Bcdata

------------------
Major revision, nothing is backwards compatible
- download data via WFS rather than DWDS
- functions renamed
- data downloaded to file is to geojson only
- added `bc2pg` command for mirroring BC table in postgres

* Code in `.onLoad()` that sent a request to the wfs getCapabilities endpoint could cause the package to fail to load. This was moved into an internal function `bcdc_get_capabilities()` that makes the request the first time it's required, and stores the result for the remainder of the session (271)
* testthat is now used conditionally to only run tests if the testthat package is installed.

IMPROVEMENTS
- Setting the `bcdata.single_download_limit` limit dynamically from the getCapabilities endpoint. 256
- `bcdc_describe_feature` now joins an object description column to the returned object to provide more information about a field directly in R. 241, 259
- Better documentation and information surrounding the `bcdata.max_geom_pred_size` option. 243, 258
- Add new function `bcdc_check_geom_size` so users can check prior to submitting a WFS request with `filter` to see
if the request will require a bounding box conversion. 243, 258
- Better documentation and messaging about when and why paginated requests are required by `bcdc_query_geodata()`. 240, 264
- Better documentation and print method for what records are suitable for use with `bcdc_query_geodata()`. 265, 267

IMPROVEMENTS
* Added `bcdc_list_groups` and `bcdc_list_group_records` to provide the ability to query on the group endpoint of the catalogue API. 234
* Added new option `bcdata.single_download_limit` to enable setting the maximum number of records an object can be before forcing a paginated download (252)

BUG FIXES
* Fixed bug in `collect.bcdc_promise` where the wrong parameter name in `crul::Paginator$new()` resulted in an error in paginated wfs requests (250, thanks meztez)
* Fixed a bug where the name of `bcdata.chunk_limit` option had a typo, so that it was not actually used properly (252)

BUG FIXES
* Remove link for pipe documentation for simplicity.
* Fixed bug where using many `as.` functions (e.g., `as.Date()`, `as.character()`, `as.numeric()`) in a filter statement would fail. (218, 219)

MAINTENANCE
* Updated internal SQL translation to use `DBI` S4 generics (`DBI::dbQuoteIdentifier()` is now used instead of
`dbplyr::sql_escape_ident()` and `DBI::dbQuoteString()` instead of `dbplyr::sql_escape_string()`), to comply
with upcoming `dbplyr` 2.0 release (225, 225; https://github.com/tidyverse/dbplyr/issues/385)
* Wrapped all examples that call web resources in `try()` to avoid spurious check failures (229).

------------------
- use DataBC Catalogue API to find data (25, 31)