The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
This is a maintenance release to ensure package compatibility with future versions of R.
This is a maintenance release.
Bugfixes:
ListProjects
now has NULL default values for limit
and offset
arguments to maintain backwards compatibility. This fixes compatibility issues with versions of DataRobot before 9.x.The datarobot
package is now dependent on R >= 3.5. This is a maintenance release.
New Features:
mode = AutopilotMode.Comprehensive
.Enhancements:
RequestFeatureImpact
now accepts a rowCount
argument, which will change the sample size used for Feature Impact calculations.datarobot:::UploadData
now takes an optional argument fileName
.ListProjects
now accepts limit
and offset
arguments, which allows users to retrieve additional projects.Bugfixes:
datarobot:::UploadData
(i.e. SetupProject
) as well as datarobot:::UploadPredictionDataset
to fail with the error No method asJSON S3 class: form_file
.datarobot
package with suppressPackageStartupMessages()
will now suppress all messages.API Changes:
ListProjects
and as.data.frame.projectSummaryList
no longer return fields related to recommender models, which were removed in v2.5.0.SetTarget
now sets autopilot mode to Quick
by default. Additionally, when Quick
is passed, the underlying /aim
endpoint will no longer be invoked with Auto
.Deprecated and Defunct:
quickrun
argument is removed from the function SetTarget
. Users should set mode = AutopilotMode.Quick
instead.Dependency Changes:
datarobot
package is now dependent on R >= 3.5 due to changes in the updated “Introduction to DataRobot” vignette.AmesHousing
package for updated “Introduction to DataRobot” vignette.MASS
package.Documentation Changes:
lendingclub.com
in vignettes due to issues with the CRAN URL checker.This is a maintenance release that was superseded by v2.18.4.
This release fixes the test suite to conditionally use the stubthat
package, which as of 2022-04-17 is no longer available on CRAN (but is still available in the Microsoft snapshot archive.) This package is listed as a Suggests dependency, and the codebase is updated to reflect this.
Bugfixes:
BuildPath
helper function.This release fixes a breaking change in the client around ListProjects
due to the removal of Spark / H20 models from the DataRobot platform.
Enhancements:
GetProject
and ListProjects
now return all available output from the API.Bugfixes:
ModelCapability
has been properly exported.library(datarobot)
no longer throws an error when executed in an RStudio session.ListProjects()
no longer throws an error about “undefined columns selected.”API Changes:
scaleoutModelingMode
, scaleoutMaxTrainPct
, and scaleoutMaxTrainRows
. These attributes had appeared in the parameters of setTarget
and startProject
along with the responses from ListProjects
and GetProject
.Deprecated and Defunct:
scaleoutModelingMode
, scaleoutMaxTrainPct
, and scaleoutMaxTrainRows
are now removed.Dependency Changes: * Client documentation is now explicitly generated with Roxygen2 v7.1.2.
Documentation Changes:
NEWS
file was renamed to NEWS.md
and formatted as Markdown.GetDeploymentSettings
and UpdateDeploymentSettings
.extdata/Friedman1.csv
and updated vignettes dependent on that dataset.extdata/anomFrame.csv
as it was unused.This release brings the R Client to parity with DataRobot API v2.18 (DataRobot 5.2), but also includes a number of features from API v2.19 (DataRobot 5.3) as well as Anomaly Assessment, a DataRobot 7.1 feature.
New Features:
GetLiftCharts
and GetAllLiftCharts
functions. This is valid only for regression models that are not time-aware.FORECAST_DISTANCE_AVG
has been added as BlendMethod$FORECAST_DISTANCE_AVG
.GetDeploymentServiceStats
retrieves metrics that track deployment utilization and performance, while GetDeploymentAccuracy
retrieves metrics that track the accuracy of a deployment’s predictions. GetDeploymentServiceStatsOverTime
and GetDeploymentAccuracyOverTime
will track changes to those metrics over a specified time interval.SubmitActuals
can now be used to submit data about actual results from a deployed model, which can be used to calculate accuracy metrics.CloneProject
. The clone will be post-EDA1 and ready for setting targets and modeling options.CreateCalendar
now supports series-specific events via the multiSeriesIdColumn
argument. An example of a series-specific event: some but not all stores being affected by a holiday.GetDeploymentAssociationId
and UpdateDeploymentAssociationId
can be used to manage a deployment’s association ID for use with SubmitActuals
and the Deployment Accuracy functions.GetDeploymentSettings
can be used to retrieve any and all settings related to a deployed model. UpdateDeploymentSettings
will allow you to make piecemeal changes as well. The convenience functions GetDeploymentDriftTrackingSettings
and GetDeploymentAssociationId
use these methods internally.RequestTransferableModel
now has a predictionIntervalsSize
parameter.InitializeAnomalyAssessment
initializes an anomaly assessment insight for the specified subsetListAnomalyAssessmentRecords
retrieves recordsGetAnomalyAssessmentExplanations
retrieves shap explanationsGetAnomalyAssessmentPredictionsPreview
retrieves predictions previewDeleteAnomalyAssessmentRecord
deletes recordsEnhancements:
RequestNewDatetimeModel
.modelId
field on the GetModelJob
or ListModelJobs
response objects.recommendedFeaturelistId
to the Blueprint
response object. If absent, there is no recommended feature list for this blueprint.Model
S3 class now exposes the modelNumber
field. This field is also exposed in the responses to GetFrozenModel
, GetDatetimeModel
, GetBlenderModel
, and GetRatingTableModel
.GetModelCapabilities
has been extended to return supportsCodeGeneration
, supportsShap
, and other newly-added capabilities. See ModelCapability
for more details.GetFeatureInfo
will now return descriptive statistics on summarized categorical features in the field keySummary
.ListDeployments
now supports sorting and searching the results using the new orderBy
and search
parameters.GetResidualsChart
and ListResidualsCharts
are now backwards-compatible with DataRobot 5.2, which does not return rowNumber.GetWordCloud
now includes a variable
field that represents the source of each ngram, as well as a class
field that represents values of the target class.GetPredictions
and Predict
when retrieving probabilities for large prediction datasets on multiclass projects, i.e. Predict(irisModel, largeDataset, type = "probability")
Bugfixes:
ListDeployments
will now return more than 20 deployments when available.ListPrimeModels
now returns an empty data frame when the API returns zero results, consistent with its documentation. Previously it would return an empty list. This response is also classed as dataRobotPrimeModels
.CreateCalendar
now terminates properly when DataRobot is unable to create the calendar. Previously, it would hang due to the R package not checking for the right error response.formatRFC3339Timestamp
now works for vectors of length > 1.API Changes:
CreateCalendar
and CreateRatingTable
is changed from file
to dataSource
to reflect that the functions can process data frames as well as CSV files.ProjectFromAsyncUrl
is replaced with ProjectFromJobResponse
; this change allowed us to simplify the package’s dependency on httr.Deprecated and Defunct:
BlendMethod$FORECAST_DISTANCE
is deprecated and will be removed in 2.19. Use BlendMethod$FORECAST_DISTANCE_ENET
instead.Dependency Changes:
testthat@>3.0.0
and devtools@>2.4.0
is now required. The test suites are being updated to meet testthat 3e requirements.Suggests: rex
as it is no longer needed for package development.Documentation Changes:
NEWS
file was renamed to NEWS.md
and formatted as Markdown.GetBlenderModel
and GetBlenderModelFromJobId
are now more consistent.StartAutopilot
and SetTarget
is clarified.curl
from Imports
since it was causing a NOTE when devtools::check_win_devel()
was run.New Features:
CreateDeployment
to create a deployment against a particular prediction server. Use ListPredictionServers
to list all the available prediction servers. Use GetDeployment
and ListDeployments
to see particular deployments that you have. You can delete a deployment with DeleteDeployment
.ReplaceDeployedModel
. Use ValidateReplaceDeployedModel
first to test that the deployment replacement is valid, if desired.GetDeploymentDriftTrackingSettings
to get drift tracking settings for a deployment. You can update the drift tracking using UpdateDeploymentDriftTrackingSettings
.GetFeatureAssociationMatrix
. Relative pairwise feature association statistics can be retrieved with GetFeatureAssociationMatrixDetails
.BatchFeaturesTypeTransform
.Enhancements:
doNotDerive
in the featureSettings
of CreateDatetimePartition
to disable DataRobot’s automatic time series feature engineering for a particular feature (e.g., so you can derive lags yourself manually).UploadComplianceDocTemplate
) using keyword tags.Bugfixes:
GetPredictionExplanationsRowsAsDataFrame
previously did not work with numeric labels. This has been fixed.API Changes:
Deprecated and Defunct:
defaultToAPriori
parameter in CreateDatetimePartitionSpecification
has been renamed to defaultToKnownInAdvance
. defaultToAPriori
has now been fully removed.aPriori
flag in the featureSettings
parameter in CreateDatetimePartitionSpecification
as been renamed to knownInAdvance
. aPriori
has now been fully removed.SetupProjectFromMySQL
, SetupProjectFromOracle
and SetupProjectFromPostgreSQL
have now been removed. Use SetupProjectFromDataSource
instead.GetTransferrableModel
, ListTransferrableModels
, UpdateTransferrableModel
, DeleteTransferrableModel
, DownloadTransferrableModel
, and UploadTransferrableModel
have been removed and replaced with their correctly spelled counterparts (GetTransferableModel
, ListTransferableModels
, UpdateTransferableModel
, DeleteTransferableModel
, DownloadTransferableModel
, and UploadTransferableModel
).Dependency Changes:
Documentation Changes:
New Features:
GetSeriesAccuracy
to retrieve the accuracy. You can also download it as a CSV with DownloadSeriesAccuracy
.Enhancements:
includePredictionIntervals = TRUE
in calls to Predict
. For each model, prediction intervals estimate the range of values DataRobot expects actual values of the target to fall within. They are similar to a confidence interval of a prediction, but are based on the residual errors measured during the backtesting for the selected model.ListPredictions
now returns metadata on prediction intervals. includesPredictionIntervals
is TRUE if there are prediction intervals in the predictions and FALSE
otherwise. predictionIntervals
specifies the size (in percent) of intervals or is NULL
if there are no intervals.GetDatetimeModel
). The new attributes are effectiveFeatureDerivationWindowStart
, effectiveFeatureDerivationWindowEnd
, forecastWindowStart
, forecastWindowEnd
, and windowsBasisUnit
.DownloadComplianceDocumentation
and GetSeriesAccuracy
now support a maxWait
parameter to customize the amount of time to wait before raising a timeout error.Deprecated and Defunct:
RecommendedModelType$Recommended
type for GetModelRecommendation
and GetRecommendedModel
has been removed and replaced with RecommendedModelType$RecommendedForDeployment
.Documentation Changes:
New Features:
StartTuningSession
for details.CreateCalendar
, GetCalendar
, ListCalendars
, UpdateCalendar
, and DeleteCalendar
.Share
for details.Share
for details.Enhancements:
UploadPredictionDataset
and UploadPredictionDatasetFromDataSource
will now return dataQualityWarnings
that mention any potential problems with the uploaded dataset.UploadPredictionDataset
and UploadPredictionDatasetFromDataSource
now have a parameter relaxKIAFeaturesCheck
. If TRUE
, uploaded datasets for time series projects will allow missing values for the Known in Advance features in the forecast window at prediction time.fractionPredictedAsPositive
, fractionPredictedAsNegative
, liftPositive
, and liftNegative
) with cumulative gains and lift data.filter
option to ListProjects
that supports filtering retrieval of project lists by name using the projectName
filter.GetCalendarFromProject
can be used to get the calendar associated with a project.StartProject
to quickly create a project from a data source.SetupProjectFromDataSource
.crossSeriesGroupByColumns
has been added to datetime partitioning to allow users the ability to indicate how to further split series in to related groups.Bugfixes:
windowBasisUnit
has been renamed to the correct windowsBasisUnit
.Deprecated and Defunct:
DeleteReasonCodes
, DeleteReasonCodesInitialization
, DownloadReasonCodes
, GetAllReasonCodesRowsAsDataFrame
, GetReasonCodesInitialization
, GetReasonCodesInitializationFromJobId
, GetReasonCodesMetadata
, GetReasonCodesMetadataFromJobId
, GetReasonCodesRows
, ListReasonCodesMetadata
, RequestReasonCodes
, and RequestReasonCodesInitialization
have all been removed and replaced with appropriately renamed functions and a new workflow. See GetPredictionExplanations
for more.SetupProjectFromMySQL
, SetupProjectFromOracle
, SetupProjectFromPostgreSQL
, SetupProjectFromHDFS
are now deprecated. They will be removed in v2.17. Use SetupProjectFromDataSource
instead.RequestPredictionsForDataset
has been renamed to RequestPredictions
. The original RequestPredictionsForDataset
has been removed.GetDatetimeModelObject
has been renamed to GetDatetimeModel
. The original GetDatetimeModelObject
has been removed.defaultToAPriori
parameter in CreateDatetimePartitionSpecification
has been renamed to defaultToKnownInAdvance
. defaultToAPriori
is now removed.aPriori
flag in the featureSettings
parameter in CreateDatetimePartitionSpecification
as been renamed to knownInAdvance
. aPriori
is now removed.GetTransferrableModel
, ListTransferrableModels
, UpdateTransferrableModel
, DeleteTransferrableModel
, DownloadTransferrableModel
, and UploadTransferrableModel
have all been deprecated and replaced with their correctly spelled counterparts (GetTransferableModel
, ListTransferableModels
, UpdateTransferableModel
, DeleteTransferableModel
, DownloadTransferableModel
, and UploadTransferableModel
). The misspelled versions will be removed in v2.17.StartAutopilot
has now been fully removed.Dependency Changes:
curl
at version 2.7 or higher is now required.Documentation Changes:
Enhancements:
SeriesID
, forecastPoint
, and forecastDistance
.GetDatetimePartition
now returns isCrossSeries
to indicate whether the datetime partition uses cross-series features.ScoreBacktests
now accepts a parameter wait = TRUE
to wait for job completion.Predict
and GetPredictions
no longer return positiveProbability
for non-binary problems.Predict
and GetPredictions
no longer return seriesId
for non-multiseries problems.Documentation Changes:
knownInAdvance
was defined in the featureSettings
in the “time series” vignette.Bugfixes:
New Features:
DownloadComplianceDocumentation
can be used to download compliance documentation. Compliance documentation also can be created with default or custom templates - use GetComplianceDocTemplate
to get particular templates and UploadComplianceDocTemplate
to use your own. See the vignette on “Compliance Documentation” for more information.Share
to share a data source or data store. Use ListSharingAccess
to see current access rights. Use UpdateAccess
for more complex access right modification operations.useCrossSeries = TRUE
in CreateDatetimePartitionSpecification
to enable.GetFeatureHistogram
.GetModelCapabilities
.Enhancements:
StartProject
and UpdateProject
that up the worker count can now set the worker count to "max"
, which uses the maximum available number of workers.fallbackToParentInsights
is now available as a parameter on all insights functions (GetRocCurve
, ListRocCurves
, GetLiftChart
, ListLiftCharts
, GetConfusionChart
, ListConfusionCharts
). When TRUE
, a frozen model with missing insights will attempt to retrieve the missing insight data from its parent model.windowBasisUnit
and setting it to "ROW"
.DataSubset$AllBacktests
for requesting the predictions for all backtest validation folds.Bugfixes:
StartTuningSession
.StartProject
without defining a project name, it now correctly uses the name of the variable passed (like SetupProject
) rather than erroneously just calling it “dataSource”.Deprecated and Defunct:
GetAllLiftCharts
has now been removed (use ListLiftCharts
instead).GetModelJobs
has now been removed (use ListModelJobs
instead).GetProjectList
has now been removed (use ListProjects
instead).GetAllRocCurves
has now been removed (use ListRocCurves
instead).RecommendedModelType$Recommended
type for GetModelRecommendation
and GetRecommendedModel
has been deprecated and replaced with RecommendedModelType$RecommendedForDeployment
. It will be removed in v2.16.PeriodicityTimeUnits
has been renamed to TimeUnits
. PeriodicityTimeUnits
still exists for backwards compatibility.Documentation Changes:
Predict
in the “Introduction to DataRobot” vignette was previously inaccurate. It has been fixed.RecommendedModelType
and GetModelRecommendation
now have more documentation about the model recommendation process.StartProject
throughout.New Features:
GetTuningParameters
and start a tuning session with StartTuningSession
. You can interactively iterate through all the parameters for a model using RunInteractiveTuning
. These advanced tuning features are currently generally available for Eureqa models. To use this feature with other model types, contact your CFDS for more information.Predict
can now be used to create predictions directly from a model and a test dataset, bypassing the need to UploadPredictionDataset
, RequestPredictions
, and GetPredictions
.ListPredictions
can be used to summarize all the different predictions available for a particular project, model, and/or prediction dataset.GetPredictionExplanations
now supports a single workflow to get prediction explanations (previously called reason codes, see “Deprecated and Defunct”) for a model and a dataset without needing all the various intermediary steps.GetPredictionDataset
can be used to get metadata on a particular prediction dataset.SetPredictionThreshold
.GetFeatureImpact
works like GetFeatureImpactForModel
, but will also request the feature impact if it has not already been requested.GetTrainingPredictionsForModel
retrieves training predictions for a given model object, requesting them in the process.StarModel
will star a model. UnstarModel
will unstar it. ToggleStarForModel
will toggle the star status. ListStarredModels
will list all the starred models for a particular project. Model objects will also have an isStarred
parameter returned to tell whether they are starred or not. (All models are unstarred by default.)DeleteFeaturelist
and DeleteModelingFeaturelist
can now be used to delete featurelists and modeling featurelists respectively.UpdateFeaturelist
can be used to change the name and description of a featurelist. UpdateModelingFeaturelist
works for modeling featurelists as well.Enhancements:
type = "raw"
in GetPredictions
(or Predict
) will return the raw dataframe of predictions metadata.ListModels
now can take an orderBy
parameter to sort the output list by metric
or samplePct
.ListModels
now can take a filter
parameter to filter output by samplePct
, name
, and/or isStarred
.StartProject
and SetupProject
can now work without a projectName
StartProject
can now take a workerCount
parameter to set the worker count for the project.StartProject
can now take wait = TRUE
to automatically wait for the autopilot to complete (thus making an explicit call to WaitForAutopilot
unnecessary).ProjectStage
can now be used to get a list of the available project stages.RequestMultiSeriesDetection
manually for a multiseries project.isBlocked
that specifies whether a job is blocked from execution because one or more dependencies have not yet been met.ListModelJobs
now returns the trainingRowCount
key.GetPredictions
now can get predictions using a projectId and a predictionId (see ListPredictions
) in addition to its prior ability to retrieve predictions using a predictionJobId
.GetFeaturelist
and GetModelingFeaturelist
) now return a created
value with the timestamp, a isUserCreated
value explaining whether or not the feature was created by a user (as opposed to DataRobot automation), numModels
showing how many models use the featurelist, and description
which gives a text description of the featurelist.dataRobotPredictionDataset
class in addition to being list
class.GetPredictions
now can get predictions using a projectId and a predictionId (see ListPredictions
) in addition to its prior ability to retrieve predictions using a predictionJobId
.GetFeaturelist
and GetModelingFeaturelist
) now return a created
value with the timestamp, a isUserCreated
value explaining whether or not the feature was created by a user (as opposed to DataRobot automation), numModels
showing how many models use the featurelist, and description
which gives a text description of the featurelist.GetDatetimePartition
now reports information on the number of “known in advance” features.GetDatetimePartition
reports if the partition was drawn from a time series project and/or a multiseries project.Bugfixes:
API Changes:
RequestMultiSeriesDetection
(which is now no longer necessary to invoke directly) now blocks until the multiseries request is complete and returns details about which multiseries columns were detected.Deprecated and Defunct:
DeleteReasonCodes
, DeleteReasonCodesInitialization
, DownloadReasonCodes
, GetAllReasonCodesRowsAsDataFrame
, GetReasonCodesInitialization
, GetReasonCodesInitializationFromJobId
, GetReasonCodesMetadata
, GetReasonCodesMetadataFromJobId
, GetReasonCodesRows
, ListReasonCodesMetadata
, RequestReasonCodes
, and RequestReasonCodesInitialization
have all been deprecated (and will be removed in v2.15). These functions have been replaced with appropriately renamed functions and a new workflow. See GetPredictionExplanations
for more.GetDatetimeModelObject
has been renamed to GetDatetimeModel
. The original GetDatetimeModelObject
has been deprecated and will be removed in v2.15.Documentation Changes:
Predict
workflow.GetRecommendedModel
instead of calculating the best mode instead of calculating the best model.Enhancements:
plot
on ListOfModels returns an error if the desired percent passed to pct
is not found.Bugfixes:
as.data.frame
to handle missing featurelist IDs when simple = FALSE
.as.data.frame
to handle a list of prediction datasets.as.data.frame
to handle a list of models when samplePct
is not set.plot
on ListOfModels to work with missing featurelist IDs.Deprecated and Defunct:
simple
parameter in as.data.frame
now produces an error instead of a warning.New Features:
StartProject
, combines both SetupProject
and SetTarget
into one function.GetMissingValuesReport
.UploadPredictionDatasetFromDataSource
to create a prediction dataset from a DataRobot data source (introduced in v2.11).Enhancements:
RequestNewModel
and passing ""
(an empty string) as the value for each monotonic constraint featurelist you wish to override.Bugfixes:
AutopilotMode$Quick
“mode” for SetTarget
had been broken and no longer triggered the quick mode (instead it ran the full autopilot). This has been fixed.Deprecated and Defunct:
Dependency Changes:
httr
at version 1.2.0 or higher is now required.New Features:
ListModelRecommendations
has been added to get all the model recommendations, GetModelRecommendation
can return a particular recommendation, and GetRecommendedModel
returns the particular model object corresponding with a particular recommendation.ListDrivers
and GetDriver
to get available drivers, CreateDataStore
to create a data store from a driver, and CreateDataSource
to create a data source from a data store.SetupProjectFromDataSource
.GetTimeSeriesFeatureDerivationLog
can now be used to retrieve a lot of information on details for derived features for time series projects. DownloadTimeSeriesFeatureDerivationLog
can download the log to a text document.Enhancements:
GetFeatureInfo
and ListFeatureInfo
now report targetLeakage
, specifying whether a feature is considered to have target leakage or not.CrossValidateModel
on your model object.ConnectToDataRobot
now works with environment variables. Set DATAROBOT_API_ENDPOINT
and DATAROBOT_API_TOKEN
to connect to DataRobot. Note that previously the R client unofficially used DataRobot_URL
and DataRobot_Token
as environment variables to facilitate connecting to DataRobot, but these variables are now no longer supported.Bugfixes:
as.data.frame
to handle missing featurelist IDs.API Changes:
UploadPredictionDataset
to support bulk predictions upload for time series projects.Bugfixes:
as.data.frame
to handle multiple featurelists.partitionKeyCols
parameter in CreateGroupPartition
to more clearly error if more than one partition key is passed.Deprecated and Defunct:
quickrun
parameter on SetTarget, the ability to use GetFeatureInfo
with feature IDs, the GetRecommendedBlueprints
function, GetModelObject
, GetAllModels
, GetBlueprintDocuments
, and the RequestPredictions
function.defaultToAPriori
parameter in CreateDatetimePartitionSpecification
is being deprecated and has been renamed to defaultToKnownInAdvance
. defaultToAPriori
will be fully removed in v2.15.aPriori
flag in the featureSettings
parameter in CreateDatetimePartitionSpecification
is being deprecated and has been renamed to knownInAdvance
. The aPriori
will be fully removed in v2.15.New features:
RequestModelDeployment
, get information on a specific deployment using GetModelDeployment
, and list information on all deployments across all projects via ListModelDeployments
. You can also get more information on the service health of a particular deployment using GetModelDeploymentServiceStatistics
or get the action log for a deployed model using GetModelDeploymentActionLog
.Enhancements:
maxTrainRows
, scaleoutMaxTrainPct
, and scaleoutMaxTrainRows
have been added to projects retrieved by GetProject
. maxTrainRows
specified the equivalent value to the existing maxTrainPct
as a row count. The scaleout fields can be used to see how far scaleout models can be trained on projects, which for projects taking advantage of scalable ingest may exceed the limits on the data available to non-scaleout blueprints.trainingRowCount
argument, specifying a desired amount of rows instead of a desired percentage of the dataset (via the current samplePct
parameter). Specifying model size by row count is recommended when the float precision of sample_pct could be problematic, e.g. when training on a small percentage of the dataset or when training up to partition boundaries. This new approach is available for RequestNewModel
, RequestFrozenModel
, and RequestSampleSizeUpdate
. RequestFrozenDatetimeModel
already had this feature.GetPredictions
now returns a more informative error message when the async service times out.featureSettings
attribute when setting the target or specifying datetime partitioning settings on time series projects. Any features not specified in the featureSettings
parameter will be assigned according to the defaultToAPriori
value.DatetimePartitioningSpecification
to fine-tune how time-series projects derive modeling features. treatAsExponential
can control whether data is analyzed as an exponential trend and transformations like log-transform are applied. differencingMethod
can control which differencing method to use for stationary data. periodicities
can be used to specify periodicities occurring within the data. All are optional and defaults will be chosen automatically if they are unspecified.SetTarget
.Bugfixes:
UploadPredictionDataset
and GetPredictions
. These functions have now been fully tested to handle data up to 1GB, and likely can handle more than that. If you run into issues, try incrementing the maxWait
parameter.ssl_verify: FALSE
in drconfig.yaml
to not verify SSL when connecting with DataRobot.DownloadRatingTable
when it meant to read DownloadTrainingPredictions
.reasonCodeId <- GetReasonCodesMetadataFromJobId(projectId, jobId)
when it should read reasonCodeId <- GetReasonCodesMetadataFromJobId(projectId, jobId)$id
.API Changes:
trainingRowCount
is available on non-datetime models as well as “rowCount” based datetime models. It reports the number of rows used to train the model (equivalent to samplePct
).GetParetoFront
to get the Pareto Front details and AddEureqaSolution
to add a new solution to the leaderboard.New features:
SetTarget
function, setting targetType = TargetType$Multiclass
. If DataRobot recognizes your data as categorical, and it has up to 10 classes, using multiclass will create a project that classifies which label the data belongs to.GetFeatureInfo
and ListFeatureInfo
now return the EDA summary statistics (i.e., mean, median, minum, maximum, and standard deviation) for features where this is available (e.g., numeric, date, time, currency, and length features). These summary statistics will be formatted in the same format as the data it summarizes.scaleoutModelingMode
when setting a project target. It can be used to control whether scaleout models appear in the autopilot and/or available blueprints. Scaleout models are only supported in the Hadoop enviroment with the corresponding user permission set.accuracyOptimizedBlueprints
when setting a project target. Accuracy optimized blueprints are longer running model blueprints that provide increased accuracy over the normal blueprints that run during autopilot.GetModelBlueprintChart
and model blueprint documentation via GetModelBlueprintDocumentation
. These are like regular blueprint charts and blueprint documentation, except for model blueprints, which are a reduced representation of the blueprint run by the model to only include the relevant branches actually executed by the model.Enhancements:
CreateDatetimePartitionSpecification
now includes the optional disableHoldout
flag that can be used to disable the holdout fold when creating a project with datetime partitioning.offset
and exposure
to allow specifying offset and exposure columns to apply to predictions generated by models within the project. See the user guide documentation in the web app for more information on offset and exposure columns.eventsCount
to allow specifying the events count column. See the user guide documentation in the webapp for more information on events count.ConnectToDataRobot
now supports an option sslVerify
that turns off SSL verification if set to FALSE.Bugfixes:
GetReasonCodesMetadataFromJobId
from being called with a project directly (instead of a project id).RequestNewModel
from being called when options(stringsAsFactors = TRUE)
is set.GetBlueprintDocuments
(now named GetBlueprintDocumentation
).Deprecated and Defunct:
GetRecommendedBlueprints
function, and the RequestPredictions
function were all originally planned to be deprecated in version 3.0. These features and functions will now be deprecated in v2.10 instead.GetBlueprintDocuments
is replaced by GetBlueprintDocumentation
and deprecated (and will be removed in v2.10).Documentation Changes:
modelwordcloud
package is now available on CRAN, so the documentation has been updated to reflect CRAN installation instructions.New features:
GetWordCloud
function.GetLiftCharts
and GetAllLiftCharts
function.GetRocCurve
and GetAllRocCurves
GetPredictJob',
GetModelJob,
GetJobfunctions. Any job can be retrieve via
GetJobwhich is less specific. Only prediction jobs can be retrieved with
GetPredictJoband only modeling jobs can be retrieved with
GetModelJob`.Enhancements:
GetModelParameters
now includes an additional key showing the coefficients for individual stages of multistage models (e.g. Frequency-Severity models).DatetimeModel
on a window of data, a timeWindowSamplePct
can be specified to take a uniform random sample of the training data instead of using all data within the window.Bugfixes:
Dependency Changes:
jsonlite
at version 1.0 or higher and curl
at version 1.1 or higher are now required.Deprecated and Defunct:
New features:
3
will be cast as the string 3
and the value 3.14
will also be cast as the string 3
. Further, the value -3.6
will become the string -3
. Missing values will still be recognized as missing.)DatetimePartitioningSpecification
using CreateDatetimePartition and CreateBacktestSpecification function and passing it into GenerateDatetimePartition, inspect the results and adjust as needed for the specific project dataset by adjusting the DatetimePartitioningSpecification
and re-generating, and then set the target by passing the final DatetimePartitioningSpecification
object to the partitioning_method parameter of SetTarget.Enhancements:
API Changes:
New features:
Enhancements:
NA
.Bugfixes:
API Changes:
Deprecated and Defunct:
Documentation Changes:
New features:
Enhancements:
Bugfixes:
API Changes:
Deprecated and Defunct:
Documentation Changes:
Enhancements:
Deprecated and Defunct:
New features:
Bugfixes
Bugfixes
API Changes
datarobot
.New features:
Enhancements:
holdoutLevel
to be NULL (which results in not sending the holdout level, in line with backend API changes to allow user partitions to be created without a holdout level).[
from objects of type listOfBlueprints, listOfFeaturelists, and listOfModels will now retain the appropriate type.Deprecated and Defunct:
Enhancements
Bugfixes
Deprecated & Defunct
Bugfixes
Bugfixes
Enhancements:
New features:
Bugfixes:
Deprecated and Defunct:
API Changes Summary:
Enhancements:
New Features:
Bugfixes: None
Deprecated and Defunct: None
API Changes: None
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.