com.workfusion.vds.sdk.api.nlp.fe

Interface FeatureExtractor<T extends Element>

Type Parameters:

T - determines what kinds of elements to perform feature extraction on

All Known Implementing Classes:

AreFollowingTokensCapitalizedBeforeNerFE, BaseKeywordsInElementFE, BaseNerInFocusElementFE, FocusElementKeywordsFE, FocusElementKeywordsFrequencyFE, GenericPythonFE, IsFirstNerInFocusElementFE, IsFocusElementExactSameToNerFE, IsFocusElementWithNerFE, IsLastNerInFocusElementFE, LowerCaseWordshapeFE, PositionInCoveringFE, PreviousElementSameLineKeywordsMatchFE, SimilarityToKeywordsFE

public interface FeatureExtractor<T extends Element>

The FeatureExtractor interface represents the component that performs feature extraction. It accept document for processing and element to perform feature extraction on. Implementation of that interface can have some lifecycle methods annotated by follow annotations:

OnInit - method annotated by this annotation will be invoked after constructor

OnDocumentStart - method annotated by this annotation will be invoked after constructor and init methods, but before feature extraction. It will be invoked one time for one Document.

OnDocumentComplete - method annotated by this annotation will be invoked once for one Document after all features extraction. OnDocumentStart and OnDocumentComplete executed once for one document in one thread. Both these annotations used on methods to prepare/finalize some data from Document. extract(Document, Element) method executed besides this 2 steps and can be multi threaded. All data prepared in OnDocumentStart step should be read-only in extract method.

OnDestroy - method annotated by this annotation will be invoked once for one class instance after all another events. The method annotated with Destroy is typically used to release resources that it has been holding.

Also feature extractors support follow TYPE annotations:

DependsOn - to predefine array of element types that need to be defined in document before feature extraction usage.

FeatureName - to define feature name.

Indexes - to define indexes for performance optimization.

Since:

8.5.7

Author:

Alena Zaitsava, Artem Strok, Alexey Gaponik, Heorhi Stsefanovich

Method Summary

Modifier and Type	Method and Description
Collection<Feature>	extract(Document document, T element) Analyze giving document structure and extract feature for specified element.

Method Detail

extract

Collection<Feature> extract(Document document,
                            T element)

Analyze giving document structure and extract feature for specified element.

Parameters:

document - document to proceed

element - annotated element in document that need to be analyzed for feature extraction

Returns:

collection of extracted features