File size: 829 Bytes
a26008b
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
{
  "version": 1,
  "language": "it",
  "_comment": "Per-bundle language facts (data-driven; see C# LanguageRules / ADR-004). stripTrailingTerminator: strip a single trailing . ! ? before the WordPiece model and re-attach it after (the Italian bundle is brittle to a trailing period). elision.nonElidingArticles: full (un-elided) articles the tokenizer treats as glued-elision prefixes (gli'avvocato -> gli avvocato); overrides ItalianElision.DefaultNonElidingArticles. Absent file => None (stripTerminator false, default article list).",
  "stripTrailingTerminator": true,
  "elision": {
    "nonElidingArticles": ["gli", "le", "i", "il", "la", "lo", "una", "uno"]
  },
  "clauseConjunctions": ["ma", "e", "ed", "o", "oppure", "però", "mentre", "quindi", "poi", "perché", "anzi", "dunque", "infatti", "tuttavia", "perciò"]
}