Incorporating the latest round of API changes

[mustberuss]

Things to note:

• The new API field shorthand is allowed in the fields parameter and is used by get_fields(). Without it the patent endpoint has so many attributes that GET requests can fail (> 4K url) on some queries (if all fields were fully qualified). From their jupyter notebook:

The fields for related entities can be requested in the API request's fields parameter as a group by using the group name in the fields parameter, or individually by specifying the required field as "{entity_type}.{subfield}".

• It's still test-heavy for now, at lot has changed and the API has problems (see test-api-bugs.R)

Should we bump the version to 1.0.0 because of the breaking changes?

[crew102]

Hey Russ, can you summarize what this PR does? Also I see that the tests on CI aren't passing, can you comment and/or make sure they pass?

[crew102]

Can we also drop all those tests that we talked about?

[mustberuss]

It's the remaining code to catch up with the API:

• paging by primary key only and sorting in R; deprecating the page and per_page search_pv() parameters; padding the patent_id's after value when patent_id is used as the sort field; exporting the padding function and adding after as a search_pv parameter for custom paging; removal of paging limits
• accepting group names in the field parameter (API's new shorthand); get_fields() returning group names as fields rather than fully qualifying (group.member notation) each member of the group
• unnesting and casting changes now that the endpoint names (singular, some nested) and returned entities (mostly plural) aren't interchangeable; dealing with the two endpoints that return rel_app_texts (from patent/rel_app_text and publication/rel_app_text)
• switching to httr2, httr says it's superseded: only changes necessary to keep it on CRAN will be made.
• there's an additional inclulde_pk boolean on get_fields() and encoded_url boolean on retrieve_linked_data()
• there is a new qry_fun in_range() that isn't strictly needed, it's not fronting anything in the API but would generate the tedious _and/_gte/_lte needed ex: qry_funs$in_range(patent_date = c("1976-01-01", "1983-02-28"))
• field name/endpoint name changes in the examples
• bug fix on posts when fields or sort wasn't specified
• bug fix when there was more than a primary sort
• bug fix to avoid a coercion warning

We can remove tests but some of the suggestions you made have code changes now, especially get_fields() and, argument changes to search_pv()

I'm not sure what's wrong with the build. I bumped versions in the workflow to get rid of deprecated warnings but didn't intentionally change anything, I can try reverting. It's working in my repo but it's trying to make API calls here. There are skip_on_cran()s in the tests but they're running and so are the dont_run examples.