g2p-barranquenho 0.1.2a3

Rule-based grapheme-to-phoneme converter for Barranquenho

g2p_barranquenho - Barranquenho Phonemizer

This repository provides a rule-based Grapheme-to-Phoneme (G2P) converter for the Barranquenho language.

Barranquenho is an Ibero-Romance language (often classified as a dialect) spoken in the municipality of Barrancos, Portugal, which shares many features with the nearby Spanish dialects of Extremadura and Andalusia.

The rules implemented in the phonemize function are based on the three official 2025 normative sources listed in the Sources section below.

Architecture

phonemize builds on the shared orthography2ipa pronunciation lattice rather than a private tokenizer:

• Tokenization is done by the language-agnostic orthography2ipa.phonetok.PhonetokTokenizer — a maximal-munch trie over the Barranquenho grapheme set (the consonant multigraphs tch ch nh lh rr ss plus the single letters). There is no hand-rolled digraph detection or index arithmetic.
• Vowel identity is decided by orthography2ipa.vowels.is_orthographic_vowel (which recognises the ẽ ĩ ũ nasal vowels), so the engine shares one vowel inventory with the spec instead of keeping a private copy.
• Barranquenho-specific rules — coda-conditioned nasalization, nasal diphthongs, stress-conditioned vowel quality, ⟨qu⟩/⟨gu⟩ glide elision, ⟨s⟩ → [h] coda aspiration, and the three-way ⟨x⟩ rule — are each expressed as an orthography2ipa.rescorer.LatticeRescorer re-costing the shared per-grapheme lattice slots, instead of a parallel private cascade.

Nasal vowels are emitted as a single atomic phoneme carrying the IPA/Unicode nasalization diacritic COMBINING TILDE (U+0303), e.g. [ɐ̃]. The coda ⟨m⟩/⟨n⟩ is absorbed into the preceding nasal vowel rather than surfacing as its own segment, so a nasal vowel is one list item, not a vowel followed by a separate tilde.

🚀 Usage

The core functionality is provided by the phonemize function, which takes a word (string) and returns a list of International Phonetic Alphabet (IPA) phonemes.

from g2p_barranquenho import phonemize

for word in [...]:
    phonemes = phonemize(word)
    print(word, phonemes)
    
# paraba ['p', 'ɐ', 'ɾ', 'a', 'b', 'ɐ']
# pássaru ['p', 'a', 's', 'ɐ', 'ɾ', 'u']
# biba ['b', 'i', 'b', 'ɐ']
# cahtelu ['k', 'ɐ', 'h', 't', 'e', 'l', 'u']
# boca ['b', 'o', 'k', 'ɐ']
# ambu ['ɐ̃', 'b', 'u']
# cantá ['k', 'ɐ̃', 't', 'a']
# manhán ['m', 'ɐ', 'ɲ', 'ɐ̃']
# que ['k', 'ɨ']
# aquí ['ɐ', 'k', 'i']

Sources

The phoneme rules are derived from the coordinated normative suite produced under the Programa de Preservação e Valorização da Língua e Cultura Barranquenhas:

• Convenção Ortográfica do Barranquenho (June 2025). Câmara Municipal de Barrancos / II Congresso Barranquenho working group. Available at: https://cm-barrancos.pt/upload_files/1/3/Noticias/2025/II%20Congresso%20Barranquenho/Conven%C3%A7%C3%A3o%20Ortogr%C3%A1fica%20do%20Barranquenho%20junho%202025%20final.pdf
• Gramática Básica de Barranquenho (July 2025). Maria Filomena Gonçalves (Universidade de Évora / CIDEHUS-UÉ / FCT / Cátedra UNESCO em Património Imaterial e Saber-Fazer Tradicional), María Victoria Navas (Universidad Complutense de Madrid / CLUL), Victor M. Diogo Correia (Universidade de Évora / CIDEHUS-UÉ / FCT). Universidade de Évora, 1ª edição. ISBN 978-972-778-464-6.
• Dicionário de Barranquenho (2025). Maria Filomena Gonçalves (Universidade de Évora / CIDEHUS-UÉ / FCT / Cátedra UNESCO), María Victoria Navas (Universidad Complutense de Madrid / CLUL), Vera Ferreira (CIDLeS — Centro de Documentação Linguística e Social). Universidade de Évora, 1ª edição. ISBN 978-972-778-460-8.