Portuguese API

Portuguese G2P provides rule-based phoneme conversion for Brazilian Portuguese, designed for Kokoro TTS models.

Main Class

class kokorog2p.pt.PortugueseG2P(language: str = 'pt-br', use_espeak_fallback: bool = False, use_spacy: bool = False, spacy_model: str = 'pt_core_news_sm', mark_stress: bool = True, affricate_ti_di: bool = True, expand_abbreviations: bool = True, enable_context_detection: bool = True, dialect: str = 'br', version: str = '1.0', **kwargs: Any)

Bases: G2PBase

Brazilian Portuguese G2P converter using rule-based phonemization.

This class provides grapheme-to-phoneme conversion for Brazilian Portuguese text using Portuguese orthographic rules.

Example:

>>> g2p = PortugueseG2P()
>>> tokens = g2p("Olá, como está?")
>>> for token in tokens:
...     print(f"{token.text} -> {token.phonemes}")

__init__(language: str = 'pt-br', use_espeak_fallback: bool = False, use_spacy: bool = False, spacy_model: str = 'pt_core_news_sm', mark_stress: bool = True, affricate_ti_di: bool = True, expand_abbreviations: bool = True, enable_context_detection: bool = True, dialect: str = 'br', version: str = '1.0', **kwargs: Any) → None

Initialize the Portuguese G2P converter.

Args:

language: Language code (default: ‘pt-br’). use_espeak_fallback: Reserved for future espeak integration. use_spacy: Whether to use spaCy for tokenization and POS tagging.

Defaults to False to preserve existing behavior.

spacy_model: spaCy Portuguese model package to load when use_spacy=True

(e.g., “pt_core_news_sm”, “pt_core_news_md”, “pt_core_news_lg”).

mark_stress: Whether to mark primary stress with ˈ. affricate_ti_di: Whether to affricate /t d/ before /i/ (Brazilian feature). expand_abbreviations: Whether to expand common abbreviations. enable_context_detection: Context-aware abbreviation expansion. dialect: “br” for Brazilian, “pt” for European Portuguese.

Affects number pronunciation (dezesseis vs dezasseis)

version: Target model version.

__call__(text: str) → list[GToken]

Convert text to a list of tokens with phonemes.

Args:

text: Input text to convert.

Returns:

List of GToken objects with phonemes assigned.

property nlp: object

Lazily initialize spaCy.

property spacy_tokenizer: SpacyTokenizer

Lazily initialize the spaCy tokenizer.

lookup(word: str, tag: str | None = None) → str | None

Look up a word’s phonemes.

Args:

word: The word to look up. tag: Optional POS tag (ignored for Portuguese).

Returns:

Phoneme string or None.

phonemize(text: str) → str

Convert text to phonemes.

Args:

text: Input text to convert.

Returns:

Phoneme string.

get_target_model() → str

Get the target Kokoro model variant for this G2P instance.

Returns:

Model identifier: version string (“1.1” or “1.0”).

Examples

from kokorog2p.pt import PortugueseG2P

g2p = PortugueseG2P(language="pt-br")
tokens = g2p("Olá mundo!")

for token in tokens:
    print(f"{token.text} -> {token.phonemes}")

Phonology Features

Brazilian Portuguese phonology includes:

• 7 oral vowels (a, e, ɛ, i, o, ɔ, u) with open/closed e/o variants

• 5 nasal vowels (ã, ẽ, ĩ, õ, ũ)

• Nasal diphthongs (ãw̃, õj̃, etc.)

• Palatalization: lh [ʎ], nh [ɲ], x/ch [ʃ]

• Affrication: t+i [ʧ], d+i [ʤ] (Brazilian Portuguese feature)

• Sibilants: s [s/z], x [ʃ], z [z]

• Liquids: r [ʁ/x/h] (varies by dialect), rr [ʁ/x], single r [ɾ]

• No θ sound (unlike European Portuguese)