doperables
Operation that is able to take dataframe and split its timestamp column to many columns containing timestamp parts.
Can create a Transformer of type T based on a DataFrame.
Evaluates a DataFrame.
This is a trait that lets define a method for loading model, and if it fails, it falls back to default load implementation.
This is a trait that lets define a method for loading model, and if it fails, it falls back to default load implementation. It is especially useful for supporting two spark versions.
Metric value.
Metric value.
name of the metric (e.g. RMSE).
value.
MultiColumnEstimator is a ai.deepsense.deeplang.doperables.Estimator that can work on either a single column or multiple columns.
MultiColumnEstimator is a ai.deepsense.deeplang.doperables.Estimator that can work on either a single column or multiple columns. Also, it can also work in-place (by replacing columns) or not (new columns will be appended to a ai.deepsense.deeplang.doperables.dataframe.DataFrame).
Parent type of the returned transformers.
The type of the returned transformer when working on multiple columns.
The type of the returned transformer when working on a single column.
This class is returned from an Estimator when multiple column mode was selected during fit.
This class is returned from an Estimator when multiple column mode was selected during fit. A model created in this way can be used to transform multiple columns ONLY. It holds a sequence of SingleColumnModels.
MultiColumnTransformer is a ai.deepsense.deeplang.doperables.Transformer that can work on either a single column or multiple columns.
MultiColumnTransformer is a ai.deepsense.deeplang.doperables.Transformer that can work on either a single column or multiple columns. Also, it can also work in-place (by replacing columns) or not (new columns will be appended to a ai.deepsense.deeplang.doperables.dataframe.DataFrame). When not working in-place and when working with a single column one has to specify output column's name. When working with multiple columns and in not in-place mode one has to specify output column names' prefix.
Sorts the input Dataframe according to selected columns.
Wrapper for creating deeplang Estimators from spark.ml Estimators.
Wrapper for creating deeplang Estimators from spark.ml Estimators. It is parametrized by model and estimator types, because these entities are tightly coupled.
We assume that every ml.Estimator and SparkModelWrapper has a no-arg constructor.
Type of wrapped ml.Model
Type of wrapped ml.Estimator
Type of used model wrapper
Wrapper for creating deeplang Evaluators from spark ml Evaluators.
Wrapper for creating deeplang Evaluators from spark ml Evaluators. It is parametrized by evaluator type.
Type of wrapped ml.evaluation.Evaluator
Wrapper for creating deeplang Transformers from spark.ml Models.
Wrapper for creating deeplang Transformers from spark.ml Models. It is parametrized by model and estimator types, because these entities are tightly coupled.
Every SparkModelWrapper should have a no-arg constructor.
type of wrapped ml.Model
type of wrapped ml.Estimator
SparkMultiColumnEstimatorWrapper represents an estimator that is backed up by a Spark estimator.
SparkMultiColumnEstimatorWrapper represents an estimator that is backed up by a Spark estimator. The wrapped estimator (and it's model) must operate on a single column. SparkMultiColumnEstimatorWrapper allows to create (basing on a Spark estimator) an estimator that is capable of working on both single columns and multiple columns. Depending on the mode it returns different types of models (SingleColumnModel or MultiColumnModel). Both of the returned models have to have a common ancestor ("the parent model").
Spark model used in Single- and MultiColumnModel.
The wrapped Spark estimator.
A common ancestor of the single and multi column models produced by the SparkMultiColumnEstimatorWrapper.
Type of the model returned when the estimator is working on a single column.
Type of the single column estimator.
Type of the model returned when the estimator is working on multiple columns.
This class creates a Deeplang MultiColumnTransformer from a Spark ML Transformer that has inputCol and outputCol parameters.
This class creates a Deeplang MultiColumnTransformer from a Spark ML Transformer that has inputCol and outputCol parameters. We assume that every Spark Transformer has a no-arg constructor.
Wrapped Spark Transformer type
This class creates a Deeplang Transformer from a Spark ML Transformer.
This class creates a Deeplang Transformer from a Spark ML Transformer. We assume that every Spark Transformer has a no-arg constructor.
Wrapped Spark transformer type
Able to transform a DataFrame into another DataFrame.
Able to transform a DataFrame into another DataFrame. Can have mutable parameters.
Provides helper methods for automatic conversion of double columns to vector columns.
Operation that is able to take dataframe and split its timestamp column to many columns containing timestamp parts. Client can choose timestamp parts from set: {year, month, day, hour, minutes, seconds} using parameters. Choosing $part value will result in adding new column with name: {original_timestamp_column_name}_$part of IntegerType containing $part value. If a column with that name already exists {original_timestamp_column_name}_$part_N will be used, where N is first not used Int value starting from 1.