---
title: "Computational benchmarking and synthetic stress testing"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Computational benchmarking and synthetic stress testing}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(collapse = TRUE, comment = "#>")
```

Scientific validity and computational feasibility are separate questions. `eye_benchmark_design()` measures runtime/scaling under declared dataset sizes, while synthetic corruption plans probe robustness to missingness, pupil dropout, calibration offsets, timestamp jitter, AOI label noise, device shifts, and trial imbalance.

```{r, eval=FALSE}
plans <- list(
 synthetic_corruption_plan(missingness=.05),
 synthetic_corruption_plan(missingness=.20, sampling_jitter_sd=2),
 synthetic_corruption_plan(pupil_dropout=.30, gaze_offset_x=.02)
)
st <- stress_test_process_pipeline(data, plans, analysis_fun)
stress_test_summary(st)
plot(st, severity="missingness", metric="effect")
```

Stress tests describe sensitivity to the perturbations actually supplied. They do not replace validation on independent empirical data.