The exported _data() functions ppc_data(), ppd_data(), ppc_ribbon_data(), and ppd_ribbon_data() have no direct test coverage. These functions are part of the public API, intended for users who want raw computed data for custom visualizations.
ppc_data() / ppd_data() melt prediction matrices into long-format data frames via the internal .ppd_data() helper, handling y/yrep stacking, group joining, and label generation.
ppc_ribbon_data() / ppd_ribbon_data() are exported aliases for ppc_intervals_data() / ppd_intervals_data().
While these functions are exercised indirectly through their corresponding plot functions, bugs in the data transformation logic (incorrect columns, wrong row counts, broken group handling) would silently propagate to plots and user pipelines with no warning.