04_rlang.Rmd

---
title: "Touring tidyverse"
subtitle: "rlang"
output:
  xaringan::moon_reader:
    lib_dir: libs
    nature:
      highlightStyle: github
      highlightLines: true
      countIncrementalSlides: false
---

```{r setup, include=FALSE}
options(htmltools.dir.version = FALSE)
```

background-image: url(https://raw.githubusercontent.com/r-lib/rlang/master/man/figures/rlang.png)
background-size: 100px
background-position: 90% 6%

# rlang

The rlang package provides tools to work with core language features of R and the tidyverse:

1. The tidy eval framework, which is a well-founded system for non-standard evaluation built on quasiquotation (`!!`) and quoted arguments (`enquo()`).
1. User-friendly error reporting with backtraces and chained errors (`abort()`, `trace_back()`, `with_abort()`).
1. A consistent API for working with base types. Note that overall this is a work in progress that is still in flux.

---
# Initial commit of rlang/lazyeval

First commit: `2014-08-12 21:15` by Hadley Wickham.

Current iteration of rlang is not the first attempt to introduce quasiquotation.
```{r, eval=FALSE}
library(lazyeval)
f = function(col1, col2, new_col_name) {
    mutate_call = lazyeval::interp(~ a + b, 
                                   a = as.name(col1), 
                                   b = as.name(col2))
    mtcars %>% mutate_(.dots = setNames(list(mutate_call), 
                                        new_col_name))
}
```
---
# Current state

1. Current version - `r packageVersion("rlang")` (dev version!).
1. https://github.com/r-lib/rlang
1. Developed by __Lionel Henry__, Hadley Wickham.
1. 3200+ commits by 31 contributors.

---

# Why tidyeval

1. Programming over quoting `tidyverse` functions (e.g., `dplyr`, `ggplot2`) <sup>1</sup>.

.footnote[
[1] although it's likely you don't need it for most tasks
]

---
background-image: url(https://raw.githubusercontent.com/r-lib/rlang/master/man/figures/rlang.png)
background-size: 100px
background-position: 90% 6%

# 5 big ideas of tidyeval <sup>1</sup>

1. R code is a tree.
1. You can capture the tree by quoting.
1. Unquoting makes it easy to build trees.
1. Quoting + unquoting to write functions.
1. Quosures capture expression and environment.

.footnote[
[1] https://www.youtube.com/watch?v=nERXS3ssntw
]

---
```{r, echo = FALSE}
countdown::countdown_fullscreen(minutes = 4, update_every = 30,
                     padding = "20%", margin = "5%",
                     font_size = "8em", line_height = "1.25")
````


---
```{r, echo = FALSE}
countdown::countdown_fullscreen(minutes = 4, update_every = 30,
                     padding = "20%", margin = "5%",
                     font_size = "8em", line_height = "1.25")
````

---
# Resources

1. `tidyeval` book - https://tidyeval.tidyverse.org/ or `tidyeval` post - https://rpubs.com/lionel-/tidyeval-introduction
1. `tidyeval` webinar - https://www.rstudio.com/resources/webinars/tidy-eval/
1. "Tidy evaluation in 5 minutes" by Hadley Wickham - https://www.youtube.com/watch?v=nERXS3ssntw
1.  Metaprogramming chapters in "Advanced R" - https://adv-r.hadley.nz/meta.html
1. `tidyeval` cheatsheet - https://www.rstudio.com/resources/cheatsheets/
1. Why you might not need tidyeval - https://resources.rstudio.com/rstudio-conf-2019/working-with-names-and-expressions-in-your-tidy-eval-code and https://resources.rstudio.com/rstudio-conf-2019/lazy-evaluation