Type Members

1.  trait AnchorExtractor[T] extends AnchorExtractorBase[T] with SparkRowExtractor

   Provides feature values based on some "raw" data element

   Provides feature values based on some "raw" data element

   T

   raw data type

2.  trait AnchorExtractorBase[T] extends Serializable

   base class for feature extractor class, e.g., AnchorExtractor, SimpleAnchorExtractorSpark and GenericAnchorExtractorSpark

   base class for feature extractor class, e.g., AnchorExtractor, SimpleAnchorExtractorSpark and GenericAnchorExtractorSpark

   T

   row data type

3.  class AutoTensorizableTypes extends AnyRef

   Mappings from Feathr FeatureTypes to their "auto-tensorized" tensor type.

   Mappings from Feathr FeatureTypes to their "auto-tensorized" tensor type. In other words, when the feature owner does not specify a tensor type, Feathr will pick one based on the feature's FeatureType.

   These mappings define how Feathr will represent features having types NUMERIC, CATEGORICAL, etc., as Quince tensors.

   Annotations

   @Experimental()

4.  trait CanConvertToAvroRDD extends AnyRef

   If an AnchorExtractor only works on a Avro record, it should extends this trait, and use convertToAvroRdd to do a one-time batch conversion of DataFrame to RDD of their choice.

   If an AnchorExtractor only works on a Avro record, it should extends this trait, and use convertToAvroRdd to do a one-time batch conversion of DataFrame to RDD of their choice. convertToAvroRdd will be called by Feathr engine before calling getKeyFromRow() and getFeaturesFromRow() in AnchorExtractor.

5.  class CoercingTensorData extends TensorData

   Coerce types in any TensorData to the one requested by the client.

6.  class ColumnUtils extends AnyRef

   Utility function to build internal flat array representation based on the types of the input.

7.  case class DateParam(startDate: Option[String] = None, endDate: Option[String] = None, dateOffset: Option[String] = None, numDays: Option[String] = None) extends Product with Serializable

   Date can be specified in two ways: 1.

   Date can be specified in two ways: 1. startTime plus endTime, or 2. dateOffset plus numDays in the case of 2, the startTime will be "today_in_pst - dateOffset - numDays + 1", endTime will be "today_in_pst - dateOffset" This case class is used both by feature and observation data.

   startDate

   in yyyyMMdd or yyyyMMddHH

   endDate

   in yyyyMMdd or yyyyMMddHH

   dateOffset

   non-negative integer (either in days or hours, resolution will determine if it is in days or hours). Used to calculate the endTime.

   numDays

   positive integer (either in days or hours, resolution will determine if it is in days or hours). Used together with dateOffset to calculate the startTime

8.  class Equal extends AnyRef

   Compare whether two tensor data are equal.

   Compare whether two tensor data are equal.

   Note: order does not matter here, and we assume that there is no duplicate rows in left & right tensor data

   Return true if the tensors are equal.

9.  class ErasedEntityTaggedFeature extends Serializable

   A com.linkedin.feathr.common.TaggedFeatureName whose String tags have been "erased", i.e.

   A com.linkedin.feathr.common.TaggedFeatureName whose String tags have been "erased", i.e. replaced with integers. The mapping between the integer identifiers and their original string names is unknown by this class and must be maintained externally.

10.  class Experimental extends Annotation with Annotation with ClassfileAnnotation

    Annotations

    @Documented() @Retention()

11.  class FeathrJacksonScalaModule extends Module with JacksonModule with IteratorModule with EnumerationModule with OptionModule with SeqModule with IterableModule with TupleModule with MapModule with SetModule with UntypedObjectDeserializerModule
12.  sealed abstract final class FeatureAggregationType extends Enum[FeatureAggregationType]

    Feature aggregation types

13.  class FeatureDependencyGraph extends AnyRef

    A dependency graph for feature anchors and feature derivations.

    A dependency graph for feature anchors and feature derivations.

    Purpose 1: Given a list of features' dependencies and which features are anchored, build a graph that can determine which features are reachable and can resolve features' transitive dependencies on demand. Purpose 2: Given a list of features with entity key bindings (a.k.a. key tags), provide fully expanded list of all transitive dependencies including entity key bindings. E.g. if feature2 depends on feature1, and the request was for [ (a1):feature2, (a2):feature2 ], then the expanded dependencies would be: [ (a1):feature2, (a2):feature2, (a1):feature1, (a2):feature1]

    Annotations

    @InternalApi()

14.  trait FeatureDerivationFunction extends FeatureDerivationFunctionBase

    Produces a derived feature given some input features

15.  abstract class FeatureDerivationFunctionBase extends Serializable

    this is the common trait for derivation function for Feathr internal use if Feathr user need to implement customized derivation, please use FeatureDerivationFunction or FeatureDerivationFunctionSpark

16.  class FeatureError extends AnyRef

    An error associated with a particular feature containing an error msg and the corresponding FeatureErrorCode

17.  sealed abstract final class FeatureErrorCode extends Enum[FeatureErrorCode]

    Error code associated with a feature request

18.  trait FeatureExtractor[SOURCE_DATA] extends AnyRef

    A feature extractor that extracts FeatureValue from source data.

    A feature extractor that extracts FeatureValue from source data. The class must have 0-argument constructor so Feathr engine can instantiate the class with 0-argument constructor. We expect this to replace Extractor eventually.

    WARNING: This interface is in beta testing and is not approved for public use yet.

    Annotations

    @Experimental()

19.  class FeatureInfo extends AnyRef

    information about a feature in the feathr feature join and feature generation output dataframe

20.  class FeatureRef extends Serializable

    Represents a fully-qualified reference to a feature.

21.  class FeatureTypeConfig extends Serializable

    This class encapsulates the type definition for a feature.

    This class encapsulates the type definition for a feature. This implementation supports Feathr's high level semantic type system (e.g. NUMERIC, BOOLEAN, VECTOR) and the new Tensor type system which includes additional metadata such as category, dimensions, shape, value type and etc.

    When the feature type annotation from the feature definition is processed, the following scenarios are expected: 1. No type annotated --> (FeatureType=UNSPECIFIED, TensorType=null) 2. High level semantic NON-TENSOR feature types --> (FeatureType=<some value>, TensorType=<mapped from feature type via auto-TZ>) 3. Parameterized Tensor type annotated --> (FeatureTypes = TENSOR, TensorType = <some value>)

    Annotations

    @JsonDeserialize()

22.  class FeatureTypeConfigDeserializer extends JsonDeserializer[FeatureTypeConfig]

    Config deserializer for FeatureType config.

    Config deserializer for FeatureType config. It mainly supports parsing two type configs: 1. Simple type config: type with a enum. Example: type: BOOLEAN 2. Type config with more specifications. For example: DENSE_VECTOR with size: type: { type: DENSE_VECTOR shape: [10] doc: "documentation" } TENSOR type with more details: type: { type: DENSE_TENSOR shape: [10, 10] dimensionType: [INT, INT] valType: FLOAT doc: "documentation" }

23.  sealed abstract final class FeatureTypes extends Enum[FeatureTypes]

    An Enum that defines the supported feature types in feathr.

    An Enum that defines the supported feature types in feathr. This list will be extended in the future.

24.  class FeatureValue extends Serializable with KryoSerializable

    Describes the basic representation of a feature value in Feathr.

    Describes the basic representation of a feature value in Feathr. This is the OLD FeatureValue API. Feathr is migrating to a new FeatureValue API defined at com.linkedin.feathr.common.value.FeatureValue. This implementation of the old API is a wrapper around the new API, with lots of glue to maintain compatibility.

    SERIALIZATION: For use with Spark RDD API, this class is serializable both by Java Object Serialization and by Kryo. Note that because the serialized representation is based on NTV (name-term-value), serialization will NOT work for tensor features having rank greater than 1.

    MUTABILITY: For legacy compatibility, instances of this class are mutable under the following conditions:

    1. When constructed with zero-argument constructor 2. When constructed with Map<String,Float> constructor 3. When constructed specifically as a term-vector using any createStringTermVector factory methods 4. When deserialized with Kryo or Java-Object Serialization When mutable, instances may be mutated by:
    2. Using modifier methods the Map objects returned by .getValue or .getAsTermVector 2. Using modifier methods on ._value public member field. 3. Using the .put method. All mutability support in FeatureValue is DEPRECATED.

    NOTE: This class will be deprecated soon in favor of com.linkedin.feathr.common.value.FeatureValue

25.  class FeatureVariableResolver extends SimpleValueResolver

    FeatureVariableResolver takes a FeatureValue object for member variable during MVEL expression evaluation, and then resolve the value for that variable.

26.  class GenericTypedTensor extends TypedTensor

    This is the base tensor class, may be created from the basic properties, from a feature or operator applied to another tensor

27.  class Hasher extends AnyRef

    Generate hash for a tensor based on column types and values stored in the tensor.

    Generate hash for a tensor based on column types and values stored in the tensor. This hash should be same for all tensor instances with same column types and values even if they are implemented differently under the hood.

28.  class Header extends AnyRef

    header of the result of the Feathr feature join and feature generation API

29.  class InternalApi extends Annotation with Annotation with ClassfileAnnotation

    Annotations

    @Documented() @Retention()

30.  case class JoiningFeatureParams(keyTags: Seq[String], featureName: String, dateParam: Option[DateParam] = None, timeDelay: Option[String] = None, featureAlias: Option[String] = None) extends Product with Serializable

    All parameters relevant to a feature which is requested to be joined in the join config.

    All parameters relevant to a feature which is requested to be joined in the join config.

    keyTags

    sequence of keyTags of the feature.

    featureName

    name of the feature.

    dateParam

    time parameters to restrict the size of the joining feature's data. User can set a date range on the feature data.

    timeDelay

    the override time delay parameter, which would override the global simulate time delay. This simulates a delay on observation data's timestamp by subtracting this value from observation data's timestamp while joining this specific feature.

    featureAlias

    feature alias name, this would be the output of the feature.

31.  trait ParameterizedFeatureExtractor[SOURCE_DATA, PARAM] extends FeatureExtractor[SOURCE_DATA]

    This extension of the FeatureExtractor rely on extra parameters to extract features.

    This extension of the FeatureExtractor rely on extra parameters to extract features. The expectation is that the Feathr engine will call init when constructing an instance of this Extractor and implementation of this interface is responsible for maintaining any state for the parameters as needed.

32.  trait Params extends Serializable

    parameter map used in various of classes, such as anchor extractor, derived feature, source key extractor, etc.

    parameter map used in various of classes, such as anchor extractor, derived feature, source key extractor, etc.

    This helps user to pass arbitrary parameter into the related classes when needed.

33.  implicit class RichConfig extends AnyRef
34.  trait SparkRowExtractor extends AnyRef

    An extractor trait that provides APIs to transform a Spark GenericRowWithSchema into feature values

35.  class TaggedFeatureName extends Serializable

    A tuple of (key tag, feature name)

    A tuple of (key tag, feature name)

    Essentially this is a feature name annotated with info on how we plan to query it.

36.  class TaggedFeatureUtils extends AnyRef
37.  final class TensorUtils extends AnyRef

    INTERNAL debugging utility functions.

    INTERNAL debugging utility functions.

    In the future, this class may be deprecated.

    Annotations

    @InternalApi()

38.  trait TypedTensor extends AnyRef

    TypedTensor builds on top of @link{TensorType} and @{TensorData} and can be generated from tensor operators that change the shape.