plot_lint() compiles a plot and reports the design problems a static render
hides from a green test suite: text too small to read, colour contrast below
the WCAG threshold, two palette colours a colour-blind reader cannot tell
apart, labels that overlap or fall off the panel, and grammar-level mistakes
such as an encoding with a single level or a legend too long to read. It is
the "flag it" step of the accessibility workflow — pair it with
render_plot(cvd = ) to see a failing palette and scale_pattern() /
pattern_hatch() to fix it with a redundant non-colour encoding.
Arguments
- x
A PlotSpec (or anything
vellum::as_vellum_scene()accepts).- ...
Passed to
vellum::vl_lint():rules,exclude,severity,cvd,min_text_pt,max_overplotand the rest of the thresholds.- min_text_px
Minimum legible text height in pixels (default
7); text below it is flagged.- min_contrast
Minimum acceptable contrast ratio between a mark and its background (default
3, the WCAG AA threshold for graphical objects).
Value
A data frame (class vellum_lint) with one row per finding: rule,
severity ("warning"/"note"), node, a human message, and the
device-px box x0/y0/x1/y1 — NA for a grammar finding, which is
about a scale and so has no box. Zero rows when the plot is clean.
Details
The geometric rules come from the engine's vellum::vl_lint(), which judges
them in resolved device pixels. The grammar rules are registered into the same
registry by vellumplot, so they are not special: vellum::vl_lint_rules()
lists them, rules = selects them, and vellum::vl_lint() on a compiled plot
reports them too. Findings are returned most-severe first.
Because every finding carries the box it refers to,
vellum::vl_lint_overlay() can draw the report onto the plot, and
vellum::vl_lint_assert() can fail a test on it.
A composition or table has no single set of trained scales, so it reports the geometric findings only — lint the cells individually for the grammar ones (vellumplot#147).
Examples
# a tiny-text, single-level-scale plot trips the linter
p <- vplot(transform(mtcars, grp = "one group")) |>
mark_point(x = wt, y = mpg, color = grp) |>
theme(axis.text = element_text(size = 2))
plot_lint(p)
#> 9 lint findings (8 warnings):
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ℹ [single_level_scale] scale:color: The color scale has a single level (one
#> group): the encoding conveys nothing and its legend is redundant.
# the engine's arguments come through, so a project can accept a finding
plot_lint(p, exclude = "scale:color")
#> 8 lint findings (8 warnings):
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor
#> ✖ [tiny_text] text: 2.7 px tall - below the 7 px legibility floor