Skip to content

Creating functions

Overview

Creating a function for use in FaaSr entails the following steps:

  • Select a GitHub repository to store your function code; we'll use MyGitHubAccount and MyFunctionRepo as a GitHub account name and repository name, respectively, in examples
  • Develop the code for your function. A best practice is to have one file per function; we will use compute_sum.R and compute_sum.py as examples
  • Add FaaSr API calls where appropriate, e.g. faasr_get_file() to get an input file from an S3 data server, faasr_put_file() to put an output file to an S3 data server, faasr_log() to write a message to the log. Refer to the FaaSr R APIs and FaaSr Python APIs documents for a complete list
  • Add the function to a workflow. This is done using the FaaSr Workflow Builder Web UI by clicking on an Action in the workflow DAG

Example

Let's say you develop a function compute_sum.R (e.g. the one used in the FaaSr tutorial) as follows:

compute_sum <- function(folder, input1, input2, output) {

  # FaaSr API calls to get inputs from S3 (two CSV files)
  faasr_get_file(remote_folder=folder, remote_file=input1, local_file="input1.csv")
  faasr_get_file(remote_folder=folder, remote_file=input2, local_file="input2.csv")

  # Function's main implementation (compute a sum and write the output)
  frame_input1 <- read.table("input1.csv", sep=",", header=T)
  frame_input2 <- read.table("input2.csv", sep=",", header=T)
  frame_output <- frame_input1 + frame_input2
  write.table(frame_output, file="output.csv", sep=",", row.names=F, col.names=T)

  # FaaSr API call to put the output file in the S3 bucket
  faasr_put_file(local_file="output.csv", remote_folder=folder, remote_file=output)

  # Log a message
  log_msg <- paste0('Function compute_sum finished; output written to ', folder, '/', output, ' in default S3 bucket')
  faasr_log(log_msg)
}

Say you commit compute_sum.R to repository MyGitHubAccount/MyFunctionRepo.

To use this function in a workflow, in the FaaSr Workflow Builder Web UI proceed as follows:

alt text

  • Create an Action (e.g. compute_sum)
  • Select it to edit using the left pane
  • Under Function Name, enter compute_sum; this is the name of the function declared in the code above
  • Under Compute Server, select your compute server from the drop-down menu (e.g. GH for GitHub Actions)
  • Under Arguments, enter names and values of the arguments matching those used by the function: folder, input1, input2, output
  • Under Function's Git Repo/Path, enter MyGitHubAccount/MyFunctionRepo

Important notes

  • If you provide a GitHub repo (e.g. MyGitHubAccount/MyFunctionRepo) Function's Git Repo/Path, FaaSr will clone and source all source code files from it; e.g. if your repository has files compute_sum.R, compute_mult.R, etc, each will be sourced
  • You can, alternatively, provide a path to a file in a GitHub repo, e.g. MyGitHubAccount/MyFunctionRepo/compute_sum.R; this will only fetch and source one function