fr.french-persons-ne.regex Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of uima-tokens-ner Show documentation

Uploads all artifacts belonging to configuration ':archives'

The newest version!

import fr.univnantes.termsuite.types.TermSuiteTypeSystem;
use fr.univnantes.termsuite.types.WordAnnotation;
java-matcher: IsInFirstNameLexicon;
java-matcher: IsInPersonTitlesLexicon;
set inline = false;

matcher FirstName: [IsInFirstNameLexicon];
matcher Title: [IsInPersonTitlesLexicon];
matcher LastName: [Capitalized | TitleCased];
matcher De as D: [text == "de"];
matcher Du as D: [text == "du"];
matcher La as L: [text == "la"];
matcher Le as L: [text == "le"];
matcher Des as D: [text == "des"];

rule "simple": FirstName LastName;
rule "noble1": FirstName Du LastName;
rule "noble2": FirstName De La LastName;
rule "noble3": FirstName De Le LastName;
rule "noble4": FirstName Des LastName;
rule "title": Title LastName;