Character Filters

Character filters preprocess raw text before tokenization.

Base Class

CharacterFilter

Abstract base class for implementing custom character filters.

Available Character Filters

ICUNormalizer

Unicode text normalization using ICU normalization forms.

Normalization Forms:

• NFKC (default): Compatibility Composition - Recommended for search
• NFC: Canonical Composition - For text storage and display
• NFD: Canonical Decomposition - For accent-insensitive search
• NFKD: Compatibility Decomposition - Maximum normalization

Best for: General search applications, normalizing ligatures and full-width characters, accent-insensitive search