README.md

In personalscience/taster: Personal Science CGM Analysis Web App

Taster

This is a Golem-based implementation of an R Shiny app to view, explore, and analyze CGM glucose results obtained from Tastermonial users.

Requirements

Develop the app as a package in RStudio. Like other Golem-based packages, by convention the dev directory is reserved for non-package-related setup and miscellaneous.

Docker

You should be able to run the app from the Dockerfile, which will take care of all the software versions and setup. You'll need to connect to a database (see below), the credentials of which are kept in the local environment variables.

Make a Docker version using the Dockerfile.

I copy all the files from my local machine to the remote machine like this:

scp -r * myname@remoteserver:/path/to/remote/dir

Assuming Docker is already available on the remote machine, build the container:

docker build -t tastermonial .

Keep the database credentials as system environment variables in a file .env that looks like this:

TASTER_HOST=12.34.56.78
TASTER_USER=postgres
TASTER_PASSWORD=<your password>
TASTER_DBNAME=<db name, usually qsdb>
TASTER_PORT=<port number>

Run the docker container while loading the environment variables

docker run -p 8085:80 --env-file .env --name taster -d tastermonial

Or use Docker Compose

version: "3"
services:
  web:
    build: .
    ports:
      - "8086:80"
      env_file:
      - taster.env
      volumes:
      - ./config:/inst

Assuming you have the above .env file in taster.env, you can run like this

docker-compose up -d

and the app will appear on port 8086 of your host.

Database

The app requires a database. For testing purposes, a pre-prepared Sqlite database will work fine, though it is currently not included in this repo.

Pre-loading and other Tastermonial-specific database features are in the separate package tasterdb.

Testing

Run unit tests with devtools::test() (or type CMD-SHIFT-T)

Deploy

See the script dev/psi_deploy.R for specific deployment prerequisites.

I'm able to load and host the web app on DigitalOcean by simply loading this Github repo without changes directly as a DigitalOcean app. You will of course need to set environment variables in order to run the database.

personalscience/taster documentation built on Feb. 5, 2022, 9:27 p.m.