Load package imports in generated shinylive app entrypoint

[JZauner]

Motivation

• The previously generated app.R entrypoint sourced R/*.R and called LightLogWeb() without attaching package imports listed in DESCRIPTION, causing runtime errors in a clean Shinylive environment for unqualified calls like page_navbar() and addResourcePath().
• Ensure exported Shinylive bundles reliably start by preloading the package dependencies the app needs at runtime.

Description

• Added parse_description_field() to parse and clean package names from DESCRIPTION fields.
• Added package_dependencies_for_entrypoint() to extract Depends and Imports from DESCRIPTION and filter out R and the package itself.
• Added build_shinylive_entrypoint() and updated export_shinylive_bundle() so the generated app.R first attaches dependencies with library(..., character.only = TRUE), then sources R/*.R, and finally runs LightLogWeb().
• Expanded unit tests in tests/testthat/test-export-shinylive-script.R to cover DESCRIPTION parsing, dependency extraction/filtering, and that the generated entrypoint includes dependency loading.

Testing

• Added unit tests in tests/testthat/test-export-shinylive-script.R that exercise parse_description_field(), package_dependencies_for_entrypoint(), and build_shinylive_entrypoint() and verify expected output.
• No automated R package tests were executed in this environment because Rscript / package test execution is unavailable, so the new tests were not run here.

---

Codex Task

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4912711123

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]

Guard missing DESCRIPTION dependency fields

Handle absent Depends/Imports columns before indexing desc, because read.dcf() omits columns that are not present and desc[1, "Depends"] (or "Imports") then throws subscript out of bounds. In that case export_shinylive_bundle() fails before generating app.R, so repos that drop one of these fields (which is valid in DESCRIPTION) cannot be exported even though parse_description_field() already supports empty inputs.

Useful? React with 👍 / 👎.