Testing FaaSr locally on your desktop

Overview

While FaaSr is designed to allow you to execute workflows and access data in the cloud, in an automated and unmanaged way, it is useful to test workflow executions locally so you can verify desired functionality and debug errors. The FaaSr provides a package that allows you to run a workflow locally in your desktop, in one of two ways:

• Native execution, where the functions run in your native R/Rstudio environment in your desktop. This is the default.

• Docker execution, where the functions run in Docker containers in your desktop

In both native and Docker execution, both the data used by your workflows and the R functions are stored in folders/files in your computer

Setting up folders and code

You need to create, in your working directory, a folder named faasr_data (if it doesn’t already exist). Inside this folder, you store the payload and credentials file (payload.json and faasr_env in the example below), and you need to create another folder named R. You then need to copy your R functions to this folder. So, in your working directory, you should have a folder structure like this:

payload.json
faasr_env
faasr_data/R/sum.R
            /mult.R
            /div.R

You should also copy any input file(s) needed by your workflow within the faasr_data folder. This folder is essentially akin to an S3 bucket in your computer, and should be configured as such.

Preparing for execution

Just as in the companion vignette for single function, you begin by using the faasr() main function to load a payload.json configuration and faasr_env credential file into an R list. Note that, for local execution, the credentials in the credential file are ignored. We will use faasr_example as the name of the list:

faasr_example <- faasr(json_path="payload.json", env="faasr_env")

Execute the workflow locally

The process is similar to the one described in the companion vignette for single function, but instead of using $register_workflow and $invoke_workflow for cloud registration/execution, you use $faasr_test, as shown below:

faasr_example$faasr_test()

If you prefer to execute the workflow using Docker, you need to provide, as an argument to $faasr_test, a list with the following elements (note: “latest” and “faasr/local-test” are the default values for the container name and version to use; you may specify your own container if needed):

use_docker = list(use=TRUE, version="latest", image="faasr/local-test")
faasr_example$faasr_test(use_docker)

Checking the outputs

Your outputs will be stored locally inside the faasr_data folder you created.