
Fast pixel-by-pixel image comparison for R, powered by odiff.
batch_report()expect_images_match() and
expect_images_differ() for visual regression testingOdiffr requires the Odiff binary to be installed on your system:
# npm (cross-platform, recommended)
npm install -g odiff-bin
# Or download binaries from GitHub releases
# https://github.com/dmtrKovalenko/odiff/releases# Install from CRAN (when available)
install.packages("odiffr")
# Or install the development version from GitHub
# install.packages("pak")
pak::pak("BenWolst/odiffr")If you cannot install Odiff system-wide:
odiffr::odiffr_update() # Downloads to user cachelibrary(odiffr)
# Compare two images
result <- compare_images("baseline.png", "current.png")
result$match
#> [1] FALSE
result$diff_percentage
#> [1] 2.45
# Generate a diff image
result <- compare_images("baseline.png", "current.png", diff_output = "diff.png")# High-level API (returns tibble if available)
result <- compare_images("img1.png", "img2.png")
# Low-level API (returns detailed list)
result <- odiff_run("img1.png", "img2.png")# Adjust sensitivity threshold (0-1, lower = more precise)
result <- compare_images("img1.png", "img2.png", threshold = 0.05)
# Ignore antialiased pixels
result <- compare_images("img1.png", "img2.png", antialiasing = TRUE)
# Fail immediately if dimensions differ
result <- compare_images("img1.png", "img2.png", fail_on_layout = TRUE)# Ignore specific areas (e.g., timestamps, dynamic content)
result <- compare_images("img1.png", "img2.png",
ignore_regions = list(
ignore_region(0, 0, 200, 50), # Header
ignore_region(0, 500, 800, 600) # Footer
)
)# Compare multiple image pairs
pairs <- data.frame(
img1 = c("baseline/page1.png", "baseline/page2.png"),
img2 = c("current/page1.png", "current/page2.png")
)
results <- compare_images_batch(pairs, diff_dir = "diffs/")
# Extract failures or passes
failed_pairs(results)
passed_pairs(results)
# Compare entire directories
results <- compare_image_dirs("baseline/", "current/", recursive = TRUE)
# Get summary statistics
summary(results)
#> odiffr batch comparison: 50 pairs
#> Passed: 42 (84.0%)
#> Failed: 8 (16.0%)
# Use parallel processing (Unix only)
results <- compare_images_batch(pairs, parallel = TRUE)# One-liner: compare directories and generate HTML report
compare_dirs_report("baseline/", "current/")
# -> Creates diffs/ directory with diff images and report.html
# Or step-by-step for more control
results <- compare_image_dirs("baseline/", "current/", diff_dir = "diffs/")
batch_report(results, output_file = "qa-report.html")
# Self-contained report with embedded images
batch_report(results, output_file = "qa-report.html", embed = TRUE)
# Portable report with relative image paths
batch_report(results, output_file = "output/report.html", relative_paths = TRUE)Run visual regression tests in GitHub Actions and upload diff artifacts:
# .github/workflows/visual-regression.yaml
name: Visual Regression
on: [push, pull_request]
jobs:
visual-test:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: r-lib/actions/setup-r@v2
- name: Install dependencies
run: |
install.packages(c("odiffr", "webshot2"))
odiffr::odiffr_update()
shell: Rscript {0}
- name: Generate screenshots
run: Rscript scripts/generate-screenshots.R
- name: Compare images
run: |
library(odiffr)
results <- compare_dirs_report("baseline/", "current/")
if (any(!results$match)) stop("Visual regression detected!")
shell: Rscript {0}
- name: Upload diffs
if: failure()
uses: actions/upload-artifact@v4
with:
name: visual-diffs
path: diffs/library(magick)
# Compare magick-image objects directly
img1 <- image_read("baseline.png") |> image_resize("800x600")
img2 <- image_read("current.png") |> image_resize("800x600")
result <- compare_images(img1, img2)library(testthat)
library(odiffr)
test_that("dashboard renders correctly", {
expect_images_match(
"screenshots/current.png",
"screenshots/baseline.png",
threshold = 0.1
)
})
test_that("button changes on hover", {
expect_images_differ(
"button_normal.png",
"button_hover.png"
)
})On failure, diff images are automatically saved to
tests/testthat/_odiffr/.
# Check if Odiff is available
odiff_available()
# Get version and configuration info
odiff_info()
# Update to latest version (downloads to user cache)
odiffr_update()
# Use a specific binary
options(odiffr.path = "/path/to/odiff")options(odiffr.path = "...") - User overrideSys.which("odiff"))odiffr_update()| Type | Formats |
|---|---|
| Input | PNG, JPEG, WEBP, TIFF |
| Output | PNG only |
Cross-format comparison is supported (e.g., compare JPEG to PNG).
Odiffr is designed for use in validated pharmaceutical and clinical research:
options(odiffr.path = ...)odiff_version() to
document binary version for audit trails# Pin to a specific validated binary
options(odiffr.path = "/validated/bin/odiff-4.1.2")
# Document in validation scripts
info <- odiff_info()
sprintf("Using odiff %s from %s", info$version, info$source)Odiff is approximately 6x faster than ImageMagick for pixel comparison, thanks to SIMD optimizations. Performance scales well with image size.
MIT