Tags the rendered Victory SVG(s) inside container and builds the MAIDR
subplot grid.
Single-panel figures keep the original flat tagging
(data-maidr-victory-<n> attributes on the container's first svg) and the
original [[{ layers, legend }]] shape, byte-identical to previous output.
Multi-panel figures resolve one svg per panel (bound via each panel's
svgIndex ordinal among all top-level Victory components, so non-chart
Victory siblings such as a standalone scatter or a shared legend cannot
shift the binding), stamp each panel svg with
data-maidr-victory-panel="<i>", and emit panel-scoped attribute names and
selectors so panels can never cross-highlight. Each panel gets its own
claimed element set so tagging never leaks across panels.
Panels whose charts contain no supported data are dropped (the core model
cannot represent an empty subplot inside a grid), but only AFTER grid
chunking: every remaining panel keeps the grid cell of the visual CSS
arrangement, and only rows left entirely empty are removed (the core cannot
represent empty rows; ragged rows navigate fine).
Tags the rendered Victory SVG(s) inside
containerand builds the MAIDR subplot grid.Single-panel figures keep the original flat tagging (
data-maidr-victory-<n>attributes on the container's first svg) and the original[[{ layers, legend }]]shape, byte-identical to previous output.Multi-panel figures resolve one svg per panel (bound via each panel's
svgIndexordinal among all top-level Victory components, so non-chart Victory siblings such as a standalone scatter or a shared legend cannot shift the binding), stamp each panel svg withdata-maidr-victory-panel="<i>", and emit panel-scoped attribute names and selectors so panels can never cross-highlight. Each panel gets its ownclaimedelement set so tagging never leaks across panels.Panels whose charts contain no supported data are dropped (the core model cannot represent an empty subplot inside a grid), but only AFTER grid chunking: every remaining panel keeps the grid cell of the visual CSS arrangement, and only rows left entirely empty are removed (the core cannot represent empty rows; ragged rows navigate fine).