Closed christophscheuch closed 2 days ago
This PR got bigger than expected because I decided to do some ad hoc refactorings to make code test coverage work somehow and because I discovered redundancies and potential for improvements. There is only a small part in the ids_bulk()
function where I just gave up (see the #nocov tags).
My biggest learning is that from now on, every PR should also include a check for test coverage. It really helps to remove redundant code and question the implementation. I created a dev workflow write-up here that includes the process: https://github.com/Teal-Insights/r-wbids/wiki/Development-Workflow.
I guess the PR is fine since all tests and examples pass. Maybe you can just glance over it @chriscarrollsmith if you find something fishy.
Wow, this is a pretty kick-ass PR! Great work! And thank you for the documentation, too.
Is there a reason you prefer the
cli
library's stop, message, and warning functions over the standard base R ones?
cli
has better formatting and more options than base and it basically comes as a "free depedency" anyway if you use dplyr
or tidyr
.
Includes updating the tests to get 100% coverage. Current status:
wbids Coverage: 100.00% R/ids_bulk.R: 100.00% R/ids_bulk_files.R: 100.00% R/ids_bulk_series.R: 100.00% R/ids_get.R: 100.00% R/ids_list_counterparts.R: 100.00% R/ids_list_geographies.R: 100.00% R/ids_list_series_topics.R: 100.00% R/ids_list_series.R: 100.00% R/perform_request.R: 100.00% R/read_bulk_info.R: 100.00%
With this PR, we should also get a 100% codecov badge as here: https://github.com/tidy-intelligence/r-wbwdi