Skip to contents

Installation

You can install the latest version of epoxy with remotes

# install.packages("remotes")
remotes::install_github("gadenbuie/epoxy")

or from gadenbuie.r-universe.dev.

options(repos = c(
  gadenbuie = "https://gadenbuie.r-universe.dev/",
  getOptions("repos")
))

install.packages("epoxy")

Setup

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

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 styles

epoxy provides inline styling inspired by cli’s inline markup. This styling 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.

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
)

We can use the inline styles to format the replacement text as we build up a description from this data.

```{epoxy echo = FALSE}
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 styles in ?epoxy_style_inline.

Style replaced values

You can use the epoxy_style_wrap() with the epoxy_style chunk option to wrap the evaluated R expression in formatting or templating text. Or you can use the pre-set epoxy_style_bold(), epoxy_style_italic(), or epoxy_style_code() style transformers or with epoxy_style().

```{epoxy, epoxy_style = epoxy_style("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_style_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, epoxy_style = epoxy_style("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_style_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.

Raw Blocks

HTML Blocks

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>
  • Chevrolet Malibu
  • Dodge Caravan
  • Ford Expedition

LaTeX Blocks

Similarly, you can also use epoxy_latex blocks to epoxy R and LaTeX together. By default, expressions in these types of blocks are wrapped in < and >.

\begin{itemize}
```{epoxy_latex, data = mpg}
\item \textbf{<manufacturer>} \textit{<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}
\item \textbf{Chevrolet} \textit{Malibu} gets 19 city and 27 highway miles per gallon.
\item \textbf{Dodge} \textit{Caravan} gets 7 city and 24 highway miles per gallon.
\item \textbf{Ford} \textit{Expedition} gets 11 city and 17 highway miles per gallon.
\end{itemize}