Encapsulate specification for feature extraction and transformation.
Attributes
- T
input record type to extract features from
- Companion:
- object
- Source:
- FeatureSpec.scala
- Graph
- Supertypes
- class Objecttrait Matchableclass Any
Members list
Value members
Concrete methods
Compose with another spec by applying a prepare function to input records first.
Compose with another spec by applying a prepare function to input records first.
Useful for reusing an existing spec for a different input record type.
Attributes
- S
input record type of the other spec
- f
function to prepare input records for the other spec
- spec
spec to compose with
- Source:
- FeatureSpec.scala
Cross feature values of two underlying transformers.
Cross feature values of two underlying transformers.
Attributes
- f
function to cross feature value pairs
- k
names of transformers to be crossed
- Source:
- FeatureSpec.scala
Extract features from an input collection.
Extract features from an input collection.
This is done in two steps, a reduce step over the collection to aggregate feature summary,
and a map step to transform values using the summary.
Attributes
- M
input collection type, e.g.
Array,List- input
input collection
- Source:
- FeatureSpec.scala
Extract features from an input collection using settings from a previous session.
Extract features from an input collection using settings from a previous session.
This bypasses the reduce step in extract and uses feature summary from settings exported
in a previous session.
Attributes
- M
input collection type, e.g.
Array,List- input
input collection
- settings
JSON settings from a previous session
- Source:
- FeatureSpec.scala
Extract features from individual records using settings from a previous session. Since the settings are parsed only once, this is more efficient and is recommended when the input is from an unbounded source, e.g. a stream of events or a backend service.
Extract features from individual records using settings from a previous session. Since the settings are parsed only once, this is more efficient and is recommended when the input is from an unbounded source, e.g. a stream of events or a backend service.
This bypasses the reduce step in extract and uses feature summary from settings exported
in a previous session.
Attributes
- settings
JSON settings from a previous session
- Source:
- FeatureSpec.scala
Extract features from an input collection using a partial settings from a previous session.
Extract features from an input collection using a partial settings from a previous session.
This bypasses the reduce step in extract and uses feature summary from settings exported
in a previous session.
Attributes
- M
input collection type, e.g.
Array,List- input
input collection
- settings
JSON settings from a previous session
- Source:
- FeatureSpec.scala
Extract features from individual records using partial settings. Since the settings are parsed only once, this is more efficient and is recommended when the input is from an unbounded source, e.g. a stream of events or a backend service.
Extract features from individual records using partial settings. Since the settings are parsed only once, this is more efficient and is recommended when the input is from an unbounded source, e.g. a stream of events or a backend service.
This bypasses the reduce step in extract and uses feature summary from settings exported
in a previous session.
Attributes
- settings
JSON settings from a previous session
- Source:
- FeatureSpec.scala
Creates a new FeatureSpec with only the features that respect the given predicate.
Creates a new FeatureSpec with only the features that respect the given predicate.
Attributes
- predicate
Function determining whether or not to include the feature
- Source:
- FeatureSpec.scala
Add an optional field specification.
Add an optional field specification.
Attributes
- A
extracted feature type
- default
default for missing values
- f
function to extract feature
Option[A]from recordT- t
Transformer for extracted feature
A- Source:
- FeatureSpec.scala
Add a required field specification.
Add a required field specification.
Attributes
- A
extracted feature type
- f
function to extract feature
Afrom recordT- t
Transformer for extracted feature
A- Source:
- FeatureSpec.scala