shiny-cromwell

Shiny app for interacting with the Fred Hutch instances of Cromwell.

The app lives at a URL only accessible to Fred Hutch folks.

This app can also be run on your local machine for any reason, e.g., if you notice memory limitations in the deployed version or you want to run it at another institution.

Running the app

Without Docker

You can run this Shiny app locally. First you'll need to install required packages:

(note: run make pkg_deps_cmd to update the below code block)

pak::pak(c("getwilds/[email protected]", "getwilds/[email protected]", "dplyr", "DT", "ggplot2", "glue", "httr", "jsonlite", "lubridate", "magrittr", "memoise", "purrr", "rclipboard", "RColorBrewer", "rlang", "shiny", "shinyBS", "shinydashboard", "shinydashboardPlus", "shinyFeedback", "shinyjs", "shinylogs", "shinyvalidate", "shinyWidgets", "testthat", "tibble", "uuid"))

And the above yourself in R.

Or you can run pkg_deps_install which should find and install required packages all in one step.

After installing required packages run the app. There's a few different ways to do that:

1. If you are familiar with using make commands, run make run
2. If you don't or can't use make, you can run it on the command line like Rscript -e 'shiny::runApp("app", launch.browser = TRUE)'
3. Inside of R you can run shiny::runApp("app", launch.browser = TRUE)'

All of these should open the Shiny app in your default browser. If that does not happen get in touch!

With Docker

The make command make run_docker will attempt to run the app with Docker. You'll need Docker installed and running for this to work.

Note that CTRL+C doesn't work to kill the container after running make run_docker. Yo have to use a separate terminal window or Docker Desktop to kill the container (basically the docker process doesn't receive the signal to kill its process).

Also note that docker commands can be specific to the host operating system, so if you run into errors you may need to modify the docker commands.

Run a feature branch locally

The make command make run_branch will run a remote branch locally on your machine. Whereas make run_docker runs the current branch you're on, make run_branch pulls down a docker image of the remote branch specified and runs that.

Make sure to run docker login nexus-registry.fredhutch.org before you try this make command. You only need to run this once. Contact scicomp to get the username and password for this step.

A full example: make branch=inputs-viewer run_branch

Note that the branch specification has to be before the target name (run_branch).

You must be on FH campus or be on a VPN to use this make command.

Debugging

• Use print statements as needed ...
• Instead of just make run, one can run the app with SHINY_LOG=true make run which activates the package shinylogs

App structure

├── app
│   ├── about.md
│   ├── buttons.R
│   ├── iputs_utils.R
│   ├── modals.R
│   ├── proof.R
│   ├── server.R
│   ├── sidebar.R
│   ├── tab-servers.R
│   ├── tab-submission.R
│   ├── tab-tracking.R
│   ├── tab-troubleshoot.R
│   ├── tab-validate.R
│   ├── tab-welcome.R
│   ├── ui.R
│   ├── ui_components.R
│   ├── utils.R
│   ├── validators.R
│   └── www
│       ├── favicon.ico
│       ├── fred-hutch.svg
│       └── js
│           └── keyup.js

Notes about the structure:

• the main app files are at app/server.R and app/ui.R
• files in www/ are automatically sourced
• code in other files in app/ - e.g., app/proof.R - are only available if pulled in with source()

We will continue to give the app more structure, possibly including Shiny modules and maybe as an R package.

License

MIT