pygoruut 0.6.4

Text-to-IPA converter and phonetic translator for Python, powered by the Goruut phonemization engine

pygoruut

Getting started

from pygoruut.pygoruut import Pygoruut

pygoruut = Pygoruut()

print(str(pygoruut.phonemize(language="EnglishAmerican", sentence="fast racing car")))

# Prints: fˈæst ɹˈeɪsɪŋ kˈɑɹ

# Now, convert it back

print(str(pygoruut.phonemize(language="EnglishAmerican", sentence="fˈæst ɹˈeɪsɪŋ kˈɑɹ", is_reverse=True)))

# Prints: fast racing car

ℹ️ For English, we recommend using EnglishBritish or EnglishAmerican instead of English. These dialect-specific models use high-quality Kokoro Misaki dictionaries and produce better results, especially for reversing IPA back to text.

---

Uyghur language, our highest quality language

print(str(pygoruut.phonemize(language="Uyghur", sentence="قىزىل گۈل ئاتا")))

# Prints: qizil gyl ʔɑtɑ

# Now, convert it back

print(str(pygoruut.phonemize(language="Uyghur", sentence="qizil gyl ʔɑtɑ", is_reverse=True)))

# Prints: قىزىل گۈل ئاتا

The quality of translation varies across the 136 supported languages.

---

Advanced Use

Multi-lingual sentence handling

Use comma (,) separated languages in language. The first language is the preferred language:

print(pygoruut.phonemize(language="EnglishBritish,Slovak", sentence="hello world ahojte notindictionary!!!!"))

# Prints: həlˈoʊ wˈɜɹld aɦɔjcɛ ŋətandəktɪnˈɑːɪ!!!!

---

Numerics handling (English, Arabic)

print(str(pygoruut.phonemize(language="EnglishBritish", sentence="100 bottles")))

# Prints: wˈʌn hˈʌndɹəd bˈɒtəlz

---

Homograph handling (Hebrew3)

print(str(pygoruut.phonemize(language="Hebrew3", sentence="השרים ביקשו מהשרים לפתוח את הדלתות של בית השרים.")))

# Prints: hasaʁˈim bikʃˈu mehasaʁˈim liftˈoaχ ʔˈat hadlatˈot ʃˈel bˈet hasaʁˈim.

---

No punctuation

print(str(pygoruut.phonemize(language="EnglishBritish", sentence="hello world!!!!", is_punct=False)))

# Prints: həlˈəʊ wˈɜːld

---

Force a specific version

You can pin a specific version. It will translate all words in the same way forever:

from pygoruut.pygoruut import Pygoruut

pygoruut = Pygoruut(version='v0.6.2')

---

Use an online inference api

You can use an inference api. The model will not be downloaded:

from pygoruut.pygoruut import Pygoruut

pygoruut = Pygoruut(api='https://hashtron.cloud')

---

Configure a model download directory for faster startup

To cache models in a user-specified directory:

from pygoruut.pygoruut import Pygoruut

pygoruut = Pygoruut(writeable_bin_dir='/home/john/')

To cache in the user's home subdirectory .goruut:

from pygoruut.pygoruut import Pygoruut

pygoruut = Pygoruut(writeable_bin_dir='')