Tracking coverage of OA entities

[yjunechoe]

New file openalexR-coverage.R tracks oa2df()'s coverage of OA entity fields (currently just one draft for the Work entity)

New function get_coverage() introduces an interface to the coverage information, given an entity

This could also use some tests to ensure a match between what oa2df() actually returns vs. what we say it returns.

[yjunechoe]

@trangdata Could you help me double check my understanding of how we process variables from Work objects? And feel free to suggest anything to add to the comments column

[trangdata]

Thank you for putting this together @yjunechoe 💯 We needed this a long time ago. You're as thorough as always! 🚀

I will make a PR if I think of other things to add to comments.

[rkrug]

I like this functionality.

One point: in the case of e.g. biblio, where there are values extracted, it would be great if the field could be specified, e.g. biblio.volume, volume and also biblio, NA to indicate where the value is coming from and that biblio itself is removed.

That would make it clearer to understand and also make it possible to rename sub fields after extraction.

[rkrug]

Oh - I would also suggest to add the sorting of the fields in this function (or a separate function).

[trangdata]

@yjunechoe I can't add you as a reviewer because you opened the PR. But I would love to get your feedback when you have a minute. (no rush though!)

• add warnings for the breaking changes
• add keywords as an entity, remove documentation for concepts
• add coverage for all entities, update test
• update readme and news

[yjunechoe]

This looks great @trangdata !! Thanks so much for finishing up the coverage data - I know that's a lot of manual labor!

One thing that stands out is our current handling of data, e.g.:

get_coverage <- function(entity = NULL) {
  utils::data("oa2df_coverage", envir = environment(), package = "openalexR")
  oa2df_coverage <- oa2df_coverage

I don't mean to overthink fixing something that's not broken (and I can't even come up with something better ATM) but just want to mull over this a bit. An alternative to utils::data() that's less side-effect-y is to use oa2df_coverage <- get("oa2df_coverage", envir = asNamespace("openalexR")). Anyways I'm attempting to nerd-snipe some R wizards with a mastodon post so let me come back to this but otherwise the PR LGTM!

---

P.S. - Now that we have {rlang}, we're 1-step removed from also adding {cli} which I think may be worth considering later as QoL improvement: ex: warn() -> cli_warn(). The cli::cli_progress_*() functions can also replace our dependency to {progress}!

[trangdata]

Thank you @yjunechoe! So are you saying oa2df_coverage <- get("oa2df_coverage", envir = asNamespace("openalexR")) would work? Or it's still not working?

P.S. - Now that we have {rlang}, we're 1-step removed from also adding {cli} which I think may be worth considering later as QoL improvement: ex: warn() -> cli_warn(). The cli::cli_progress_*() functions can also replace our dependency to {progress}!

Great point! I'm happy to add cli to Imports (and remove progress)!

[rkrug]

Just another idea (after reading the problem with the data): why not have a function returning the data, and also provide the function with one argument, the version of openalexR which defaults to the currently installed version. By doing this, one could easily go back in time and see what changed between an older version and the new version in regards to coverage - very useful if one want’s to re-run a script after some time again and it does not work.

[trangdata]

@yjunechoe OK I have moved these points to #297 and #296. Let's continue our discussions there.

@rkrug Thanks for your input. I have added the warning in multiple places and described the changes in NEWS.md. We understand that breaking changes require adjustment, but they are often essential for keeping the package aligned with best practices and maximizing functionality over the long term.