rgbif v3.3.0

MINOR IMPROVEMENTS

• added two new occurrence issues to gbif_issues(): GEOREFERENCED_DATE_INVALID and GEOREFERENCED_DATE_UNLIKELY (#430)

BUG FIXES

• fixed an error in occ_data() caused by GBIF adding a new field of data to the output of /occurrence/search/: gadm. cleaned up internals of occ_data() to drop gadm, and other fields that are complex and take time to parse (use occ_search() if you want all the data fields) (#427)
• gbif_names() fix: was ending up with invalid URLs to GBIF species pages because we had taxon keys with leading spaces somehow. now all leading and trailing spaces in taxon keys removed before making URLs (#429)

rgbif v3.2.0

MINOR IMPROVEMENTS

• gbif_issues() changes: three new occurrence issues added; one name issue removed that's deprecated (#423)
• gbif_citation() rights field was empty unless pulling from a downloaded file; now fill in with license key; also a fix for when occurrence key passed to the function (#424)
• establishmentMeans now supported in occ_download/pred (#420)

BUG FIXES

• fix for occ_download_get(): response content-type header changed recently, fixed (#422)

rgbif v3.1.0

MINOR IMPROVEMENTS

• finally delete code originally extracted from plyr::rbind.fill - use data.table::rbindlist in all cases (#417)
• fix failing test on cran for dataset_search() (#418)
• fix xd refs note on cran (non-file package anchored links) for curl pkg function (#419)

BUG FIXES

• occ_download_cancel_staged() fix: was broken cause we were indexing to a column in a table with [,"key"] (#416)

rgbif v3.0.0

BREAKING CHANGE

• Many functions (occ_search, occ_get, name_usage, name_lookup, name_suggest, name_backbone, and dataset_search) have a return parameter to toggle what is returned from the function call. To simplify rgbif maintainence, we've deprecated the return parameter. We've left it in each of the functions, but it no longer does anything, other than raising a warning if used. This means that function calls to these functions now always return the same data structure, making it easier to reason about for the user, as well as for us developers trying to make sure the package works as expected under a variety of conditions. If you have been using the return parameter, do the same function call as before, but now index to the output you need. This is a breaking change, thus the major version bump (#413)

NEW FEATURES

• new function occ_download_cached(), which takes the same input as occ_download(), but instead of starting a query, it checks if you've recently made the same request (with configureable settings for what "recent" means). This can save time when you're doing occurrence download requests that you may have done in the recent past (#308)

MINOR IMPROVEMENTS

• configured package to be able to use two different base urls, api.gbif-uat.org and api.gbif.org. We have only used the latter previously, but now can configure rgbif to use the former, mostly for testing purposes (#398)
• occ_download_import() gains encoding parameter that is passed down to data.table::fread to make it very clear that encoding can be configured (even though you could have before via ...) (#414)

BUG FIXES

• fix tibble construction (#412)

rgbif v2.3

MINOR IMPROVEMENTS

• max records you can return for /occurrence/search route is now 100,000 (used in occ_data() and occ_search()). updated docs throughout accordingly (#405)
• improved docs in occ_download_queue() for how we determine when a job is done. see new section "When is a job done?" (#409)
• print methods print.occ_download_prep and print.occ_download improved. previously well-known text strings were printed in their entirety. now they are handled to only print so many characters; also applies to any download predicate string that's long (#407)
• occ_download_get() now supports using a progress bar by passing in httr::progress() (#402)
• occ_data() and occ_search() gain two new parameters: recordedByID and identifiedByID (#403)

BUG FIXES

• fix in occ_download_queue: an empty occ_download_meta() lead to problems; now removing any NULL's from a list of occ_download_meta() outputs before further work (#408)
• fix in occ_download_queue: we were not accounting for job status "cancelled" (#409)
• occ_download_import() fix: fill parameter was set to TRUE by default, changed to FALSE. improved docs for this fxn on passing down parameters to data.table::fread (#404)

rgbif v2.2.0

MINOR IMPROVEMENTS

• add a section Download status to the ?downloads manual file listing all the different download status states a download can have and what they mean (#390)
• fix gbif_issues/gbif_issues_lookup: added four missing occurrence issues to the package (COORDINATE_PRECISION_INVALID, COORDINATE_UNCERTAINTY_METERS_INVALID, INDIVIDUAL_COUNT_INVALID, and INTERPRETATION_ERROR) (#400)
• doing real tests now for occ_download() via vcr (#396)

BUG FIXES

• fix name_lookup(): we were attempting to rearrange columns when no results found, leading to an error (#399)

rgbif v2.1.0

DEFUNCT

• the spellCheck parameter has been removed from the occurrence routes; thus, the occ_spellcheck() function is now defunct - and the parameter spellCheck has been removed from occ_data() and occ_search() (#397)

MINOR IMPROVEMENTS

• docs fix for occ_data(): remove ... parameter definition as it wasn't used in the function (#394)

BUG FIXES

• download predicate fxns fix: "within" wasnt being handled properly (#393) thanks @damianooldoni

rgbif v2.0.0

NEW FEATURES

• The download query user interface for occ_download() has changed in a breaking fashion (thus the major version bump). After installation, see ?download_predicate_dsl. Much more complex queries are now possible with occ_download(). TL;DR: you now construct queries with functions like pred("taxonKey", 3119195) rather than passing in strings like taxonKey = 3119195, and pred_gt("elevation", 5000) instead of "elevation > 5000" (#362)
• gains new function occ_download_wait() to re-run occ_download_meta() until the download is ready - kinda like occ_download_queue() but for a single download (#389)
• occ_download_dataset_activity() gains pagination parameters limit and start to paginate through results (#382)
• gbif_citation() now works with the output of occ_data() in addition to the other existing inputs it accepts (#392)

MINOR IMPROVEMENTS

• typo fix in the geometry section of the occ_download() manual file (#387)
• vignettes fixes (#391)

BUG FIXES

• gbif_citation() tests needed preserve body bytes for vcr (#384)
• fix to occ_count() and count_facet(): isGeoreferenced/georeferenced variable needed booleans converted to lowercase before being sent to GBIF (#385) (#386)

rgbif v1.4.0

NEW FEATURES

• gains new function mvt_fetch() for fetching Map Vector Tiles (MVT). mvt used to be an option in map_fetch(), but we only returned raw bytes for that option. With mvt_fetch() we now leverage the protolite package, which parses MVT files, to give back an sf object (#373) thanks to @jeroen for the protolite work to make this work
• associated with above, map_fetch() loses the format = ".mvt" option; and thus now only returns a RasterLayer
• occ_issues() and name_issues() reworked. Both now use the same underlying internal logic, with occ_issues pulling metadata specfic to occurrence issues and name_issues pulling metadata specific to name issues. name_issues used to only be a data.frame of name issues, but can now be used similarly to occ_issues; you can pass the output of name_usage() to name_issues to filter/parse name results by their associated name issues. Associated with this, new function gbif_issues_lookup can be used to lookup either occurrence or name issues by their full name or code (#363) (#364)

MINOR IMPROVEMENTS

• fix examples and tests that had WKT in the wrong winding order (#361)
• parsing GBIF issues in the output of name_usage() wasn't working (#328) (#363) (#364)
• name_lookup() gains an additional parameters issue for filtering name results by name issues (#335) (#363) (#364)
• fixed definitions of x, y, z parameters in map_fetch() manual file (#375)
• added examples to gbif_citation() manual file for accessing many citations (#379)
• fixed a test for occ_download_queue() (#365)
• name_* function outpus have changed, so be aware if you're using those functions

BUG FIXES

• fixed issue with map_fetch(): when srs was EPSG:3857, the extent we set was incorrectly set as raster::extent(-180, 180, -85.1, 85.1). Now the extent is raster::extent(-20037508, 20037508, -20037508, 20037508 (#366) (#367) thanks @dmcglinn for reporting and @mdsumner for fixing!
• fix for Windows platforms for gbif_citation() for occ_download_get objects. we weren't correctly creating the path to a file on windows (#359)
• fix to print.gbif_data (#370) (#371)
• occ_download() was erroring with a useless error when users try to use the fxn with the same parameter input types as occ_search/occ_data; when this happens now there is a useful error message (#381)
• fix to occ_download(): when type = "in" was used, we weren't creating the JSON correctly, fixed now (#362)

rgbif v1.3.0

NEW FEATURES

• occ_download() and occ_download_prep() gain a new parameter format for specifying the type of download. options are DWCA (default), SIMPLE_CSV, or SPECIES_LIST. SIMPLE_CSV and SPECIES_LIST are csv formats, while DWCA is the darwin core format (#352)
• now throughout the package you can pass NA in addition to NULL for a missing parameter - both are removed before being sent to GBIF (#351)

MINOR IMPROVEMENTS

• replace tibble::as_data_frame/tibble::data_frame with tibble::as_tibble (#350)
• key and gbifID in the output of occ_data/occ_search/occ_get have been changed so that both are character class (strings) to match how GBIF encodes them (#349)
• fix some test fixtures to use preserve exact bytes so that cran checks on debian clang devel don't fail (#355)

BUG FIXES

• fix to occ_download: fail with useful message when user does not pass in queries as character class (#347)
• fix to occ_download: fail with useful message now when user/pwd/email not found or given (#348)