Write R Code

background-image: url(https://i.imgur.com/CVIZGyY.jpg)
background-position: top
background-size: 100%
class: inverse, shadow

body {
  font-family: 'Fira Sans','Droid Serif', 'Palatino Linotype', 'Book Antiqua', Palatino, 'Microsoft YaHei', 'Songti SC', serif;
}

a, a > code {
  color: #EC99C6;
  text-decoration: none;
}

strong {
  color: #E69F00;
}

em {
  color: #56B4E9;
  font-style: normal;
  font-weight: bold;
}

del {
  color: #E5E5E5;
  text-decoration: none;
  font-weight: bold;
}

.remark-code {
  font-family: 'IBM Plex Mono', 'Lucida Console', Monaco, monospace;
  font-size: 100%;
}

.remark-inline-code {
  font-family: 'Fira Sans', 'Lucida Console', Monaco, monospace;
  font-weight: 400;
  font-size: 100%;
}

.remark-code-line-highlighted {
  background-color: #CEE9FF;
  font-weight: 500;
}

.large { font-size: 130% }
.medium { font-size: 115% }
.small { font-size: 70% }

.remark-slide-content {
  color: #474747;
  font-weight: 300;
  font-weight: 300;
  padding: 1em 2em 1em 2em
}

h1 {
  color: #56B4E9;
  font-weight: 500;
}

h2 {
  font-weight: 500;
}

.remark-slide-number {
  font-size: 20px;
}

.title-slide .remark-slide-number {
  display: none;
}

.inverse.title-slide {
  background-size: cover;
  color: #EDEEEF;
}

.inverse.title-slide h1  {
  color: #E69F00;
  font-size: 72px;
  text-shadow: none;
  text-align: left;
  vertical-align: bottom;
}
.inverse.title-slide h2  {
  color: #56B4E9;
  text-shadow: none;
  font-size: 48px;
  text-align: left;
  font-weight: bold;
}
.inverse.title-slide h3  {
  color: #EDEEEF;
  text-shadow: none;
  font-size: 36px;
  text-align: left;
  margin-bottom: 10px;
}

.inverse.title-slide h4  {
  color: #EDEEEF;
  text-shadow: none;
  font-size: 24px;
  text-align: left;
  margin-bottom: 10px;
}

.inverse {
  background-size: cover;
  background-color: #23373B;
  color: #EDEEEF;
  font-weight: bold;
  text-shadow: none;
}

.inverse-ns {
  background-size: cover;
  background-color: #23373B;
  color: #EDEEEF;
  text-shadow: none;
  font-weight: bold;
}

.takeaways {
  padding-top: 80px;
}

.inverse h2, .inverse h3  {
  color: #EDEEEF;
  font-weight: 500;
}

.inverse del {
  color: #6C7B7F;
}

.shadow { text-shadow: 0 0 20px rgba(51,51,51,0.35) }

img {
  display: block;
  margin-left: auto;
  margin-right: auto;
}

ul {
  font-size: 48px;
  list-style-type: none;
  text-align: left;
  font-weight: 500;
  padding-top: 40px;
}

ul li {
  padding-bottom: 40px;
}

ol {
  counter-reset: my-counter;
  list-style: none;
  padding-left: 40px;
  font-size: 45px;
  font-weight: bold;
  text-align: left;
}

ol li {
  counter-increment: my-counter;
  padding-left: 40px;
  position: relative;
  font-size: 45px;
  margin: 20px 0;
  display: block;
  margin-block-start: 0.83em;
  margin-block-end: 0.83em;
  margin-inline-start: 0;
  margin-inline-end: 0;
}

ol li::before {
  content: counter(my-counter);
  color: #fff;
  font-size: 40px;
  font-weight: bold;
  position: absolute;
  left: -25px;
  line-height: 50px;
  width: 50px;
  height: 50px;
  top: 0;
  background: #56B4E9;
  border-radius: 50%;
  text-align: center;
}</style>

# My Organization's First R package
## Write R Code
## `rstudio::conf(2020L)`

---

---

# **Writing Functions: Review**

---

## Writing functions

---

---

---

---

## Your Turn 1

#### Re-write this ggplot2 theme as a function. Call it `theme_avalanche_h()`.

#### Run this code to test that your function works

---

## Your Turn 1

```r
*theme_avalanche_h <- function() {
* theme_minimal(base_size = 14) +
*   theme(panel.grid.minor = element_blank(), panel.grid.major.y = element_blank())
*}

residents_per_sector <- 
  data.frame(
    sector = as.factor(1:8),
    residents = c(1000, 2034, 4594, 2304, 8093, 1200, 300, 2398)
  )

ggplot(residents_per_sector, aes(forcats::fct_reorder(sector, residents), residents)) +
  geom_col() + 
  coord_flip() + 
  xlab("sector") + 
  theme_avalanche_h()
```
]
---

---

# **We're going to start writing functions to our package, not the global environment. This will require a change in workflow!**

---

# `use_r("file_name")`

## Write a new file to `R/`

---

---

---

# devtools: loading vs. building

## **`load_all()`**: fast, all functions available
## **`build()`**: builds and installs the package

---

# ~~devtools: loading vs. building~~

## ~~`load_all()`: fast, all functions available~~
## ~~`build()`: builds and installs the package~~

<br>

# *Keyboard shortcuts*
## `load_all()`: **`CMD/CTRL + Shift + L`**
## `build()`: **`CMD/CTRL + Shift + B`**

---

---

## Our new workflow

---

## Your Turn 2

#### Create a new file with `use_r()` called "db_con"
#### Put this function in the file and save it
#### Use `load_all()` to load the package function.
#### Run this code to make sure it works:

---

## Your Turn 2

```r
use_r("db_con")
```

### in `R/db_con.R`:

```r
db_con <- function(dbname = "residents_per_sector") {
  dbname <- match.arg(dbname)
  # We will set up real database connections later. For now, 
  # we'll just return some hard-coded data instead.
  data.frame(
    sector = as.factor(1:8),
    residents = c(1000, 2034, 4594, 2304, 8093, 1200, 300, 2398)
  )
}
```

---

## Your Turn 2

---

# Using other packages

1. Import a package with `use_package()`
2. Use the package with `pkg::fun()`
3. **DO NOT** use `library()`. Avoid it completely while developing.

---

---

---

---

---

---

---

# Testing for suggested packages

```r
if (requireNamespace("shiny", quietly = TRUE)) {
  # ... shiny code
}
```

---

---

---

## Your Turn 3

#### Fix the code in `R/themes.R` to use `ggplot2::` instead of `library(ggplot2)`
#### Run `use_package("ggplot")` to add ggplot2 to `Imports`
#### Re-load the package (`Cmd/Ctrl+Shift+L`) and run this code to make sure it works

---

## Your Turn 3

```r
use_package("ggplot2")
```

### in `R/themes.R`:

```r
theme_avalanche <- function() {
* ggplot2::theme_minimal(base_size = 14) +
*   ggplot2::theme(panel.grid.minor = ggplot2::element_blank())
}
```

---

## Your Turn 3

---

# Packages so good they get their own functions

---

# Packages so good they get their own functions

## `use_tibble()`
## `use_data_table()`
## `use_pipe()`
## `use_rcpp*()`

---

## Your Turn 4

#### We need roxygen2 for this exercise. We'll learn more about it in the next module. For now, just run `use_roxygen_md()`
#### Run `use_tibble()` and `use_data_table()`
#### In `R/get_data.R`, edit the function to be able to return a data table: Add the argument `data_table = FALSE`. If `data_table` is `TRUE`, convert the data frame with `data.table::as.data.table()`
#### Run this code to make sure it works

---

## Your Turn 4

```r
use_roxygen_md()
use_tibble() 
use_data_table()
```

```r
*get_resident_data <- function(data_table = FALSE) {
  residents_per_sector <- db_con("residents_per_sector")

* if (data_table)
*   return(data.table::as.data.table(residents_per_sector))

tibble::as_tibble(residents_per_sector)
}
```

---

## Your Turn 4

```
##    sector residents
## 1:      1      1000
## 2:      2      2034
## 3:      3      4594
## 4:      4      2304
## 5:      5      8093
## 6:      6      1200
## 7:      7       300
## 8:      8      2398
```

```r
data.table::is.data.table(res_data)
```

```
## [1] TRUE
```

---

## Your Turn 4: Stretch goal

#### Run `use_pipe()` to add the magrittr pipe to your package. What changed?

---

## Organizing `.R` files

---

## Organizing `.R` files

## *Less than 1:1, more than all:1*

---

## Organizing `.R` files

## ~~Less than 1:1, more than all:1~~
## *`utils.R`*

---

## Organizing `.R` files

## ~~Less than 1:1, more than all:1~~
## ~~`utils.R`~~
## *`zzz.R`*, *`.onLoad`*

See [R Packages, ed. 2](https://r-pkgs.org/r.html#when-you-do-need-side-effects) for more.