DefaultParamsWriter

class pyspark.ml.util.DefaultParamsWriter(instance: Params)[source]

Specialization of MLWriter for Params types

Class for writing Estimators and Transformers whose parameters are JSON-serializable.

New in version 2.3.0.

Methods

extractJsonParams(instance, skipParams)

option(key, value)

Adds an option to the underlying MLWriter.

overwrite()

Overwrites if the output path already exists.

save(path)

Save the ML instance to the input path.

saveImpl(path)

save() handles overwriting and then calls this method.

saveMetadata(instance, path, sc[, …])

Saves metadata + Params to: path + “/metadata”

session(sparkSession)

Sets the Spark Session to use for saving/loading.

Attributes

sc

Returns the underlying SparkContext.

sparkSession

Returns the user-specified Spark Session or the default.

Methods Documentation

static extractJsonParams(instance: Params, skipParams: Sequence[str]) → Dict[str, Any][source]
option(key: str, value: Any)pyspark.ml.util.MLWriter

Adds an option to the underlying MLWriter. See the documentation for the specific model’s writer for possible options. The option name (key) is case-insensitive.

overwrite()pyspark.ml.util.MLWriter

Overwrites if the output path already exists.

save(path: str) → None

Save the ML instance to the input path.

saveImpl(path: str) → None[source]

save() handles overwriting and then calls this method. Subclasses should override this method to implement the actual saving of the instance.

static saveMetadata(instance: Params, path: str, sc: pyspark.context.SparkContext, extraMetadata: Optional[Dict[str, Any]] = None, paramMap: Optional[Dict[str, Any]] = None) → None[source]

Saves metadata + Params to: path + “/metadata”

  • class

  • timestamp

  • sparkVersion

  • uid

  • paramMap

  • defaultParamMap (since 2.4.0)

  • (optionally, extra metadata)

Parameters
extraMetadatadict, optional

Extra metadata to be saved at same level as uid, paramMap, etc.

paramMapdict, optional

If given, this is saved in the “paramMap” field.

session(sparkSession: pyspark.sql.session.SparkSession) → RW

Sets the Spark Session to use for saving/loading.

Attributes Documentation

sc

Returns the underlying SparkContext.

sparkSession

Returns the user-specified Spark Session or the default.