- fixed a bug in building the documentation with readthedocs :issue:`138`, :issue:`129`
- tests now run against live GBIF APIs :issue:`101`, :issue:`128`
- updated caching.py since the
remove_expired_responses
method is deprecated. :issue:`126`
- added support for predicates:
isNull
,isNotNull
,in
andnot
:issue:`92`, :issue:`102` and :issue:`103` - added support for nested queries/dictionaries :issue:`104`
- deprecated the
add_predicate
function and addedadd_pred_dict
to accomodate for newly supported predicates to ensure that the arguments that are sent are added in the payload function :issue:`108` - added support for multiple download formats :issue:`105`
- updated operators and look-up tables :issue:`107`
- included documentation on newly supported predicates and dictionaries :issue:`106`
- update to fix requesting GBIF downloads
- minor documentation updates :issue:`95` and :issue:`99`
- update to fix broken dependencies :issue:`93`
- minor documentation updates
- Fix for occurrences.download when giving geometry as a string rather than using add_geometry; predicates were being split on whitespace, which doesn't work for WKT :issue:`81` :issue:`84`
- Moved to using the logging module instead of print() for giving information on occurrence download methods :issue:`78`
- Clarify that occurrences.count for length 1 inputs only; see occurrences.search for > 1 value :issue:`75` :issue:`77`
- Improved documentation for species.name_usage method, mostly for the language parameter :issue:`68`
- Gains download method download_cancel for cancelling/deleting a download request :issue:`59`
- occurrences.search now supports recordedByID and identifiedByID search parameters :issue:`62`
- clean up the Contributing file, thanks @niconoe :issue:`64`
- clean up internal imports in the library, thanks @niconoe :issue:`65`
- fix usage of is and ==, was using them inappropriately sometimes (via https://realpython.com/python-is-identity-vs-equality/), :issue:`69`
- remove redundant parameter in a doc string, thanks @faroit :issue:`71`
- make a test for internal fxn gbif_GET_write more general to avoid errors if GBIF changes content type response header slightly :issue:`72`
- changed base url to https for all requests; was already https for maps and downloads in previous versions
- occurrences, species, and registry modules gain docstrings with brief summary of each method
- pygbif gains ability to cache http requests. caching is off by default. See ?pygbif.caching for all the details :issue:`52` :issue:`56` via @nleguillarme
- made note in docs that if you are trying to get the same behavior as the GBIF website for name searching, species.name_backbone is likely what you want :issue:`55` thanks @qgroom
- for parameters that expect a bool, convert them to lowercase strings internally before doing HTTP requests
- pygbif is Python 3 only now :issue:`19`
- Gains maps module with maps.map method for working with the GBIF maps API :issue:`41` :issue:`49`
- Gains new module utils with one method wkt_rewind :issue:`46` thanks @aubreymoore for the inspiration
- Fixed bug in registry.installations: typo in one of the parameters identifierTyp instead of identifierType :issue:`48` thanks @data-biodiversity-aq
- Link to GitHub issues from Changelog 🎉
- Fix a occurrence download test :issue:`47`
- Much more thorough docs :issue:`25`
- Download methods much improved :issue:`16` :issue:`27` thanks @jlegind @stijnvanhoey @peterdesmet !
- MULTIPOLYGON now supported in geometry parameter :issue:`35`
- Fixed docs for occurrences.get, and occurrences.get_verbatim, occurrences.get_fragment and demo that used occurrence keys that no longer exist in GBIF :issue:`39`
- Added organizations method to registry module :issue:`12`
- Added remainder of datasets methods: registry.dataset_search (including faceting support :issue:`37`) and registry.dataset_suggest, for the /dataset/search and /dataset/suggest routes, respectively :issue:`40`
- Added remainder of species methods: species.name_lookup (including faceting support :issue:`38`) and species.name_usage, for the /species/search and /species routes, respectively :issue:`18`
- Added more tests to cover new methods
- Changed species.name_suggest to give back data stucture as received from GBIF. We used to parse out the classification data, but for simplicity and speed, that is left up to the user now.
- start parameter in species.name_suggest, occurrences.download_list, registry.organizations, registry.nodes, registry.networks, and registry.installations, changed to offset to match GBIF API and match usage throughout remainder of pygbif
- Added many new occurrence.search parameters, including repatriated, kingdomKey, phylumKey, classKey, orderKey, familyKey, genusKey, subgenusKey, establishmentMeans, facet, facetMincount, facetMultiselect, and support for facet paging via **kwargs :issue:`30` :issue:`34`
- Fixes to **kwargs in occurrence.search so that facet parameters can be parsed correctly and requests GET request options are collected correctly :issue:`36`
- Added spellCheck parameter to occurrence.search that goes along with the q parameter to optionally spell check full text searches :issue:`31`
- Added variable types throughout docs
- Changed default limit value to 300 for occurrences.search method
- tox now included, via @xrotwang :issue:`20`
- Added more registry methods :issue:`11`
- Started occurrence download methods :issue:`16`
- Added more names methods :issue:`18`
- All requests now send user-agent headers with requests and pygbif versions :issue:`13`
- Bug fix for occurrences.download_get :issue:`23`
- Fixed bad example for occurrences.get :issue:`22`
- Fixed wheel to be universal for 2 and 3 :issue:`10`
- Improved documentation a lot, autodoc methods now
- Fixed distribution for pypi
- First release