Metadata-Version: 2.4
Name: lectura-g2p
Version: 4.1.3
Summary: Pipeline G2P complet du français — Tokeniseur + Formules + Phonémiseur + Groupes de lecture
Author-email: Max Carriere <admin@lectura.world>
License: AGPL-3.0-or-later
Project-URL: Homepage, https://www.lectura.world/developpement/modules/
Project-URL: Repository, https://github.com/maxcarriere/lectura-modules/tree/main/G2P-Pipeline
Project-URL: Issues, https://github.com/maxcarriere/lectura-modules/issues
Keywords: g2p,grapheme-to-phoneme,text-to-phoneme,french,nlp,phonemizer,phonetics,ipa,tts-frontend
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: GNU Affero General Public License v3 or later (AGPLv3+)
Classifier: Natural Language :: French
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Text Processing :: Linguistic
Requires-Python: >=3.10
Description-Content-Type: text/markdown
Requires-Dist: lectura-tokeniseur>=2.2
Requires-Dist: lectura-formules>=3.7
Requires-Dist: lectura-phonemiseur>=4.0
Requires-Dist: lectura-lexique>=1.4
Provides-Extra: aligneur
Requires-Dist: lectura-aligneur>=4.0; extra == "aligneur"

# lectura-g2p

Complete French grapheme-to-phoneme pipeline — text to IPA phonetics in one call.
Combines tokenizer, formula reader, and neural phonemizer.

*Pipeline G2P complet du francais — texte vers phonetique IPA en un appel.
Combine tokeniseur, lecteur de formules et phonemiseur neural.*

## Installation

```bash
pip install lectura-g2p              # pipeline complet
pip install lectura-g2p[aligneur]    # + alignement grapheme-phoneme et syllabation
```

## Usage

```python
from lectura_g2p import analyser, creer_engine

engine = creer_engine()

# Texte → phonetique IPA
result = analyser("Bonjour, il est 14h30.", engine=engine)
print(result["ipa"])
```

### Avec alignement et syllabation

```python
from lectura_g2p import analyser, creer_engine
from lectura_aligneur.lectura_aligneur import LecturaSyllabeur

engine = creer_engine()
syl = LecturaSyllabeur()

result = analyser("Bonjour !", engine=engine, aligner=syl)
for a in result["alignments"]:
    print(a.syllabes)
```

## Architecture

- **Couche 1** : `lectura-phonemiseur` — modele G2P neural BiLSTM (1.8 Mo ONNX INT8, 98.5% word accuracy)
- **Couche 2** : `lectura-g2p` — pipeline complet (tokeniseur + formules + phonemiseur)
- **4 backends** : API, ONNX Runtime, NumPy, Pure Python (zero dependance)

## Licence

AGPL-3.0-or-later — voir [LICENCE.txt](LICENCE.txt).

**Licence commerciale et modeles locaux disponibles** — contacter [admin@lectura.world](mailto:admin@lectura.world).
