Struct StreamingG2P 

pub struct StreamingG2P { /* private fields */ }

The core engine for streaming Text-to-Phoneme (G2P) conversion.

StreamingG2P processes text incrementally, expanding abbreviations/numbers and resolving phonemes dynamically. It supports multi-language text using internal heuristics.

Implementations

impl StreamingG2P

pub fn new(lang: Language) -> Result<Self>

Constructs a single-language streaming G2P engine.

pub fn with_languages( languages: &[Language], default_language: Language, ) -> Result<Self>

Constructs a multi-language streaming G2P engine.

It intelligently detects the language of the incoming stream from the provided languages list. If detection fails or is ambiguous, it falls back to the default_language.

pub fn new_session(&self) -> StreamingG2pSession

Creates a new isolated phonemization session.

The session must be mutated with push_text and finalized with finish.

pub fn push_text( &self, session: &mut StreamingG2pSession, text: &str, ) -> Result<Vec<G2pToken>>

Pushes a chunk of text into the streaming g2p processor.

Yields any phonemes that have been fully resolved (typically at word boundaries). Partial words or unresolved expansions are buffered internally within the session.

pub fn finish(&self, session: &mut StreamingG2pSession) -> Result<Vec<G2pToken>>

Finalizes the stream and flushes any pending text buffered in the session.

This should be called exactly once at the end of the text stream to guarantee the final phonemes are emitted.