LTLA / ActionsTesting

Testing for GitHub Actions

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

iSEE screenshots

Overview

This repository contains code to automatically generate screenshots of iSEE instances, using the appshot() function from the webshot package to generate screenshots in an automated manner. This provides a location from which other packages can access these screenshots (e.g., during vignette compilation).

Organization

We organize the repository into these subdirectories:

  • data/, a directory of R scripts to generate a SummarizedExperiment object.
  • apps/, a directory of R scripts to create an iSEE instance.
  • images/, a directory of screenshots for each iSEE instance.

Note that images/ is only populated with PNGs in the compiled branch of this repository. This allows us to delete and regenerate images at any point without swamping the master with large binaries.

Instructions

To avoid constructing the same SummarizedExperiment multiple times for different iSEE instances, we centralize data processing into a common data/ directory. Here, each script is expected to generate an RDS file containing a SummarizedExperiment object. The RDS file should contain the same prefix as the script, e.g., allen.R should generate allen.rds.

Each script in data/ should load an SE object from an RDS file and generate a Shiny app object named app. We suggest using rprojroot to access the RDS file, to ensure that the file can be run from anywhere inside the repository. The script may also generate delay, a numeric scalar specifying the number of seconds to wait before taking the screenshot. Longer delays may be necessary to ensure that more complex plots are properly built.

The images/compile.R file will execute each script in apps/ and take the screenshot with the specified delay. A PNG file is produced with the same prefix as its originating source file. Note that it is necessary to run webshot::install_phantomjs() before running this script. You should also change the branch to compiled before generating and committing new images.

Contribute

Contributors should add files to data/ and apps/ as necessary; we will take care of recompiling them.

About

Testing for GitHub Actions


Languages

Language:R 65.1%Language:Dockerfile 20.2%Language:Shell 14.7%