03_functions.Rmd

---
title: "Functions"
subtitle: "Part 3 of the RStudio Tips and Tricks Series"
author: "Jeremy Allen"
date: "Nov 2, 2021, updated on `r Sys.Date()`"
output:
  html_document:
    highlight: zenburn
    theme:
      bg: "#303436"
      fg: "#9da9af"
      primary: "#ff8080"
      base_font:
        google: "Roboto"
      code_font:
        google: "JetBrains Mono"
editor_options: 
  chunk_output_type: console
self_contained: true
---
---

<br>  
<br>  

```{r packages, message=FALSE, warning=FALSE, include=FALSE}

library(palmerpenguins)
library(tidyverse)
library(janitor)
library(here)

```

# Review

- we made a new function - clean_if_bad_names()
- shift+cmd+p for command palette
- fun/if/for/lib/ec then shift+tab to add snippets
- esquisse addin to generate plot code if new to ggplot2
- alt+dash for assignment arrow 
- cmd+up in console to see history

# Functions - let's inspect!

```{r }

clean_if_bad_names <- function(x) {
  if (any(str_detect(names(x), "[\\sA-Z()/-]"))) {
    x <- clean_names(x)
    x
  }
  x
}

# reset df
df <- penguins_raw

clean_if_bad_names(x = df)

```

<br>

# But we don't really need that silly function

clean_names() from the janitor package looks for more bad elements in names\
How do we know?\
`r emo::ji("smile")` Pro tip: cmd+click or F2 on a function name to inspect it's body

```{r echo=TRUE, eval=FALSE}

janitor::clean_names()

```


```{r}

# we don't really need to test first, just clean the column names
df <- penguins_raw %>% janitor::clean_names()

```

![](images/inspect1.png)\

Select the data.frame method from the drop down ![](images/inspect2.png)

Then cmd click the internal function make_clean_names()\
![](images/inspect3.png)

Now we see the internals of janitor's make_clean_names(), we see it cleans up apostrophes and other things our function didn't test for.\

![](images/inspect4.png)

<br>

# What other functions do you want to inspect?

```{r echo=TRUE, eval=FALSE}

dplyr::arrange()

```

<br>

## Water cooler chat: Chunk options in gear icon

In the chunk above, the options in the gear icon do not provide a way for:\
`{r echo=TRUE, eval=FALSE}`\
which means to show the code but do not run the code.

<br>

## Function and snippet mashup
`r emo::ji("smile")` combine some functions into one function using purrr::compose()  

```{r}

not_in <- purrr::compose( # by default it runs bottom to top
  `!`,
  `%in%`
)

```


```{r}

not_in(3, c(1:2))
not_in(1, c(1:2))
not_in("January", month.name)

```

```{r compose, message=FALSE, warning=FALSE}
library(broom)
library(gt)

lm_peek <- purrr::compose(
  .dir = "forward",
  lm,
  broom::tidy,
  gt::gt,
  ~tab_header(.x, title = "Results Preview")
)

# this
lm_peek(culmen_depth_mm ~ flipper_length_mm + body_mass_g, data = df)

```

<br>  

## Got code?
`r emo::ji("smile")` Pro tip: highlight code, then in Code menu, select extract function
```{r echo=TRUE, eval=FALSE}

# in the code menu, select extract function

df %>% 
    filter(Sex == "MALE") %>% 
    summarise(mean_body_mass = mean(`Body Mass (g)`, na.rm = TRUE))

```