plotfit

Automatic plot dimension optimization for PDF exports of many ggplots.

plotfit analyzes ggplot objects on a PDF device, to return a suggested layout for the 'patchwork' package, 'patchwork::plot_layout()'. The goal is for individual plots to have optimal dimensions and sizing. The space between plots is not optimized. One can then manually assemble the plots into publication figures in a vector graphics editor like Illustrator or Inkscape.

What It Does

• Accepts a list of ggplot2 plots.
• Infers plot footprints automatically from axes, labels, legends, facets, and aspect constraints.
• Chooses one or more pages.
• Searches plot arrangements, row and column sizes for sensible physical plot dimensions.
• Returns an editable patchwork layout and R code.

Installation

remotes::install_github("jjfroehlich/plotfit")

Basic Usage

Load the installed package:

library(plotfit)

Optimize and inspect the suggested patchwork layout:

layout_result <- suggest_patchwork_layout(
  plots = list(p1, p2, p3, p4, p5, p6),
  page_width_in = 8.27,
  page_height_in = 11.69
)

# Main artifacts for review, manual adjustment, and reuse:
layout_result$pages[[1]]$patchwork
cat(layout_result$pages[[1]]$patchwork_code)

Patchwork Output

layout_engine = "patchwork" is the default output. It returns an optimized layout that remains easy to inspect, copy, paste, and adjust in R.

layout_result <- suggest_patchwork_layout(plots)

# Draw or further customize the editable patchwork object.
layout_result$pages[[1]]$patchwork

# Copy/paste this generated code into your own script as a starting point.
cat(layout_result$pages[[1]]$patchwork_code)

Constrained Groups

By default, plotfit decides which plots share a page. Use page_groups when the report's narrative requires a particular page assignment. Each element of the list defines one page, using the names from the plots list:

plots <- list(
  p1 = p1,
  p2 = p2,
  p3 = p3,
  p4 = p4,
  p5 = p5,
  p6 = p6,
  p7 = p7,
  p8 = p8, 
  p9 = p9, 
  p10 = p10
)

report_result <- suggest_patchwork_layout(
  plots,
  page_groups = list(
    group_1 = c("p1", "p2", "p3", "p4", "p5", "p6"),
    group_2 = "p7",
    group_3 = c("p8", "p9", "p10")
  )
)

Faster Search

Use the fast preset for interactive iteration. Explicit limits will still override the preset defaults:

quick_result <- suggest_patchwork_layout(plots, search_mode = "fast")
quick_result$performance_diagnostics$stages
quick_result$performance_diagnostics$candidates
quick_result$performance_diagnostics$fit_cache

Grid Output

layout_engine = "grid" renders the optimizer's physical row, column, and plot footprint measurements more exactly. It is mainly available for diagnostic comparisons when you want to distinguish sizing decisions from patchwork rendering behavior.

layout_result <- suggest_patchwork_layout(plots, layout_engine = "grid")
pdf("optimized-preview.pdf", width = 8.27, height = 11.69)
draw_layout_pages(layout_result)
dev.off()

Demo

Run the single demo script:

& 'C:\Program Files\R\R-4.2.2\bin\Rscript.exe' scripts\demo.R

Default output:

• demo_output/layout_feedback.pdf
• demo_output/previous/layout_feedback.pdf
• demo_output/archive/structural-scaling-baseline.pdf
• man/figures/readme-layout-comparison.png

The unified feedback PDF numbers every plot globally in its title (p1, p2, and so on). The previously existing PDF version is copied to demo_output/previous/ for side-by-side review.

Write optional diagnostics, warnings, and generated patchwork code:

& 'C:\Program Files\R\R-4.2.2\bin\Rscript.exe' scripts\demo.R --diagnostics

Development

Run tests:

& 'C:\Program Files\R\R-4.2.2\bin\Rscript.exe' -e "testthat::test_local('.', reporter = 'summary')"

Support

Get in touch if you have feedback or want to contribute.