Skip to content

use_brand(), the #67 decision document, fixes for #101 and #102 - #114

Merged
larnsce merged 8 commits into
mainfrom
dev
Sep 2, 2026
Merged

use_brand(), the #67 decision document, fixes for #101 and #102#114
larnsce merged 8 commits into
mainfrom
dev

Conversation

@larnsce

@larnsce larnsce commented Sep 2, 2026

Copy link
Copy Markdown
Contributor

Summary

Closes #101. Closes #102. Refs #67, #109, #113.

Test plan

  • devtools::test() passes on the merge commit (67 passed and 2 CRAN-only skips on dev at 1a9b4c9)
  • pkgdown:::data_reference_index(pkgdown::as_pkgdown()) validates with use_brand listed
  • devtools::check() is clean
  • setup_readme() on a fresh package writes a license link containing the package name and no %7B
  • update_citation() on a fresh package adds ^CITATION\.cff$ to .Rbuildignore, once, across two runs
  • (Visual) a data package site built after use_brand() renders with the brand fonts and palette

larnsce and others added 8 commits August 19, 2026 09:31
First draft of the one-page decision settling canonical sources, field
mappings for the four dataspice files, the keywords home, artifact
purposes, and consequences for #68, #69, #70, #71, #87 and #47.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…e package (#67)

Amendments on acceptance: the staging layer keeps dataspice's file
format while the dataspice package leaves Imports; the keywords
derivation via X-schema.org-keywords is verified against cffr 1.4.1;
retiring the CSVs entirely is recorded as a v1.2.0 question. Adds the
Imports consequence for #72 (16 to 13 from the metadata side).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
use_brand() copies _brand.yml and the logo files it references from
the central openwashdata/brand repository (raw GitHub content at a
configurable ref, or a local/alternative source for tests and offline
work) into the active package. Re-running refreshes idempotently and
reports written and updated files by md5 comparison; unchanged files
stay untouched. With pkgdown = TRUE an existing _pkgdown.yml is wired
to the brand through template.bslib.brand so the next site build
renders with the brand fonts and colors; without a config the wiring
is skipped with a setup_website() hint. Brand values are never edited
locally, keeping the change-here-first discipline of the brand repo.

Five tests cover install, idempotent refresh with source change,
pkgdown wiring and re-run stability, the missing-config skip, and the
missing-source error. Full suite passes at 63.

Refs #109

Assisted-by: Claude claude-fable-5
…trial

bslib resolves the brand through the brand.yml package at site build
time; without it pkgdown::build_site() stops. Surfaced by the first
end-to-end run on a real data package.

Refs #109

Assisted-by: Claude claude-fable-5
use_brand(): install and refresh the openwashdata brand
inst/templates/README.Rmd carried the packagename placeholder in URL
encoded form ({{{ had become %7B%7B%7B), so whisker never substituted
it and every README that setup_readme() generated linked to a dead
LICENSE.md URL. The v1.0.2 CRAN release ships the encoded line.

Regression test added to test_setup_readme.R; it fails against the
old template and passes with the fix.

Closes #101

Assisted-by: Claude claude-fable-5-1
update_citation() writes CITATION.cff at the top level of the data
package but never added it to .Rbuildignore, so the next R CMD check
reported "Non-standard file/directory found at top level:
'CITATION.cff'". cffr 1.4.1 adds the entry itself only when
cff_write() is handed a file path; washr hands it a cff object, and
cffr's cff_update_rbuildignore() returns early for objects.

usethis::use_build_ignore() is idempotent, so re-running
update_citation() adds the entry once. Regression test added to
test_update_citation.R; it fails against the old code (the fixture
package has no .Rbuildignore until this call creates one) and passes
with the fix.

Closes #102

Assisted-by: Claude claude-fable-5-1
_pkgdown.yml lists every topic explicitly, and use_brand() landed in
PR #110 without an entry, so pkgdown::build_site() stopped on dev
with "1 topic missing from index: use_brand". Listed under
"Publishing your data" next to setup_website(), which it extends.

Refs #109

Assisted-by: Claude claude-fable-5-1
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

update_citation() leaves CITATION.cff out of .Rbuildignore README template: license link placeholder is URL-encoded and never substituted

1 participant