README.md

In datr-studio/bladerunr: bladerunr: incisive model runs across a grid of parameters

bladerunr

Overview

The goal of bladerunr is to make it easy to run long tests over a grid of search parameters. bladerunr aims to solve the following common challenges that occur with optimisation and simulation problems:

• Set up a grid of all unique combinations of parameters.
• Easily filter grid using tidyverse-style conditions.
• Run tests with built-in error-handling so that can document errors but not fail the run.
• Set time limits for runs to catch stalling or poor parameter combinations.

bladerunr provides you with a base platform you can trust to manage your testing process. That frees you up to focus on the test code itself.

Installation

You can install the development version of bladerunr from GitHub with:

# install.packages("devtools")
devtools::install_github("firthj/bladerunr")

Usage

All you need to run your tests are just three functions:

blade_setup

First, call blade_setup() to set the overall configuration of your test run.

blade_setup(
  run_name = "my_model_run",
  runr = function(...) {
    # Do cool test stuff
  }
)

There are only two required arguments: a run_name that’s memorable and a runr, which is your test function, the code you’d like to execute at each iteration.

There are also several more options that you can use to tweak your test run, including error logging, time outs, and number of attempts for each run. See the vignette “Running Tests with Bladerunnr” for more info.

blade_params

Next, we’ll also want to define a search grid.

grid <- blade_params(
  list(
    a = seq(1, 3, 0.5),
    b = 1:3,
    c = 3:1
  ),
  a > b,
  c == a * b
)

The first argument is a named list of parameters you want to combine. You may also specify any conditions about the relationships between these using tidyverse-style syntax.

blade_runr

Once your setup is complete, you’re ready for the easy part: running tests. Simply call blade_runr() with your test grid and it’ll take care of the rest.

# Not run
blade_runr(grid)

Getting help

If you encounter a clear bug, please file an issue with a minimal reproducible example on GitHub.

datr-studio/bladerunr documentation built on April 12, 2022, 6:19 p.m.