Feature/bring v0.4 into main

[AurelienNicosiaULaval]

Summary

This PR integrates the v0.4 publication-ready roadmap for tutorizeR, with a focus on making conversion workflows more robust for teacher-scale usage.

What this PR adds

• Core conversion stability:
  • deterministic chunk label handling
  • improved parsing/transform pipeline for Rmd/qmd
  • preserved setup and options behavior
• New MCQ ecosystem:
  • reusable question bank support (YAML/JSON): load_question_bank(), validate_question_bank()
  • MCQ reference blocks: {tutorizeR-mcq-ref}
• Linting and validation:
  • lint_source() with explicit severities/codes
  • richer tutorize_report + write_tutorize_report()
• i18n and teacher UX:
  • FR/EN message localization via tr()
  • clearer, localized errors and UX text
• Addins and export tooling:
  • preview addin workflow
  • LMS manifest export (export_lms_manifest())
  • optional tutorial package scaffolding (export_tutorial_package())
• Documentation and publication readiness:
  • JOSS paper scaffold (paper/paper.md, paper/paper.bib)
  • expanded vignettes (question-bank, lint-and-debug)
  • release/gov metadata and CI hardening

Validation

• R CMD check --as-cran --no-manual (pass, with expected non-blocking environment notes)
• Rscript -e "lintr::lint_package()" (No lints found)
• Rscript -e "testthat::test_local()" (98 PASS, 0 FAIL)
• CI green on matrix + lint + coverage jobs

Breaking changes

• None expected for public API:
  • convert_to_tutorial() preserved
  • convert_folder() preserved
  • new primary API entrypoint: tutorize() kept additive

Reviewer notes

• This PR is intentionally broad (v0.4 baseline + robustness + publication prep).
• If preferred, I can split into smaller follow-up PRs by layer (core/lint/i18n/exports/docs) in a second pass.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f76e705ab0

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Propagate normalized MCQ-ref strategy to caller

validate_mcq_ref_definition() tries to coerce invalid strategy values to "ordered", but this assignment is local and the function only returns TRUE; the original invalid value is still passed to select_questions_from_bank(), where match.arg() raises an error. Any tutorizeR-mcq-ref block with a typo like strategy: ordred will fail conversion instead of following the validator’s fallback logic.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Respect selected language in explicit Quarto MCQ blocks

For explicit tutorizeR-mcq blocks rendered to quarto-live, the callout builder is hard-coded with language = "en", so tutorize(..., language = "fr") still emits English quiz labels. This creates mixed-language output in localized tutorials and bypasses the i18n flow used elsewhere.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Exclude non-exercise tagged chunks from LMS manifest items

derive_manifest_items() labels every R chunk as an exercise without checking parsed tutorize tags, so chunks marked skip, narrative-only, or solution-only are still exported as learner exercises with solution_available. In files using these tags, the generated LMS manifest overstates interactive activities and can misrepresent what learners actually see.

Useful? React with 👍 / 👎.