Classify pcov gate membership for a set of points on one curve

Usage

classify_pcov_gate(
  x_log10,
  pcov,
  pcov_threshold,
  lloq_log10 = NA_real_,
  uloq_log10 = NA_real_,
  inflect_x = NA_real_
)

x_log10: Numeric vector. log10-scale position: predicted_log10_ concentration for test samples, log10_concentration for grid points.
pcov: Numeric vector. Percent CV at each x_log10. Only used in the fallback branch (when LLOQ and/or ULOQ are undefined).
pcov_threshold: Numeric scalar. Only used in the fallback branch.
lloq_log10, uloq_log10: Numeric scalars (log10 scale), or NA if undefined for this curve/model.
inflect_x: Numeric scalar (log10 scale). Only used in the fallback branch, to decide which side of the curve a failing point falls on. If NA and a point fails, that point's pcov_gate_class stays NA (we know it fails, we just can't say which side) while pcov_pass is still correctly FALSE.

A data.frame with columns:

pcov_gate_class: Factor: belowLLOQ / inDynamicRange / aboveULOQ / NA. In the fallback branch, a failing point with no inflect_x available is NA here (we know it fails, not which side).
pcov_pass: Logical, computed independently of pcov_gate_class (NOT derived from it): TRUE/FALSE whenever the point could be evaluated at all, NA only when x_log10 (or, in the fallback branch, pcov) itself is unusable. A failing point with unknown side is pcov_gate_class = NA but pcov_pass = FALSE, never NA.

Row count and order match the input vectors exactly.