occ_download_sql() for downloading occurrence data using SQL queries. (#752)occ_download_cached() bug fixed. (#748)occ_search() terms. (#698)occ_download_describe() for getting information about download formats. (#721)check_inputs()bug fixed. (#706)gbif_citation() datasetKey methods no longer supported (#716)occ_data() is no longer supported. (#718)occ_data() is soft deprecated and supported for legacy reasons only, and will no longer add new features.There have been many additions for accessing dataset metadata.
dataset_export() downloads all of the results of a dataset_search().
New functions for getting dataset metadata from a datasetkey (uuid) : dataset_get(), dataset_process(), dataset_networks(), dataset_constituents(), dataset_comment(), dataset_contact(), dataset_endpoint(),
dataset_identifier(), dataset_machinetag(), dataset_tag(), dataset_metrics().
New function for getting more obscure dataset metadata, such as machineTags: dataset().
New functions for listing dataset metadata : dataset_noendpoint(), dataset_duplicate().
dataset_doi() gets dataset metadata from the dataset's doi.
occ_count(). (#686)There are no longer static data files in rgbif. This data is better fetched fresh from the appropriate endpoints. (#690) (#688)
datasets() is soft deprecated, since the interface was overloaded and confusing. See functional replacements above.
map_fetch() now returns a base map as a magick::magick-image. This allows for the creation of high quality images from the GBIF maps API. (#675)occ_download() terms added to key lookup. (#661) (#589)pred_default() is an occ_download() pred function that allows users to easily filter out commonly unwanted occurrence records. (#611)Stream error fixed ("HTTP/2 stream 15 was not closed cleanly before end of the underlying stream"). Now map_fetch(), occ_data(), occ_search(), and occ_download_wait() have curlopts = list(http_version=2), which fixes the error. This might need to be the default setting for the whole package. (#656)
name_suggest() now gives a warning at prevents setting the limit > 100, since this is the GBIF API max. (#657)
New article Creating maps from occurrences, which explains how to use map_fetch().
occ_issues() is now deprecated, since it is difficult to maintain, and not widely used. (#651)occ_count() parameter type is now deprecated and will no longer work correctly. Please see occ_count_country(), occ_count_pub_country(), occ_count_year(), occ_count_basis_of_record() for replacements. (#622)occ_count() parameters georeferenced, type, date, to, from are no longer supported and not guaranteed to work correctly. (#622)occ_facet() and count_facet() are now deprecated use occ_count(facet="x") instead.lit_search() now supports searching the GBIF literature API. (#591)occ_count() now supports almost all occ_search() queries. (#622)occ_count() now supports the facets interface through occ_count(facet="x"). (#622)organizations() (aka publishers) now supports the use of getting lists of publishers by country. (#606)occ_download() and occ_search() now support downloading and getting occurrences a certain distance from known country/area centroids via the parameter distanceFromCentroidInMeters. (#594)occ_search() now supports more multi-valued parameters. (#617)randgeo and conditionz. (#624) (#625)New article explaining occ_count() changes and new features Getting Occurrence Counts From GBIF.
name_backbone_checklist() now accepts strict=TRUE, meaning that only non-fuzzy matches are returned. (#565)name_backbone_checklist() now accepts default values for high taxonomy, such as kingdom, phylum, family, ect. (#515)name_backbone_checklist() now returns a column is_alternative when verbose=TRUE, which lets the user know if a name was originally considered to be an alternative choice by the name matcher. (#515)sp and rgeos dependencies. (#578)name_usage now has the ability to fetch iucn red list categories using data=iucnRedListCategory. (#547)name_backbone_checklist updated definition of verbose argument. (#564)dataset_gridded bug fixed when inputting only one non-gridded dataset. (#546)occ_data() and occ_search() (#530)occ_download() terms to key lookup (#541)dataset_gridded() (#516)network_constituents() (#527)gbif_geocode() (#521)networks() is deprecated and called network() instead. (#527)parsenames() is deprecated and called name_parse() for better alignment with other name_* functions. (#504)occ_search fixed bug related to networkKey in the column names (#524)wellknown dependency and switching to wk (#512)name_backbone_checklist() : bug fix related two square brackets in url (#509)There is a new rgbif maintainer: John Waller.
derived_dataset() : New function to register a cleaned or modified dataset on GBIF for citation. (#467)name_backbone_checklist() : New function that takes a list, vector, or data.frame of scientific names and asynchronously matches them to the backbone. (#475)pred_isnull() : New predicate function that includes NULL values from a column in the download. (#489)occ_download.print() : Now prints out much more information including a DOI and citation. (#494)gbif_citation.gbif() : it is no longer considered best practice to generate a citation from occ_search() or occ_data(). We recommend occ_download() or derived_dataset() instead. (#494)occ_download_wait() and occ_download_meta() : now accept a class character download key directly. The keys does do not need to be class "occ_download". (#487)name_backbone() : now returns new columns "verbatim_name", "verbatim_genus" ect. that the user has supplied. This makes it easier for the user to track what has been matched. The verbose argument also has been un-retired. If verbose=TRUE, more results will be returned in a single data.frame. (#475)gbif_citation() : will now accept a download key directly.occ_download_get() : Does not throw an error if the data is already present and overwrite=FALSE, it will just give a warning and return the already present dataset. This allows users to run occ_download_get(key) %>% occ_download_import() multiple times without re-downloading the same file with overwrite=TRUE.download_predicate_dsl() : "publishingOrg" now added as a download key. (#496) key_lkup now includes GBIF-style uppercase keys as well. So pred("TAXON_KEY",212) and pred("taxonKey",212) will both work.Wrote new articles highlighting new features and encouraging the use of occ_download() over occ_search().
New articles:
occ_download_import() : fixed bug related to select argument. (#479)map_fetch() : fixed bug related to sp::CRS (#497)mulitpoint -> multipoint (#460) thanks @damianooldoni for catching thatstateProvince (#458), gadm (#462), and occurrenceStatus (#465)FOOTPRINT_SRS_INVALID and FOOTPRINT_WKT_INVALID (#454)occ_download_import() docs: more information on data.table::fread parameters and particular ones that would be useful to sort out data read issues (#461)occ_download_get(): downloaded files used to have a certain content type in response header we checked for, but its changed at least once even in successful responses, so that step has been removed (#464)occ_download_import(): country code for Namibia is NA - this was turning into the R missing value NA - now fixed (#463)coordinateUncertaintyInMeters added, e.g. usage: pred_lt("coordinateUncertaintyInMeters",10000) (#449)pred_and() and pred_or() slight change: now required that more than one predicate is passed to each of these functions because it doesn't make sense to do an and or or predicate with only one predicate (#452)pred_not(pred_notnull()) (#452)TAXON_MATCH_AGGREGATE) and a new name issue (BACKBONE_MATCH_AGGREGATE) (#453)pred_not(): it was not constructing the query correctly, fixed now. user facing change as well: it now expects a predicate to be passed, and only a single predicate as GBIF not predicate only accepts one predicate (#446)DIFFERENT_OWNER_INSTITUTION (#444)occ_search() (#443)occ_search() and occ_data() what parameters accept many values and which do not; in addition, we clarify which parameters accept multiple values in the same HTTP request, and those that accept multiple values but apply each in separate HTTP requests. See also ?many-values manual file (#369)gbif_issues() gains 9 new occurrence issues (#435)occ_search() and occ_data(), basisOfRecord parameter now supports multiple values, both in one request and in different requests, depending on input format (see "Multiple values passed to a parameter" section in ?occ_search) (#437)occ_download_get(): GBIF slightly altered download behavior - we now explicitly follow any redirects to get a download (#439)print.occ_download_meta (used when you run occ_download_meta()) was printing NA for number of results found if no results were ready yet - now prints 0 instead of NA (#440)count_facet() fixes: fixed internal fxn for count_facet for parsing results, was dropping values for facets; added assertions to check parameter types input by user for the fxn; changed so that keys and basisofrecord can be passed together (#436)gbif_issues(): GEOREFERENCED_DATE_INVALID and GEOREFERENCED_DATE_UNLIKELY (#430)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)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)occ_download_get(): response content-type header changed recently, fixed (#422)plyr::rbind.fill - use data.table::rbindlist in all cases (#417)dataset_search() (#418)occ_download_cancel_staged() fix: was broken cause we were indexing to a column in a table with [,"key"] (#416)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 maintenance, 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)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)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)/occurrence/search route is now 100,000 (used in occ_data() and occ_search()). updated docs throughout accordingly (#405)occ_download_queue() for how we determine when a job is done. see new section "When is a job done?" (#409)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)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)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)?downloads manual file listing all the different download status states a download can have and what they mean (#390)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)occ_download() via vcr (#396)name_lookup(): we were attempting to rearrange columns when no results found, leading to an error (#399)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)occ_data(): remove ... parameter definition as it wasn't used in the function (#394)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)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)gbif_citation() tests needed preserve body bytes for vcr (#384)occ_count() and count_facet(): isGeoreferenced/georeferenced variable needed booleans converted to lowercase before being sent to GBIF (#385) (#386)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 workmap_fetch() loses the format = ".mvt" option; and thus now only returns a RasterLayerocc_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)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)x, y, z parameters in map_fetch() manual file (#375)gbif_citation() manual file for accessing many citations (#379)occ_download_queue() (#365)name_* function outpus have changed, so be aware if you're using those functionsmap_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!gbif_citation() for occ_download_get objects. we weren't correctly creating the path to a file on windows (#359)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)occ_download(): when type = "in" was used, we weren't creating the JSON correctly, fixed now (#362)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)NA in addition to NULL for a missing parameter - both are removed before being sent to GBIF (#351)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)occ_download: fail with useful message when user does not pass in queries as character class (#347)occ_download: fail with useful message now when user/pwd/email not found or given (#348)elevation() function: the Google Maps API requires a form of payment up front, and so we've decided to move away from the service. elevation() now uses the Geonames service https://www.geonames.org/; it does require you to register to get a username, but its a free service. Geonames has a few different data models for elevation and can be chosen in the elevation_model parameter (#344) (#345)occ_data()/occ_search() output: the data.frame in the data slot now always has the first column as the occurrence key (key), and the second column is now the scientific name (scientificName). the previously used name column still exists in the data.frame, so as not to break any user code, but is simply a duplicate of the scientificName column. in a future version of this package the name column will be dropped (#329)occ_search()/occ_data()/occ_download() documentation on WKT (well-known text) with respect to winding order. GBIF requires counter-clockwise winding order; if you submit clockwise winding order WKT to occ_search() or occ_data() you should get data back but the WKT is treated as an exclusion, so returns data outside of that shape instead of within it; if you submit clockwise winding order WKT to occ_download() you will get no data back (#340)occ_download(), was failing in certain cases because of some bad code in an internal function catch_err() (#333)occ_download() was not returning user name and email in it's print method (#334)occ_issues() was failing with occ_data() or occ_search() input when type="many" (i.e., when > 1 thing was passed in) (#341)vcr package so do not require an internet connection (#306) (#327)name_usage() problem, more work coming to allow users to use the name issues data like we allow for occurrence issues through occ_issues() (#324)map_fetch() changes following changes in GBIF maps API: new parameters taxonKey, datasetkey, country, publishingOrg, publishingCountry and removed parameters search and id; note that this changes how queries work with this function (#319)map_fetch() docs that style parameter does not necessarily use the style you give it. not sure why (#302)occ_download_queue() to report an accurate number of jobs being processed; before we were just saying "kicking off first 3 requests" even if there were only 1 or 2 (#312)occ_get() when verbatim=TRUE (#318)elevation() function now fails better. when the API key was invalid the function did not give an informative message; now it does (#322)occ_download_queue(): sleep time between successive calls to check on the status of download requests is now 10 seconds or greater. This shouldn't slow down your use of occ_download_queue() much because most requests should take more than the 10 seconds to be prepared (#313)@importFrom fxns used from lazyeval package to avoid check note (#316)reshape2 and maps packages from Suggests (#317)name_usage(): we were screwing up parsing of issues column when single taxon keys passed in (#314)occ_issues() now works with download data and arbitrary data.frame's (#193)occ_download_prep() for preparing a download request without executing it, and occ_download_queue() for kicking off many download jobs while respecting GBIF's downloads rate limits. See also internal R6 classes for dealing with queuing: DownReq, GifQueue. See ?occ_download_queue to get started (#266) (#305) (#311)map_fetch() working with the GBIF maps API https://www.gbif.org/developer/maps. See ?map_fetch to get started (#238) (#269) (#284) thanks to @JanLauGe for the work on thisname_lookup() gains origin parameter (#288) (#293) thanks @peterdesmet and @damianooldoniname_lookup() and name_usage() gain internal paging - just as occ_search()/occ_data() have (#291) (see also #281) thanks @damianooldonilazyeval, and new suggests png and rasterocc_search()/occ_data() gain parameter skip_validate (boolean) to skip or not stkip WKT validation by the wicket packagename_usage() to check params that now only allow 1 value: name, language, datasetKey, rank (#287)occ_count() loses nubKey, catalogNumber, and hostCountry as those parameters are no longer accepted by GBIFname_usage(), was screwing something up internally (#286)occ_data(): curl options weren't being passed through (#297)occ_search()/occ_data() - skipping the wicket validation and constructing WKT by hand from bounding box (if bounding box given) - the validation that wicket does isn't what GBIF wants (#303)fill parameter to occ_download_import() to pass on to fill in data.table::fread, and set fill=TRUE as default. (#292)occ_download() (#300)occ_download() in which a single taxonKey passed in was failing (#283)name_usage() was ignoring datasetKey and uuid parameters (#290)gbifmap() has been removed, see the package mapr for similar functionality and map_fetch() in this package to use the GBIF map API (#298)occ_download_datasets and occ_download_dataset_activity to list datasets for a download,
and list the downloads activity of a dataset (#275) (#276)rgbif (#262)gbif_citation() function gains an S3 method for passing the
output of occ_download_meta() to it. In addition, for downloads
gbif_citation() now returns a citation for the entire download
(including) its DOI, in addition to citations for each dataset (#274)
thanks @dnoesgaardocc_count(): georeferenced had a
misleading description of what the value FALSE did (#265)gbifmap() - was failing in some cases - better
error handlingn now (#271) thanks @TomaszSuchanocc_download_cancel_staged(): it wasn't passing on authentication
parameters correctly (#280)?many-values manual file for details. added docs to
individual functions that support this, and added additional
tests (#200) (#260) (#261)V8 dependency and replaced with C++ based
WKT parser package wicket. We still use rgeos for some
WKT parsing. rgbif functions that use wicket: gbif_bbox2wkt,
gbif_wkt2bbox, check_wkt (#243)httr replaced with crul for HTTP reqeusts. As part of
this change, the ... parameter was replaced in most functions
by curlopts which expects a list. Some functions require
a ... parameter for facet inputs, so ... is retained
with the addition of curltops parameter. A result of this
change is that whereas in the past parameters that were not
defined in a function that also had a ... parameter
would essentially silently ignore that undefined parameter,
but with functions where ... was removed a misspelled
or undefined parameter will cause an error with message (#256)occ_download_import() to allow import
of csv type download in addition to darwin core archive.
additional change to occ_download_get to add format
attribute stating which format (#246)occ_download_import adding fill=TRUE to
the data.table::fread call (#257)occ_dowload gains new parameter body to allow users to pass in
JSON or a list for the query instead of passing in statements to
.... See examples in ?occ_dowload.tibble for compact data.frame output for
occ_download_import instead of bespoke internal solution (#240)https instead of http (#244)occ_download_metaocc_download to structure query correctly when
type=within and geometry used because the structure is slightly
different than when not using geometry (#242)occ_download to allow OR queries for many values of a
parameter, e.g., taxonKey=2475470,2480946 will be queried correctly
now as essentially taxonKey=2475470 or taxonKey=2480946 (#245)parsenames() caused by some slots in the list
being NULL (#237)occ_facet() tests were failing due to
changes in GBIF API (#239)gbif_oai_get_records() for slight changes in oai
dependency pkg (#236)occ_search() now has faceted search. This feature is not in occ_data()
as that function focuses on getting occurrence data quickly, so will not
do get facet data. This means that a new slot is available in the output
object from occ_search(), namely facets. Note that rgbif has had
faceted search for the species search route (name_lookup()) and the
registry search route (dataset_search()) for quite a while. (#215)occ_facet()) to facilitate retrieving only
facet data, so no occurrence data is retrieved. (#215) (#229)occ_search() and
occ_data() following addition the GBIF search API: subgenusKey,
repatriated, phylumKey, kingdomKey,
classKey, orderKey, familyKey, genusKey, establishmentMeans,
protocol, license, organismId, publishingOrg, stateProvince,
waterBody, locality (#216) (#224)spellCheck added to occ_search() and
occ_data() that if TRUE spell checks anything passed to the search
parameter (same as q parameter on GBIF API; which is a full text
search) (#227)occ_spellcheck to spell check search terms, returns
TRUE if no spelling problems, or a list with info on suggestions
if not.occ_search() and occ_data() now have ability to support
queries where limit=0, which for one should be possible and not
fail as we did previously, and second, this makes it so that you
can do faceted searches (See above) and not have to wait for occurrence
records to be returned. (#222)MULTIPOLYGON well known text features now supported in the GBIF
API. Previously, you could not query geometry with more than
one polygon (POLYGON), but now you can. (#222)occ_count(), especially for the set of
allowed parameter options that the GBIF count API supportsocc_count() gains new parameter typeStatus to indicate the
specimen type status.data slot now returns NULL instead
of a character stringgbif_photos(): 1) Mapbox URLs to their JS and CSS assets
were out of date, and API key needed. 2) In RStudio, the table view
was outputting errors due to serving files on localhost:<port>
instead of simply opening the file; fixed now by checking platform
and using simple open file command appropriate for the OS. (#228) (#235)tibble in most of the package when the output is
a data.frame (#204)is() with inherits(), no longer importing methods() (#219)data parameter, now they are (#210)gbifmap() man file (#212) thanks to @rossmounceocc_download(), in which
strings to parse were not being parsed correctly if spaces weren't in
the right place, should be more robust now, and added tests (#217). Came
from https://discuss.ropensci.org/t/rgbif-using-geometry-in-occ-download/395type was being silently ignored in a number of
registry functions. fixed that. (#211)occ_data() and occ_search() gain ability to more flexibly deal with inputs to the
geometry parameter. Previously, long WKT strings passed to occ_search() or
occ_data() would fail because URIs can only be so long. Another option is to use
the download API (see ?downloads). This version adds the ability to choose what to
do with long WKT strings via the geom_big parameter: asis (same as previous version),
bbox which detects if a WKT sting is likely too long, and creates a bounding box from the
WKT string then once data is retrieved, clips the result to the original WKT string; axe
uses the geoaxe package to chop up the input WKT polygon into many, with toggles in the
new parameters geom_size and geom_n. (#197) (#199)geom_big="axe", then
named elements of the output get names geom1, geom2, geom3, etc. instead of the
input WKT strings - this is because WKT strings can be very long, and make for very
awkward named access to elements. The original WKT strings can still be accessed via
attr(result, "args")$geometryname_usage() function, see commit e88cf01cc11cb238d44222346eaeff001c0c637etestthat fxn names, e.g., expect_gt()
instead of expect_more_than()occ_download() to parse error correctly when empty body passed from
GBIF (#202)occ_data() - its primary purpose to perform faster data requests. Whereas
occ_search() gives you lots of data, including taxonomic hierarchies and media records,
occ_data() only gives occurrence data. (#190)XML with xml2 (#192)data.table::rbindlist() for
fast list to data.frame coercion: name_lookup(), name_backbone(), name_suggest(),
name_usage(), and parsenames() (#191)httr usage to comply with changes in httr >= v1.1.0: now setting
encoding explicitly to UTF-8 and parsing all data manually, using the internal
function function(x) content(x, "text", encoding = "UTF-8") (#195)move_col() to not fail on fields that don't exist.
Was failing sometimes when no latitude or longitude columns were returned. (#196)gbif_oai_*()) for working with
GBIF registry OAI-PMH service. Now importing oai package to
make working with GBIF's OAI-PMH service easier (#183)r-curl/0.9.4 httr/1.0.0 rOpenSci(rgbif/0.9.0), with whatever
versions of each package you're using. We also pass a user-agent
string with the header X-USER-AGENT in case the useragent
header gets stripped somewhere along the line (#185)gbif_citation() helps get citations for datasets
eith using the occurrence search API via occ_search() or the
downloads API via occ_downlad() (#178) (#179)importFrom instead of import in all cases now.collectorName changed to recordedBy (#184)occ_download_meta() print method to handle 1 or more predicate results (#186)occ_issues() to work with return=data and return=all occ_search() output (#188)terraformer.js javascript code included in the package
along with an update in that codebase (#156)email parameter now NULL by default in the function
occ_download(), so that if not provided or not set in options,
then function fails. (#173)?downloads help file.elevation() to check for coordinates that
are impossible (e.g., latitude > 90), not complete (e.g., lat given,
long not given), or points at 0,0 (just warns, doesn't stop). (#176)
thanks @luisDVA/species
route, fix to function name_usage() (#174)occ_search() to remove a block of code to do synonym checking.
This block of code was used if the parameter scientificName was passed,
and checked if the name given was a synonym; if yes, we used the accepted
name according to the GBIF backbone taxonomy; if no, we proceeded with the
name given by the user. We removed the block of code because the GBIF
API now essentially does this behind the scenes server side. See
https://github.com/gbif/gbif-api for examples. (#175)gbif_photos() and gbif_names() (#170)occ_download*() for working with the GBIF download API. This is the same service as using the GBIF website, but via an API. See ?downloads. (#154) (#167)utils, methods, and stats (#166)httr v1 where empty list not allowed to pass to
the query parameter in GET (#163)/enumerations GBIF API route: enumeration()
and enumeration_country(). Many parts of the GBIF API make use of
enumerations, i.e. controlled vocabularies for specific topics - and are
available via these functions. (#152)elevation() now requires an API key (#148)V8 package an Import now, used to do WKT read/create with use of
the Javascript library Terraformer (http://terraformer.io/). Replaces
packages sp and rgeos, which are no longer imported (#155)occ_search() parameter spatialIssues to
hasGeospatialIssues (#151)/search and /count
services, and how they work. (#150)name_lookup() (#149)plyr from Imports (#159)stringr from Imports (#160)maps and grid packages from Imports (#161)limit and start parameters was in some
cases resulting in duplicate records returned. Problem fixed. (#157)\dontrun (#139)occ_search() to give correct values for default and max limit
and start parameters (#145)GET helper function to properly pass on error message (#144)assertthat::assert_that() with stopifnot() to have one less dependency (#134)occ_search() to allow ability to query by only publishingCountry, that is, with no
other parameters if desired (#137)GET() helper function to just pass NULL to the query parameter when
the list of length 0 passed, since it caused requests to fail in some cases.name_lookup() to force a logical entry for certain parameters - before this fix
if the correct logical param was not passed, the GBIF API went with its default parameter (#135)name_backbone() due to change in namelkupparser() helper function - fixes
parsing for verbose output (#136)occ_search() documentation (#140)occ_issues() to subset data from occ_search() based on GBIF issues. (#) (#122)occ_search() results, and are intentionally moved to the beginning of the column order of the data to be more obvious. (#102)occ_search() now returns all data fields by default. The default setting for the fields parameter is all - but can be changed. See ?occ_searchgbif_names() to view highlighted terms in name results from a call to name_lookup(). (#114)occ_issues_lookup() to lookup GBIF issues based on code name or full issue name, and gbif_issues() to print the entire issues table.RCurl with httrRJSONIO with jsonlite. Should see slight performance in JSON parsing with jsonlite.occ_search() now 500; was 25. (#113)organizations(), datasets(), networks(), nodes(), installations()) to data.frames when possible. (#117)occ_search() (#123)callopts parameter changed to ... throughout all functions. Now pass on options to httr as named lists or functions. (#130)dataset_search() and dataset_suggest() gain new parameter publishingOrglimit parameter changed to 100 for dataset functions: dataset_search(), dataset_suggest(), and datasets().limit parameter changed to 100 for registry functions: installations(), networks(), organizations, and nodes().networks(): name, code, modifiedsince, startindex, and maxresults gone; new parameters query, identifier, identifierType, limit, and startnodes(): new parameters identifier, identifierType, limit, and startocc_search() failed sometimes on species that were not found. Fixed. (#112)occ_search() (#119)name_lookup() (#120)gbif_photos() that caused map with no data. (#121)name_() functions according to changes in the GBIF API spec, and fixed documentation to align with GBIF API changes, and added note about maximum limit. (#124) (#127) (#129) Thanks to @willgearty !occ_search() so that user can pass in multiple values to the issue parameter. (#107)occ_search() now has a dplyr like summary output when return='all'. See ?occ_search for examples. You can still easily access all data, by indexing to meta, hierarchy, data, or media via e.g., $data, ['data'], or [['data']]. (#95)occ_search(), we now return a media slot in the output list by default.gbif_photos() to view media files (photos in the wild or of museum specimens). Two options are available, which='map' creates a single map which presents the image when the user clicks on the point, and which='table' in which a table has one row for each image, presenting the image and an interactive map with the single point. (#88)sp and whiskerrgbif changed accordingly. (#92)camelCase from under_score style - changed accordingly in rgbif.plyr::compact() instead of importing from plyr.name_lookup() removed facet_only parameter as it doesn't do anything - use limit=0 instead. Further, added two new slots of output: hierarchy and names (for common/vernacular names) (#96). The output can be determined by user via the return parameter.name_suggest(), if the field higherClassificationMap is selected to be returned via the fields parameter, a list is returned with a data frame, and a list of the hierarchies separately. If higherClassificationMap is not selected, only a data frame is returned.occ_search() gains new parameters mediatype and issue (#93), with detailed list of possible options for the issue parameter. Gains new examples for searching for images, examples of calls that will throw errors.check_wkt().facet_only parameter removed from dataset_search() function as it doesn't do anything - use limit=0 instead.gbif_bbox2wkt() and gbif_wkt2bbox() to convert a bounding box to wkt and a wkt object to a bounding box, respectively. Copied from the spocc package. Prefixes to fxn names will avoid conflicts.rgeos::readWKT or from the returned response from GBIF.gbifmap() was throwing an error because it was looking for two variables latitude and longitude, which had been changed to decimalLatitude and decimalLongitude, respectively, in other functions in this package. Fixed. (#81)occ_get() was updated to include changes in the GBIF API for this endpoint. The fix included fixing the parser for verbatim results, see rgbif::gbifparser_verbatim. (#83)elevation() - it was expecting column names to be latitude and longitude, whereas inputs from other rgbif functions have changed to decimalLatitude and decimalLongitude.count_facet() introduced b/c GBIF no longer accepts hostCountry or nubKey parameters.gist(), stylegeojson(), and togeojson() functions now listed as deprecated. Their functionality moved to the spocc package (http://cran.r-project.org/web/packages/spocc/index.html). These functions will be removed from this package in a future version. (#82)gbifmap().occ_get() for when verbatim=TRUE, which gives back different data than when verbatim=FALSE.latitude is now decimalLatitude. longitude is now decimalLongitude. clazz is now class. Code in this package changed to accomodate these changes. date is now eventDate. georeferenced is now hasCoordinate. Beware of these changes in your own code using rgbif - find and replace for these should be easy.altitude parameter in occ_search() to elevation - should have been elevation the whole time.occ_count() function with parameter changes: nubKey parameter in changed to taxonKey. New parameter protocol. Parameter catalogNumber gone. Parameter hostCountry gone. These parameters are still in the function definition, but if called they throw a useful warning telling you the correct parameter names. (#76)name_lookup() function that was labeling facet outputs incorrectly. (#77)application/json, then parse JSON ourselves using RJSONIO::fromJSON (instead of httr doing it).stringsAsFactors = FALSE to all data.frame() calls.occ_search() function. These parameters are still in the function definition, but if called they throw a useful warning telling you the correct parameter names. (#75)name_usage, name_backbone, name_lookup, and name_suggest functions.sourceId parameter in name_usage() function doesn't work so error message is thrown when used.check_wkt() to check that well known text string is the right format. (#68)occ_search() function. scientificName: search for a species by name (instead of taxonKey). continent: search by continent. lastInterpreted: search by last time GBIF modified the record. recordNumber: search by the data collector's specimen record number - this is different from the GBIF record number. typeStatus: search by specimen type status. (#74)parameterName = c(x, y, z). These parameters are: taxonKey, scientificName, datasetKey, catalogNumber, collectorName, geometry, country, recordNumber, search, institutionCode, collectionCode, decimalLatitude, decimalLongitude, depth, year, typeStatus, lastInterpreted, and continent. This isn't faceted search server side - this is just looping your different values of the parameter against the GBIF API.occ_search() now support range queries: decimalLatitude,decimalLongitude,depth,elevation,eventDate,lastInterpreted,month, and year. Do a range query for example by depth=50,100 to ask for occurrences where depth was recorded between 50 and 100 meters. Note that this syntax depth=c(50,100) will perform two separate searches, one for depth=50 and one for depth=100. (#71)DEPRECATED