Making your R-based research project portable

.title[
# Making your R-based research project portable
]
.subtitle[
## Introduction to the {renv} package
]
.author[
### Ernest Guevarra
]
.date[
### 2024-03-06
]

---

# Outline

* All about environments

* System dependencies management

* Project-local R dependencies management

---

background-color: #FFFFFF

# All about environments

.pull-right[
## **Portability** of an R workflow will depend on **reproducibility of its related environments**
]

---

# No one cares what operating system you run as long as it stays up.

## Bruce Perens

#### created the defintion of Open Source and wrote the first manifesto of Open Source

---

# Different systems, different requirements

<img src="images/Windows_logo_-_2021.svg" width="33%" /><img src="images/Apple_logo_black.svg" width="33%" /><img src="images/Tux.svg" width="33%" />
]

.pull-right[
* each operating system (and each of its versions) may/will require specific dependencies in order to install R

* each operating system (and each of its versions) may/will require specific dependencies in order to install some R packages 
]

---

# System dependencies management

.pull-right[
* solutions/approaches will depend on your use case but the most universal step is **documentation**

* **Supervisor-supervisee** or **small research team** settings - establish compatibility guidelines between members and document system requirements for each project;

* **Medium to largish research team** settings
    - use turn-key solutions that standardise R setup using cloud-based or software-as-a-service (SaaS) solutions (e.g., RStudio Cloud); or,
    - setup machines used to be consistent with team's R workflow system requirements through containerisation (e.g., [Docker](https://www.docker.com/))
]
    
---

# Using containers via Docker

* **Containers** are standardised units of packaged software that have everything the software needs to run including libraries, system tools, code, and runtime.

* **Containers** can be deployed either onto remote machines or onto a local machine using [Docker](https://www.docker.com)

* [Rocker](https://www.rocker-project.org/) is a community-organised and community-maintained hub of Docker containers that are pre-built and specified with different variants and configurations of R for various operating systems and for various types of R workflows e.g., base, RStudio, spatial analysis, machine learning, etc.

* Containers support portability because we are able to "carry" with us almost any operating system + R configuration that we might need in our R workflows

---

# Project-local R dependencies management

.pull-right[
* The more people collaborate on code and R workflows, the higher the chances that R package dependencies will increase

* The more complex the type of R analysis workflow that is being implemented, the higher the chances that R package dependencies will increase

* Management of R dependencies will be critical in ensuring portability

* The `{renv}` package facilitates this R dependencies management
]

---

# The `{renv}` package

* Work in the project as normal, installing and removing new R packages as they are needed in the project;

* Save the state of the project library to the lockfile;

* Continue working on your project, installing and updating R packages as needed; and,

* Again save the state of your project library if your attempts to update R packages were successful, or restore to a previous state as encoded in the lockfile if you encounter problems updating a package.
]

---

# Questions?

---

# Thank you!

Slides can be viewed at https://oxford-ihtm.io/open-reproducible-science/session10.html

PDF version of slides can be downloaded at https://oxford-ihtm.io/open-reproducible-science/pdf/session10-portable-r-projects.pdf

R scripts for slides available [here](https://github.com/OxfordIHTM/open-reproducible-science/blob/main/session10.Rmd)