and both the code and the output of the code are shown in the generated new code block. In the output file this will result in
> a <- 1 + 1
> a
[1] 2
Note that the .R
needs to be the first argument starting with a .
for the codeblock. For example ```{#codeblock .foo .R foo=bar}
won’t
be evaluated, while ```{#codeblock foo=bar .R .foo}
will.
By default the code in the code block is run and both the code and the output of the code are shown in the generated new code block. Arguments can be used to suppress either showing the code or the output:
```{#codeblock1 .R echo=FALSE results=TRUE}
a <- 1+1
```
Inline code is also supported. Again it should be tagged with {.R}
:
The average value of `Sepal.Width` is `mean(iris$Sepal.Width)`{.R} and
that of `Petal.Width` is `mean(iris$Petal.Width)`{.R}.
The final result of the inline code will always be included as text into the resulting markdown document. In case of code blocks the code is passed on to a function. Depending on the function used this can result in a code blocks with the evaluated code (the default), tables, figures and you can also specify your own functions.
To generate a table, we tell it to pass the code in the code block to
the function output_table
from the simplermarkdown
package. This
function will take the final result and generate a markdown table from
that. Any additional arguments of the code block are passed on to the
output_table
function.
```{.R fun=output_table caption="Sample iris"}
dta$foo <- dta$Sepal.Width/dta$Sepal.Length
dta[1:20, ]
```
To generate a figure use the output_figure
function. The function will
run the code, capture any output on the specified device and generate a
markdown image include.
```{.R fun=output_figure name="test" caption="My figure" device="pdf" width=8
height=6}
plot(dta$Sepal.Width, dta$Petal.Width)
```
The figures are saved in the folder figures
in the current folder.
By using the output_raw
filter, any other output can be generated.
This function will run the code, capture any output and put that
directly into the resulting markdown document. For example, let’s print
a list with all of the iris species:
```{.R fun=output_raw}
writeLines(paste("-", levels(iris$Species)))
```
Or you can write your own filter function. This function will get the code in the code block as character vector as it’s first argument, the language of the code block, and the id of the code block and any other arguments given.
For example:
print_in_bold <- function(code, language = "R", id = "", ...) {
cat("\n**", code, "**\n", sep = "")
}
This function could for example be defined in a block in the markdown file. Afterwards, this function can be used in your markdown document as:
```{.R fun=print_in_bold}
Hello World!
```
To use simplermarkdown as an engine for your R-package vignettes you will need to do the following:
VignetteBuilder: simplermarkdown
If your package doesn’t use simplermarkdown otherwise you can add it to
your Suggests
field in the DESCRIPTION
file:
Suggests:
simplermarkdown
Create the vignette in the vignettes
directory in your package source.
You have to add the following line to your vignette:
%\VignetteEngine{simplermarkdown::mdweave_to_html}
Instead of mdweave_to_html
you can also use mdweave_to_pdf
to
generate a vignette in PDF format. It is easiest to do this in a comment
section in your markdown file. For example, start your markdown file
with:
<!--
%\VignetteEngine{simplermarkdown::mdweave_to_html}
%\VignetteIndexEntry{The title of the vignette}
-->
---
title: [The title of the vignette]
---
[And the contents of your vignette]
By default the default templates and styling of the pandoc installation
on your machine will be used. However, you can also specify custom
styling in the header of your markdown file. See the documentation of
Pandoc for more information. For
example, if you generate HTML output and you want to use a custom
CSS-stylesheet, you can place the stylesheet in the vignettes
directory and refer to the stylesheet in the header:
<!--
%\VignetteEngine{simplermarkdown::mdweave_to_html}
%\VignetteIndexEntry{The title of the vignette}
-->
---
title: [The title of the vignette]
css: custom_styling.css
---
simplermarkdown tries to assume as little as possible about possible workflows. However, this also means that you, the user, are responsible for some things where other packages might make assumptions. One of the places where this is the case is for paths and working directories. And this is especially relevant when including figures and when generating figures using R.
As an example take the following project directory:
report/
report.md
figures/
figure1.png
report/output/
The report contains the following code:
![Figure caption](figures/figure1.png)
```{.R fun=output_figure name="figure2" caption="Caption", device="png"}
plot(1:10)
```
Assume the current working directory is the root of the project
directory and that we run mdweave
as:
mdweave("report/report.md", "report/output/report.md")
Figures are by default created in the directory figures
in the target
directory. Therefore the directory structure after running mdweave
is:
report/
report.md
figures/
figure1.png
ouput/
report.md
figures/
figure2.png
And the resulting markdown file will contain the following markdown:
![Figure caption](figures/figure1.png)
![Caption](report/output/figures/figure2.png)
As you can see, we now have two locations with figures. When running
pandoc
from the root directory of the project to create the final
output:
pandoc report/output/report.md -s -o report/output/report.html
pandoc
will not be able to find the first figure. It will find the
second figure. When you would run the final pandoc
command from the
report/output
directory. pandoc
will not be able to find any of the
figures.
There are several possible solutions for the example above:
- When working on linux or mac, you could create a symbolic link from
report/output/figures
toreport/figures
. - Copy
report/figures
toreport/output/figures
. - Path of least resistance: run
mdweave
andpandoc
from thereport
directory and also put the output in the same directory. - And probably others.
Note that the same issues occur when referencing stylsheets etc. in the meta block of the markdown file.