part3.Rmd

---
title: "Part 3"
output:
  xaringan::moon_reader:
    css: [default, rd.css, rd-font.css, "hygge"]
    lib_dir: assets
    seal: false
    nature:
      highlightStyle: dracula
      countIncrementalSlides: false
      beforeInit: "macros.js"
---


```{r setup3, include = FALSE}
knitr::opts_chunk$set(
  comment = "#>",
  collapse = TRUE,
  warning = FALSE,
  message = FALSE,
  dev = "png",
  dpi = 108,
  fig.width = 6,
  fig.height = 4.5,
  fig.align = 'center',
  width = 120
)
options(htmltools.dir.version = FALSE, htmltools.preserve.raw = FALSE)
rfa <- function(...) icons::fontawesome(...)
emj <- function(...) emo::ji(...)
kig <- function(...) knitr::include_graphics(...)
```


---

class: inverse, center, middle

# `r rfa("toolbox")` Towards the gold standard

![:custom_hr]()

## `r rfa("database")` &nbsp; `r rfa("arrow-right")` &nbsp; `r rfa("r-project")` &nbsp; `r rfa("arrow-right")` &nbsp; `r rfa("cogs")` &nbsp; `r rfa("arrow-right")` &nbsp; `r rfa("file")` &nbsp; `r rfa("arrow-right")` &nbsp; `r rfa("sync-alt")`

## `r icons::fontawesome("clock")` ~ 30min


???

A word about setting up those tools depending on the environment.
A word about that all are not needed.

---
# Gold standard

```{R, echo = FALSE, out.width = "80%", purl = FALSE}
kig("img/spectrum.png")
```

.font80[`r rfa("file")` Peng. (2011).
[Reproducible Research in Computational Science.](https://science.sciencemag.org/content/334/6060/1226)]

--

### Towards full reproducibility, technologies are getting more complex


---
# Gold standard

<br>

```{R, echo = FALSE, out.width = "90%", purl = FALSE}
kig("img/pool.png")
```


???

A tool box of software
Coding a lot


---
# Version control `r emj("hand")`

--

### **I've changed something... not sure what... but it no longer works!**

--

<br>

### **Keep track of code changes!**

--

- to talk about the same piece of code

--

- to easily go back in time

--

- to tag code versions


---
# Version control

### Local versioning

```{r, echo = FALSE, out.width = "50%", purl = FALSE}
kig("https://git-scm.com/book/en/v2/images/local.png")
```

`r rfa("book")` Chacon & Straub (2014) [Pro Git](https://git-scm.com/book/en/v2/Getting-Started-About-Version-Control)

---
# Version control

### Centralized version control

```{r, echo = FALSE, out.width = "50%", purl = FALSE}
kig("https://git-scm.com/book/en/v2/images/centralized.png")
```

`r rfa("book")` Chacon & Straub (2014) [Pro Git](https://git-scm.com/book/en/v2/Getting-Started-About-Version-Control)


---
# Version control with Git

### Distributed version control such as Git

```{r, echo = FALSE, out.width = "42%", purl = FALSE}
kig("https://git-scm.com/book/en/v2/images/distributed.png")
```

`r rfa("book")` Chacon & Straub (2014) [Pro Git](https://git-scm.com/book/en/v2/Getting-Started-About-Version-Control)


<br>

- [GitKraken](https://www.gitkraken.com/)
- [GitHub Desktop](https://desktop.github.com/)
- R Studio

---
# Version control with Git + GitHub/Gitlab

```{r, echo = FALSE, out.width = "42%", purl = FALSE}
kig("https://git-scm.com/book/en/v2/images/distributed.png")
```

### e.g. The [`r rfa("github")`](https://github.com/inSilecoInc/workshop_reproducibility) repository for this workshop !


<span class='server'>`r rfa("github")` `r rfa("gitlab")`</span>

???

push and pull
show repo
git log 
git pulll


---
# Use a pipeline toolkit: [GNU Make](https://www.gnu.org/software/make/manual/make.html)

--

### 1. Create a file `makefile` (or `Makefile`) [`r rfa("file")`](https://raw.githubusercontent.com/inSilecoInc/workshop_reproducibility/main/demo/makefile)


--


### 2. Write down what it needs to do (steps of your workflow)

```sh
all: ms code
	
ms: code ms/ms.Rmd 
	Rscript -e 'rmarkdown::render("ms/ms.Rmd")'

code: 
	Rscript -e 'lapply(list.files("R", full.names = TRUE), source); pipeline()'
```

--

### 3. Use `make` command e.g. `make` or `make ms`


--

`r rfa("blog")` Sachin Patil (2018) [What is a Makefile and how does it work?](https://opensource.com/article/18/8/what-how-makefile)


???

Not R-specific


---
# Use a pipeline toolkit: [targets](https://www.gnu.org/software/make/manual/make.html)


.pull-left[
### Successor of [drake](https://github.com/ropensci/drake)

> The targets package is a Make-like pipeline toolkit for Statistics and data science in R.

<br>

`r rfa("book")` [The targets R Package User Manual](https://books.ropensci.org/targets/index.html#motivation)

]

.pull-right[

```{r, echo = FALSE, out.width = "85%", purl = FALSE}
kig("https://docs.ropensci.org/targets/reference/figures/logo.png")
```
]


---
# Use a pipeline toolkit: [targets](https://www.gnu.org/software/make/manual/make.html) 

### 1. Create a `_targets.R` file at the root of your repository [`r rfa("folder")`](https://github.com/inSilecoInc/workshop_reproducibility/raw/main/demo/targets_workflow.zip)

```{R, eval = FALSE}
library(targets)
library(tarchetypes)
rsrc <- lapply(list.files("R", full.names = TRUE), source)
options(tidyverse.quiet = TRUE)
tar_option_set(packages = c("weathercan", "tidyverse"))
list(
  tar_target(data, get_data()),
  tar_target(tab1, create_tab1(data)),
  tar_target(fig1, create_fig1(data)),
  tar_render(report, "ms/ms.Rmd")
)
```

--

### 2. use `tar_make()`




---
# Structure your repo as an R package

## Why?

--

### 1. Because the structure of an R package is very relevant for research compendia

--

### 2. Because R developers created great packages to develop packages!


--

`r rfa("file")` Marwick et al. (2018) [Packaging Data Analytical Work Reproducibly Using R (and Friends)](https://www.tandfonline.com/doi/abs/10.1080/00031305.2017.1375986)


---
# Structure your repo as an R package

## Basic structure of an R package

```sh
.
├── DESCRIPTION
├── doc
├── data
├── LICENSE
├── man
├── NAMESPACE
├── R
├── README.md
├── tests
└── vignettes
```

---
# Structure your repo as an R package

## Basic structure of an R package

```sh
.
├── DESCRIPTION # includes info about package versions
├── doc         # include a website
├── data
├── LICENSE     # always good to specify a license
├── man         # document function
├── NAMESPACE   # no need
├── R
├── README.md
├── tests       # test your code
└── vignettes   # include your ms
```



---
# Make a R package !

## Is it an overkill?

--

### **NO!**

--

- Jim Hester (2018) [You can make a package in 20 minutes](https://rstudio.com/resources/rstudioconf-2018/you-can-make-a-package-in-20-minutes-jim-hester/)

--

- The collection of packages for developers makes every steps easy

--

- This is a very good avenue to go further


???

Yes... it might be a steep learning curve but it's worth it.


---
# Make a R package !

## Benefit from popular development tools

```{r, echo = FALSE, out.width = "42%", purl = FALSE}
kig("https://raw.githubusercontent.com/r-lib/devtools/master/man/figures/logo.svg")
```

`r rfa("github")` [r-lib/devtools](https://github.com/r-lib/devtools/)


---
# Make a R package !

### It structures your workflow

```R
library(devtools)
# load all functions
load_all()
# document functions and edit Namespace
document()
# test functions
test()
# check the package integrity
check()
```

???

shared workflow


---
# Make a R package !

## Benefit from popular development tools

```{r, echo = FALSE, out.width = "42%", purl = FALSE}
kig("https://raw.githubusercontent.com/r-lib/usethis/master/man/figures/logo.png")
```

`r rfa("github")` [r-lib/usethis](https://github.com/r-lib/usethis)



---
# Make a R package !

> usethis is a workflow package: it automates repetitive tasks that arise during project setup and development, both for R packages and non-package projects.

```R
library(usethis)
usethis::use_gpl_license()
usethis::use_citation()
```


---
# Manage your dependencies

## Different R versions, different results!


```R
# R 3.6
df <- data.frame(fact = rep(c("F1", "F2"), each = 5), val = runif(10))
plot(val ~ fact, data = df)   # works
# R 4.0
df <- data.frame(fact = rep(c("F1", "F2"), each = 5), val = runif(10))
plot(val ~ fact, data = df)   # does not work
```


---
# Manage your dependencies

```{r, echo = FALSE, out.width = "50%", purl = FALSE}
kig("https://raw.githubusercontent.com/rstudio/renv/master/man/figures/logo.svg")
```

`r rfa("github")` [rstudio/renv](https://github.com/rstudio/renv)

---
# Manage your dependencies

## Save all dependencies in renv.lock and install them later seamlessly

```R
renv::init()
renv::snapshot()
# later (on another device)
renv::restore()
```

`r rfa("file")` [Introduction to renv](https://rstudio.github.io/renv/articles/renv.html)




---
# Continuous Integration (CI)

## Check your code on a freshly installed machine

--

### e.g. The [`r rfa("github")`](https://github.com/inSilecoInc/workshop_reproducibility/actions) repository for this workshop!

--

<br>

## Again, with an R package, everything goes smoothly

```R
usethis::use_github_action_check_standard()
```

???

start thinking about OS agnosticism


---
# Use a container with [Docker `r rfa("docker")`](https://www.docker.com/)

> Containers are a standardized unit of software that allows developers to
> isolate their app from its environment, solving the “it works on my machine”
> headache. 

`r rfa("link")` https://www.docker.com/why-docker


---
# Use a container with [Docker `r rfa("docker")`](https://www.docker.com/)

```{r, echo = FALSE, out.width = "45%", purl = FALSE}
kig("img/contain1.png")
```

---
# Use a container with [Docker `r rfa("docker")`](https://www.docker.com/)

```{r, echo = FALSE, out.width = "45%", purl = FALSE}
kig("img/contain2.png")
```


---
# Use a container with [Docker `r rfa("docker")`](https://www.docker.com/)

```{r, echo = FALSE, out.width = "45%", purl = FALSE}
kig("img/contain3.png")
```


---
# Use a container with [Docker `r rfa("docker")`](https://www.docker.com/)

```{r, echo = FALSE, out.width = "45%", purl = FALSE}
kig("img/contain4.png")
```


---
# Use a container with [Docker `r rfa("docker")`](https://www.docker.com/)

```{r, echo = FALSE, out.width = "45%", purl = FALSE}
kig("img/contain5.png")
```


---
# Use a container with [Docker `r rfa("docker")`](https://www.docker.com/)

```{r, echo = FALSE, out.width = "45%", purl = FALSE}
kig("img/contain6.png")
```


---
# Use a container with [Docker `r rfa("docker")`](https://www.docker.com/)


###  docker allows you 

- to create images (via dockerfiles)

- to use images!

<br>

--

`r rfa("github")` [rocker-org/rocker](https://github.com/rocker-org/rocker)

--

```sh
$ docker run --rm -ti rocker/r-base
```

--

Run your code directly in the container!




---
# Holepunch

> `holepunch` will read the contents of your R project on GitHub, create a DESCRIPTION file with all dependencies, write a Dockerfile, add a badge to your README, and build a Docker image. Once these 4 steps are complete, any reader can click the badge and within minutes, be dropped into a free, live, RStudio server.

`r rfa("github")` https://github.com/karthik/holepunch

--

### - Uses [Binder](https://mybinder.org/) (GitHub repo ==> Docker image) + RStudio server.

--


### - Example: `r rfa("github")` [HomogenFishOntario](https://github.com/McCannLab/HomogenFishOntario)!



---
# Archive your code/data with [Zenodo](https://zenodo.org/)


```{r, echo = FALSE, out.width = "80%", purl = FALSE}
kig("img/zenodo.png")
```