io.redlink.nlp.regex.ner

Class NamedRegexDetector

java.lang.Object

io.redlink.nlp.regex.ner.RegexNamedEntityFactory

io.redlink.nlp.regex.ner.NamedRegexDetector

Direct Known Subclasses:

CsvNamedPatternNerDetector

public abstract class NamedRegexDetector
extends RegexNamedEntityFactory

Simplifies the Implementation of a RegexNamedEntityFactory in cases where

1. all created NamedEntities do use a single Token.Type
2. the created NamedEntities do use the NamedPattern#getName() as Token#getValue()

Subclasses just need to implement the loadPatterns() Method that is called once and is expected to provide the list of Regex patterns.

The acceptMatch(String, MatchResult) provides a callback that allows to filter out unwanted matches. The default implementation will filter out all blank matches

Author:

Rupert Westenthaler

Nested Class Summary

Nested classes/interfaces inherited from class io.redlink.nlp.regex.ner.RegexNamedEntityFactory

RegexNamedEntityFactory.NamedPattern

Field Summary

Fields inherited from class io.redlink.nlp.regex.ner.RegexNamedEntityFactory

log

Constructor Summary

Constructors

Constructor and Description
NamedRegexDetector(NerTag type)

Method Summary

Modifier and Type	Method and Description
protected boolean	acceptMatch(String name, MatchResult match) If the match for the Pattern with the parsed name should be accepted.
protected RegexNerProcessor.NamedEntity	createNamedEntity(String name, MatchResult match) Creates a token for the parsed MatchResult originating from the RegexNamedEntityFactory.NamedPattern with the parsed name
protected List<RegexNamedEntityFactory.NamedPattern>	getRegexes(SpanCollection section, String lang) Getter for the RegexNamedEntityFactory.NamedPattern to be used by the RegexNerProcessor
NerTag	getType()
void	init()
protected abstract Map<String,List<RegexNamedEntityFactory.NamedPattern>>	loadPatterns()

Methods inherited from class io.redlink.nlp.regex.ner.RegexNamedEntityFactory

process

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

NamedRegexDetector

public NamedRegexDetector(NerTag type)

Method Detail

getType

public NerTag getType()

init

@PostConstruct
public final void init()
                               throws IOException

Throws:

IOException

loadPatterns

protected abstract Map<String,List<RegexNamedEntityFactory.NamedPattern>> loadPatterns()
                                                                                throws IOException

Throws:

IOException

getRegexes

protected List<RegexNamedEntityFactory.NamedPattern> getRegexes(SpanCollection section,
                                                                String lang)

Description copied from class: RegexNamedEntityFactory

Getter for the RegexNamedEntityFactory.NamedPattern to be used by the RegexNerProcessor

Specified by:

getRegexes in class RegexNamedEntityFactory

Parameters:

section - the section of an AnalyzedText to be analyzed with the returned patterns

lang - the language of the parsed text section

Returns:

the list of RegexNamedEntityFactory.NamedPattern or an empty list if none

createNamedEntity

protected RegexNerProcessor.NamedEntity createNamedEntity(String name,
                                                          MatchResult match)

Description copied from class: RegexNamedEntityFactory

Creates a token for the parsed MatchResult originating from the RegexNamedEntityFactory.NamedPattern with the parsed name

Specified by:

createNamedEntity in class RegexNamedEntityFactory

Parameters:

name - the name of the RegexNamedEntityFactory.NamedPattern

match - the MatchResult

Returns:

the RegexNerProcessor.NamedEntity or null if no Token was created.

acceptMatch

protected boolean acceptMatch(String name,
                              MatchResult match)

If the match for the Pattern with the parsed name should be accepted. The default implementation accepts all none blank MatchResult.group()

Parameters:

name - the name of the pattern

match - the match

Returns:

true if a Token should be created for this match. Otherwise false