---
title: "Working with bundled examples"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Working with bundled examples}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

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

The package includes a small set of self-contained example exercises in
`inst/exercises`. These are useful for testing a fresh installation or for
bootstrapping your own authoring workflow.

## Locate the exercises

```{r}
example_dir <- system.file("exercises", package = "exams2ilias")
list.files(example_dir, full.names = TRUE)
```

## Export one exercise

```{r}
library(exams2ilias)

outdir <- tempfile("ilias-")
dir.create(outdir)

exams2ilias(
  file.path(example_dir, "stats_schoice.Rmd"),
  n = 1,
  dir = outdir,
  name = "stats_schoice",
  xmlcollapse = FALSE,
  solutionswitch = FALSE
)
```

## Export the full set

The helper script `generate_examples.R` exports each bundled exercise
individually and can also create one combined question pool.

```{r}
source(file.path(example_dir, "generate_examples.R"))

outdir <- tempfile("ilias-examples-")
dir.create(outdir)

generate_example_exports(outdir)
```

## Adapt an exercise for your own pool

A practical workflow is:

1. Start from the closest bundled exercise.
2. Replace the statistical scenario and answer logic.
3. Keep the metadata structure intact.
4. Export one question first and test the import in ILIAS.
5. Only then batch-export larger question sets.

This keeps debugging cheap and avoids importing many broken items at once.