Convenient Snapshot Testing Functions for Packages.

snapr

The goal of snapr is to provide convenient snapshot testing functions for R packages:

• expect_snapshot_data() - for data.frames (adapted with permission from the {ssdtools} package)
• expect_snapshot_object() - generalizes expect_snapshot_data() for use with any R object

Installation

You can install the released version of snapr from CRAN with:

install.packages("snapr")

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

# install.packages("pak")
pak::pak("d-morrison/snapr")

Optional: enhanced RDS diffs with diffviewer

snapr works best with the d-morrison/diffviewer fork of diffviewer, which wraps diffobj to provide rich, visual diffs when reviewing RDS snapshots via testthat::snapshot_review(). Install it with:

pak::pak("d-morrison/diffviewer")

Examples

Snapshot data.frames

library(snapr)

# In a testthat test file:
test_that("iris data is correct", {
  expect_snapshot_data(iris[1:5, ], name = "iris_sample")
})

Snapshot any R object

library(snapr)

# Snapshot a list (RDS format by default - works for any R object)
test_that("config object is correct", {
  config <- list(
    name = "my_app",
    version = "1.0.0",
    settings = list(debug = TRUE, timeout = 30)
  )
  expect_snapshot_object(config, name = "config")
})

# Snapshot with JSON format for human-readable diffs
test_that("simple data is correct", {
  data <- list(x = 1:5, y = letters[1:5])
  expect_snapshot_object(data, name = "simple_data", writer = save_json)
})

# Snapshot a model (use RDS for complex objects)
test_that("model structure is correct", {
  model <- lm(mpg ~ wt, data = mtcars)
  expect_snapshot_object(model, name = "model")
})