Changes in version 1.10 (2019-09-21)                  

    o   Avoid nameclashes with mlr3 by moving ParamHelpers to Suggests

    o   Force using httr 1.4.1 and curl 4.1

                  Changes in version 1.9 (2019-08-07)                   

    o   Fixed bug: Uploaded runs now also contain default hyperparamters.

    o   Fixed bug with multiple values for ignore.attribute.

    o   The status argument of listOML* functions can now also take the value 'all'.

    o   Allow to uploading data sets with multiple targets (e.g. multilabel tasks).

    o   New functions: uploadOMLTask, uploadOMLStudy, makeOMLStudy, getOMLStudy, listOMLStudies, extractOMLStudyIds.

                  Changes in version 1.8 (2018-03-02)                   

    o   Fixed bug in listOMLSetup.

    o   Update to curl 3.1 which fixes the error 'transfer closed with outstanding read data remaining' for large files.

    o   New function extractOMLStudyIds added, which directly extracts ids of data sets, tasks, flows and runs.

                  Changes in version 1.7 (2017-10-17)                   

    o   listOMLTask, listOMLFlow, listOMLRuns do not return the tag field anymore.

    o   New function listOMLSetup added, which enables extracting hyperparameter configurations of specific setups/flows.

    o   New function chunkOMLlist added in order to automatically do chunked listing requests.

                  Changes in version 1.6 (2017-08-14)                   

    o   Fixes some config issues.

                  Changes in version 1.5 (2017-08-11)                   

    o   New functions to list and get all information w.r.t. studies: getOMLStudy, listOMLStudies.

    o   API key not needed anymore for getOML* and listOML* functions.

                  Changes in version 1.4 (2017-06-22)                   

    o   Bugfix: Error message "Start tag expected, '<' not found" for getOMLTask and getOMLDataSet was fixed.

    o   as.data.frame methods for OMLRunParList, OMLDataSet and OMLTasks objects were added.

    o   listOMLRunEvaluations now allows to get results w.r.t. a single evaluation measure ('evaluation.measure' arg) which speeds up the request.

    o   listOMLRuns and listOMLRunEvaluations now also supports a vector of ids for task.id, flow.id and uploader.id.

    o   getOMLRun(run.id, only.xml = TRUE) can now be used to get the run without the predictions arff file (which is faster, especially when you are only interested in, e.g. getting the hyperparameters of a run).

                  Changes in version 1.3 (2017-03-30)                   

    o   Bugfixes.

    o   Updated citation.

    o   Add html vignette and update its content.

    o   listOMLTasks and listOMLDataSets now additionally show a message when the limit of results is reached.

    o   listOML* functions return an empty data frame when no results are available.

    o   listOMLRunEvaluations now returns additional columns for the flow (flow version, flow source and learner name).

    o   runTask now allows to set the 'models' option to FALSE so that resulting objects will be smaller.

                  Changes in version 1.2 (2017-02-07)                   

    o   Add support for multilabel datasets and tasks.

    o   Replace download.file with httr::GET.

    o   Add mlr 2.10 dependency (we internally use mlr::mergeBenchmarkResults and mlr::makePrediction now).

                  Changes in version 1.1 (2016-11-23)                   

    o   Setting default cache directory on package loading (fixes winbuilder).

    o   Replace internal regexps with stringi functions.
	OpenML1.0 (Release date: 2016-11-12):

    o   First submission to CRAN.