---
title: "MCQ and Assessment Modes"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{MCQ and Assessment Modes}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

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

## User goal

Assessment modes let instructors decide whether a lesson should generate coding practice, conceptual questions, or both.

## Minimal examples

```{r eval=FALSE}
library(tutorizeR)

work_dir <- file.path(tempdir(), "tutorizeR-assessment")
source_file <- file.path(work_dir, "lesson.qmd")

tutorize(source_file, output_dir = work_dir, assessment = "code")
tutorize(source_file, output_dir = work_dir, assessment = "mcq")
tutorize(source_file, output_dir = work_dir, assessment = "both")
```

## Modes

- `code`: generate exercise and solution scaffolds.
- `mcq`: generate MCQ blocks only.
- `both`: generate code exercises, solutions, and MCQs.

## Realistic source pattern

```yaml
question: "Which variable identifies the academic program?"
answers:
  - text: "program"
    correct: true
  - text: "quiz_score"
    correct: false
```

In a source lesson, this YAML would be placed inside a `tutorizeR-mcq` fenced block.

## Realistic installed example

```{r eval=FALSE}
library(tutorizeR)

example_dir <- system.file("examples", "example_course_module", package = "tutorizeR")
work_dir <- file.path(tempdir(), "tutorizeR-assessment")
dir.create(work_dir, recursive = TRUE, showWarnings = FALSE)
file.copy(file.path(example_dir, "lesson-source.qmd"), work_dir, overwrite = TRUE)
file.copy(file.path(example_dir, "student_activity.csv"), work_dir, overwrite = TRUE)

qb <- load_question_bank(file.path(example_dir, "question-bank"))

report <- tutorize(
  input = file.path(work_dir, "lesson-source.qmd"),
  output_dir = work_dir,
  assessment = "both",
  question_bank = qb,
  mcq_source = "mixed",
  overwrite = TRUE,
  verbose = FALSE
)

print(report$mcq)
```

## Limits

Generated MCQs are scaffolds. Instructors should verify that questions are aligned with learning objectives and that distractors are plausible. The package does not measure whether an MCQ improves learning.

## Reproducibility checklist

```{r eval=FALSE}
lines <- readLines(report$output_file)
any(grepl("learnr::question", lines, fixed = TRUE))
```