epoxy in Reports

Setup

library(epoxy)

Loading epoxy adds four new knitr engines, or chunk types. Each type lets you intermix text with R code or data (expr in the table below), and each is geared toward a different output context.

Engine Output Context Delimiter
epoxy all-purpose markdown {expr}
epoxy_html HTML {{expr}}
epoxy_latex LaTeX <<expr>>
whisker all-purpose mustache template language

⚠️ Caution: Previously, epoxy provided a glue engine, but this conflicts with a similar chunk engine by the glue package. You can update existing documents to use the epoxy engine, or you can explicitly use epoxy’s glue chunk by including the following in your setup chunk.

use_epoxy_glue_engine()

Using epoxy chunks

To use epoxy in your R Markdown document, create a new chunk using the engine of your choice. In that chunk, write in markdown, HTML, or LaTeX as needed, wrapping R expressions inside the delimiters for the epoxy chunk.

```{epoxy}
The average speed of the cars was **{mean(cars$speed)} mph.**
But on average the distance traveled was only _{mean(cars$dist)}_.
```

The average speed of the cars was 15.4 mph. But on average the distance traveled was only 42.98 ft.

epoxy is built around glue::glue(), which evaluates the R expressions in the { } and inserts the results into the string. The chunk above is equivalent to this call to glue::glue():

glue::glue("The average speed of the cars was **{mean(cars$speed)} mph**.
But on average the distance traveled was only _{mean(cars$dist)} ft_.")
#> The average speed of the cars was **15.4 mph**.
#> But on average the distance traveled was only _42.98 ft_.

One immediate advantage of using epoxy instead of glue::glue() is that RStudio’s autocompletion feature works inside epoxy chunks! Typing cars$ in the chunk will suggest the columns of cars.

Inline transformers

epoxy provides inline transformations inspired by cli’s inline markup. This transformer is enabled by default in epoxy(), epoxy_html() and epoxy_latex() and their respective knitr chunk engines.

Here’s an example using a small list containing data about a movie (expand the section below to see the full code for movie). We can use the inline transformer to format the replacement text as we build up a description from this data.

Movie data
movie <- list(
  year = 1989,
  title = "Back to the Future Part II",
  budget = 4e+07,
  domgross = 118450002,
  imdb_rating = 7.8,
  actors = c(
    "Michael J. Fox",
    "Christopher Lloyd",
    "Lea Thompson",
    "Thomas F. Wilson"
  ),
  runtime = 108L
)
```{epoxy}
The movie {.emph {.titlecase movie$title}}
was released in {.strong movie$year}.
It earned {.dollar movie$domgross}
with a budget of {.dollar movie$budget},
and it features movie stars
{.and movie$actors}.
```
The movie Back to the Future Part II was released in 1989. It earned $118,450,002 with a budget of $40,000,000, and it features movie stars Michael J. Fox, Christopher Lloyd, Lea Thompson, and Thomas F. Wilson.

Read more about inline transformations in ?epoxy_transform_inline.

Transform replaced values

You can use the epoxy_transform_wrap() with the epoxy_transform chunk option to wrap the evaluated R expression in formatting or templating text. Or you can use the pre-set epoxy_transform_bold(), epoxy_transform_italic(), or epoxy_transform_code() transformers or with epoxy_transform().

```{epoxy, .transformer = epoxy_transform("bold")}
All cars stopped between {min(cars$dist)} and {max(cars$dist)} feet
from a starting speed of {min(cars$speed)}---{max(cars$speed)}
```

All cars stopped between 2 and 120 feet from a starting speed of 4120 mph.

epoxy chunks are vectorized

Unlike inline R code, the epoxy chunks are vectorized. This can be something to watch out for or it can be an advantage:

```{epoxy}
{1:4}. "{letters[1:4]}" is for {c("apple", "banana", "coconut", "donut")}
```
  1. “a” is for apple
  2. “b” is for banana
  3. “c” is for coconut
  4. “d” is for donut

You can collapse fields automatically using the epoxy_transform_collapse() transformer. You can then choose how vectors are collapsed by adding *, & or | to the end of the expression.

  • * collapses with commas, e.g. {letters[1:3]*}.
  • & collapses with commas and adds " and " between the last two items
  • | collapses with commas and adds " or " between the last two items.
```{epoxy, .transformer = epoxy_transform("collapse")}
- The first three letters are {letters[1:3]*}.
- When capitalized, they are {LETTERS[1:3]&}.
- They're indexed by {1:3|}.
```
  • The first three letters are a, b, c.
  • When capitalized, they are A, B, and C.
  • They’re indexed by 1, 2, or 3.

You can change the separator between entries and between the last entry using the sep, last and the _and and _or specific arguments of the epoxy_transform_collapse() function.

Templating with epoxy chunks

It’s also possible to create a reusable template. Use the ref.label chunk option to reuse a template using the values in the .data chunk option, which can be a list or data frame.

mpg <- data.frame(
    manufacturer = c("Chevrolet", "Dodge", "Ford"),
    model = c("Malibu", "Caravan", "Expedition"),
    cty = c(19, 7, 11),
    hwy = c(27, 24, 17)
)
```{epoxy car-name, eval=FALSE}
- A {manufacturer} {model} gets {cty} city and {hwy} highway miles per gallon.
```

```{epoxy ref.label="car-name", .data = mpg}
```
  • A Chevrolet Malibu gets 19 city and 27 highway miles per gallon.
  • A Dodge Caravan gets 7 city and 24 highway miles per gallon.
  • A Ford Expedition gets 11 city and 17 highway miles per gallon.

Whisker engine

Sometimes the epoxy engine doesn’t quite deliver the template power you need. In these cases, you can use the whisker engine instead.

```{r}
contestant <- list(name = "R User", value = 1000, taxed = 600, in_ca = TRUE)
```

```{whisker .data = contestant, echo=FALSE}
Hello {{name}}:
You have just won ${{value}}!
{{#in_ca}}
Well, ${{taxed}}, after taxes.
{{/in_ca}}
```
contestant <- list(name = "R User", value = 1000, taxed = 600, in_ca = TRUE)

Hello R User: You have just won $1000! Well, $600, after taxes.

HTML and LaTeX chunks

Markdown chunks

The epoxy chunk engine can be used in any output format. In practice, it works best in markdown (i.e. generally in R Markdown or Quarto)

```{epoxy, .data = mpg}
- **{manufacturer}** _{model}_
```

where it renders as:

- **Chevrolet** _Malibu_
- **Dodge** _Caravan_
- **Ford** _Expedition_

If you’re writing for an HTML or LaTeX output, however, you may need to write literal HTML or LaTeX in your document. With the epoxy chun, you’d need to escape any { or } in your text by doubling them, otherwise the content within will be treated as a template expression. To avoid this friction, epoxy provides two additional chunk engines, epoxy_html for writing raw HTML and epoxy_latex for writing raw LaTeX.

Raw HTML chunks

Use the epoxy_html block to epoxy (glue) R and HTML together. The output is raw HTML. By default, expressions in these types of blocks are wrapped in {{ and }}, like whisker templates above.

<ul>
```{epoxy_html, .data = mpg}
  <li><strong>{{manufacturer}}</strong> <em>{{model}}</em></li>
```
</ul>
<ul>
```{=html}
  <li><strong>Chevrolet</strong> <em>Malibu</em></li>
  <li><strong>Dodge</strong> <em>Caravan</em></li>
  <li><strong>Ford</strong> <em>Expedition</em></li>
```
</ul>

Notice that the output is HTML but wrapped in a pandoc raw html block, which tells pandoc that the content is HTML that shouldn’t be modified1. It also means that the output of the chunk will only be included in HTML documents.

If your epoxy_html block is contained within another a raw html block, or if you want to force the output to appear, you can set the chunk option html_raw = FALSE.

````{=html}
<ul>
```{epoxy_html, .data = mpg, html_raw = FALSE}
  <li><strong>{{manufacturer}}</strong> <em>{{model}}</em></li>
```
</ul>
````

epoxy_html uses two custom transformers, epoxy_transform_inline() and epoxy_transform_html(), applying the html transformer before the inline transformer. With epoxy_transform_html() you can use element.class#id syntax to wrap expressions in HTML elements (all parts are optional). Let’s use this syntax to place manufacturer and model in <strong> and <em> elements, each with a custom class.

<ul>
```{epoxy_html, .data = mpg}
  <li>
    {{strong.car-make manufacturer}}
    {{em.car-model model}}
  </li>
```
</ul>
<ul>
```{=html}
  <li>
    <strong class="car-make">Chevrolet</strong><strong class="car-make">Dodge</strong><strong class="car-make">Ford</strong>
    <em class="car-model">Malibu</em><em class="car-model">Caravan</em><em class="car-model">Expedition</em>
  </li>
```
</ul>

Because the epoxy_transform_html() transformer uses .<class> to create <span class="class"> elements, epoxy_html() also recognizes @<inline> to access the inline transformers. So {{.uppercase manufacturer}} is assumed to be a CSS class and not an inline transformer class.

```{epoxy_html .data = mpg[1,]}
{{.uppercase manufacturer}}

{{@uppercase manufacturer}}
```
```{=html}
<span class="uppercase">Chevrolet</span>

CHEVROLET
```

Raw LaTeX chunks

Similarly, you can also use epoxy_latex chunks to epoxy R and LaTeX together. Wrap expressions in these types of chunks with << and >>.

\begin{itemize}
```{epoxy_latex, .data = mpg}
\item <<.strong manufacturer>> <<.emph model>> gets <<cty>> city and <<hwy>> highway miles per gallon.
```
\end{itemize}

In R Markdown knitting into a LaTeX output, this renders as:

\begin{itemize}
```{=latex}
\item \textbf{Chevrolet} \emph{Malibu} gets 19 city and 27 highway miles per gallon.
\item \textbf{Dodge} \emph{Caravan} gets 7 city and 24 highway miles per gallon.
\item \textbf{Ford} \emph{Expedition} gets 11 city and 17 highway miles per gallon.
```
\end{itemize}

Note that, like epoxy_html chunks, epoxy_latex places the output in raw latex blocks. This behavior can be disabled by setting the chunk option latex_raw = FALSE.

⚠️ Note: Prior to v1.0.0, epoxy used single < and > characters for expression delimiters in epoxy_latex() chunks. This can lead to subtle but inescapable problems if you need to use these characters inside your expression. As a result, epoxy_latex() now uses << and >> to delimit inline expressions.


  1. Without the raw html block, pandoc can do unexpected things to HTML, even if your output is HTML-friendly. If you are explicitly writing HTML markup in your R Markdown or Quarto document, you should probably wrap that markup in an {=html} block.↩︎