OpenML

Returns a Vector of NamedTuples. Receives an HTTP.Message.response that has an ARFF file format in the body of the Message.

OpenML.load(id)

Load the OpenML dataset with specified id, from those listed on the OpenML site.

Returns a "row table", i.e., a Vector of identically typed NamedTuples. A row table is compatible with the Tables.jl interface and can therefore be readily converted to other compatible formats. For example:

using DataFrames
rowtable = OpenML.load(61);
df = DataFrame(rowtable);
df2 = coerce(df, :class=>Multiclass)

Returns a list of all data qualities in the system.

• 271 - Unknown dataset. Data set with the given data ID was not found (or is not shared with you).
• 272 - No features found. The dataset did not contain any features, or we could not extract them.
• 273 - Dataset not processed yet. The dataset was not processed yet, features are not yet available. Please wait for a few minutes.
• 274 - Dataset processed with error. The feature extractor has run into an error while processing the dataset. Please check whether it is a valid supported file. If so, please contact the API admins.

Returns the qualities of a dataset.

• 360 - Please provide data set ID
• 361 - Unknown dataset. The data set with the given ID was not found in the database, or is not shared with you.
• 362 - No qualities found. The registered dataset did not contain any calculated qualities.
• 363 - Dataset not processed yet. The dataset was not processed yet, no qualities are available. Please wait for a few minutes.
• 364 - Dataset processed with error. The quality calculator has run into an error while processing the dataset. Please check whether it is a valid supported file. If so, contact the support team.
• 365 - Interval start or end illegal. There was a problem with the interval start or end.

Returns a list of all data qualities in the system.

• 412 - Precondition failed. An error code and message are returned
• 370 - No data qualities available. There are no data qualities in the system.

Returns information about a dataset. The information includes the name, information about the creator, URL to download it and more.

• 110 - Please provide data_id.
• 111 - Unknown dataset. Data set description with data_id was not found in the database.
• 112 - No access granted. This dataset is not shared with you.

List datasets, possibly filtered by a range of properties. Any number of properties can be combined by listing them one after the other in the form '/data/list/{filter}/{value}/{filter}/{value}/...' Returns an array with all datasets that match the constraints.

Any combination of these filters /limit/{limit}/offset/{offset} - returns only {limit} results starting from result number {offset}. Useful for paginating results. With /limit/5/offset/10, results 11..15 will be returned.

Both limit and offset need to be specified. /status/{status} - returns only datasets with a given status, either 'active', 'deactivated', or 'inpreparation'. /tag/{tag} - returns only datasets tagged with the given tag. /{dataquality}/{range} - returns only tasks for which the underlying datasets have certain qualities. {dataquality} can be dataid, dataname, dataversion, numberinstances, numberfeatures, numberclasses, numbermissingvalues. {range} can be a specific value or a range in the form 'low..high'. Multiple qualities can be combined, as in 'numberinstances/0..50/number_features/0..10'.

• 370 - Illegal filter specified.
• 371 - Filter values/ranges not properly specified.
• 372 - No results. There where no matches for the given constraints.
• 373 - Can not specify an offset without a limit.