Skip to content

Home

Jump to bottom
Alex Lemm edited this page Jan 31, 2020 · 8 revisions

trendminer developer Wiki

Local development setup

The steps below describe how you set up your local development environment in to order to continue developing the trendminer R package:

  • Start with installing R, RStudio and Git on your machine

  • Launch R Studio and install the following R packages via the console:

    • devtools: makes development easier by providing R functions that simplify common tasks
    • remotes: Downloads and installs R packages stored in 'GitHub', 'BitBucket'
    install.packages(c("devtools", "remotes"))
  • Next, we need to install the remaining packages the trendminer package depends on AND packages which are needed for the developing itself like, e.g., testthat for creating unit tests. Just execute the following three commands in the console:
    # Corresponds to the 'Imports' statement in the package DESCRIPTION
    install.packages(c("httr", "jsonlite", "magrittr", "dplyr", "rlang", "lubridate"))

    # Corresponds to the 'Suggests' statement in the package DESCRIPTION
    install.packages(c("testthat", "spelling", "knitr", "rmarkdown", "covr", "data.tree", "networkD3", "purrr", "ggplot2"))

    install.packages("usethis")
  • Now, we need to git clone the trendminer repo from GitHub by executing the the following command (in case you git clone via HTTPS) now outside of your R console:
    git clone https://github.com/TrendMiner/trendminer.git
  • Before we actually take a look at the development resources we just downloaded, we set the necessary R environment variables. Go ahead and execute the following command in your R console which will let you edit your .Renviron file of your home directory:
usethis::edit_r_environ()
  • Add the following environment 6 variables to .Renviron and save the file. The first 5 variables correspond to those also a standard user of the R trendminer package needs to set.
TM_client_ID = YOUR_CLIENT_ID_HERE
TM_client_secret = YOUR_CLIENT_SECRET_HERE
TM_usr = YOUR_USER_NAME_HERE
TM_pwd = YOUR_USER_PASSWORD_HERE
TM_base_url = YOUR_TM_BASE_URL_HERE

NOT_CRAN = TRUE

  • Close RStudio (new environment variables are only available after a re-launch of the R session)

  • We navigate to the repo folder we cloned from GitHub and click on the trendminer.Rproj file which will launch RStudio and the R package project

  • Optional: For a quick test to check your setup you could execute the unit tests against the TrendMiner instance you specified in the environment variables above by pressing 'CTRL + SHIFT + T' (Windows) or 'CMD + SHIFT + T' (Mac)

You can now continue developing :-). Have fun!

Vignette building

Vignettes are not part of R packages which are downloaded from GitHub. Based on the fact that trendminer is an API package which relies on specific secrets and a specific test environment, vignettes cannot be built by the user while downloading the package.

In order to make the vignettes available via GitHub downloads, the extra steps below need to be conduced on the development side when you update the vignettes:

One time steps:

  • delete inst/doc from .gitignore

Re-current procedure after every vignette update:

  • Execute the following commands:

      devtools::build_vignettes()

  dir.create("inst/doc")
  file.copy("./doc", "./inst", recursive = TRUE)

  • git commit the new/updated files in inst/doc

  • git push

Shortly before the first CRAN release inst/doc and its content need to be deleted and the change in .gitignore should be reverted as well. Those changes should, of course, also be committed.

Publishing to CRAN will ensure that the vignettes will be published as well and we can ensure that they are included in trendminer CRAN package downloads. Downloading dev versions from the master or other branches on GitHub won't come with the vignettes at that point in time anymore but that is by intention.

Clone this wiki locally