In allanecology/multiFunBetadivLuiGDM: Testing effects of multitrophic beta-diversity and land-use intensity (LUI) on ecosystem multifunctionality

installing this package

as a user

This is the way to go if you want to reproduce the work.

• Where to install the package? (In which folder?) The scripts you need are stored within the package. Therefore store it where you want to have the code. In this tutorial, the place to store is "/home/noelleschenk/Desktop/BetaDivMultifun/"

If you do not have a ".tar.gz" file

Or if you don't know what this is. 1. Download the whole folder from github. - If you use git within RStudio : Download from within RStudio: Select (at the top right corner) new project > Version control > Git and in the field "Repository URL" give the link https://github.com/allanecology/BetaDivMultifun . chose a location where to store it. - If you don't use git : manually download and place in a folder of your choice. 2. From within R, set the R working directory to the parent folder of the downloaded package. (e.g. Downloads if you stored the package folder in Downloads.)

1. Run the code below

# install.packages("devtools")
setwd("<parent-directory-of-the-package>")
devtools::install("BetaDivMultifun")

Note for above code section: - The devtools::install function also installs packages on which BetaDivMultifun depends on. - Update all packages when you are asked for by R (type 1 to the console)

TODO Noelle write about package error in windows.

Set wd back to the package

setwd("/home/noelleschenk/Desktop/BetaDivMultifun/")

If you do not with to further develop the package, but e.g. just reproduce the analysis, you are done now.

If you do have a ".tar.gz" file

If you have a ".tar.gz" file with the BetaDivMultifun package on it, you can install directly from that file. The ".tar.gz" file is called a "source"-type file.

install.packages("/home/noelleschenk/Desktop/BetaDivMultifun_0.1.tar.gz", type = "source", repos = NULL)

If you do not with to further develop the package, but e.g. just reproduce the analysis, you are done now.

install the package as a developer

This is the way to go if you want to further develop this analysis.

Literature

To create the package, I used the blogposts of: Hilary Parker, the book "R packages" and the idea of using R packages for an analysis is coming from Robert M. Flight's blogposts, where he explains why and how to use R package for an analysis.

Setup

Only if you like to work on the package (change it). I worked with following packages and RStudio, which is extremely well suited for package development.

If you are linux user, install the following packages with sudo apt-get install <packagename> - libssl-dev - libxml2 libxml2-dev - libcurl4-gnutls-dev (previously libcurl)

If you are not using Linux and there was any requirement, R will tell you in its error message as soon as you try to install the below mentioned packages.

# running these lines will take some time ...
install.packages("devtools")
library(devtools)
#devtools::install_github("klutometis/roxygen")
# the roxygen2 source possibly changed now
devtools::install_github("r-lib/roxygen2")
library(roxygen2)

Maintaining the package

changing code

After you generated/changed code of the package, the newly generated code needs to be loaded. (E.g. if you add a function and later want to use it.) Therefore, the function code as well as the functions documentation needs to be loaded as described below.

• create documentation from roxygen comments, with Ctrl-Shift-D or

devtools::document()

• load functions which are encoded in the package. devtools::load_all() or with Ctrl-Shift-L (while the cursor is placed in the console)

• install and restart the package : Ctrl-Shift-B (console R CMD INSTALL --no-multiarch --with-keep.source BetaDivMultifun)

• to rebuild the source package .tar.gz file newly : RStudio : Build > Build Source Package devtools::build()

check code

RStudio : Build > Check --> note : don't confuse with "test", this is something else.

test code

develop your own tests to prevent introducing bugs.

knit to html

By hand

rmarkdown::render('<pathtothisdocument>/1read_raw_dataset.Rmd')
rmarkdown::render('vignettes/function_imputation.Rmd', output_format = "html_document")

Note : knit will produce a .html file in a temporal directory (e.g. tmp/) to not clutter the "vignettes" folder with html files. To get a html file output, use the code above.

Update datasets

Updating the included datasets

Example : change dataset "model_names" :

# data("model_names") # loads the dataset

# IF change is done by hand
fwrite(model_names, "deleteme.csv") # write dataset to .csv in order to change by hand
model_names <- fread("deleteme.csv")

# IF change is done by hand
# just change the variable model_names

# overwrite dataset in /data folder with new version
save(model_names, "data/model_names.rda")
usethis::use_data(model_names, overwrite = T)
# document changes in the file R/data.R
# create new documentation
devtools::document()

# clean up
# in case you wrote to a .csv file, clean this up.

allanecology/multiFunBetadivLuiGDM documentation built on Nov. 12, 2023, 6:16 a.m.