Add `espeak-ng` phonemizer for Tigrinya

[yanapo]

Hello there!

In the latest release, espeak-ng support has been extended to Tigrinya among a few other languages. So, it is now possible to generate IPA transcriptions for this language:

espeak-ng --version
espeak-ng  -v ti "ዛንትኡ ግና ነዊሕ እዩ።" -qx
>>> z,anyt'y?u g'yna n'@wiC ?'yju

It has been added to phonemizer==3.3.0 too:

from phonemizer import phonemize

sentence = "ምልክታተይ ዳርጋ ጠፊኡ ኣሎ። መዓስ እየ ነቲ መድሃኒት ምውሳድ ከቋርጾ ዘለኒ፧"
print(phonemize(
        text=sentence, language='ti', strip=True, prepend_text=False, preserve_punctuation=True, with_stress=True,
        tie=False, language_switch='keep-flags', words_mismatch='ignore', njobs=4
    ))
>>> mˌɨlɨkɨtˈatəj darˈɨɡa t`əfˈiʔu ʔˈalo mˈəʕas ʔˈɨjə nˈəti mˌədɨhˈanit mɨwˈɨsad kˌək`warˈɨtso zəlˈəni

Will it be possible to update the backend phonemizer for Tigrinya to espeak-ng? I know it is possible to override transphone by generating your own IPA transcription and then passing it to the model with the input_is_phones flag but it would feel nice to be inclusive :)

Cheers!