swap to VoPho phonemizer

app.py

@@ -5,14 +5,13 @@ import re
 import numpy as np
 from scipy.io.wavfile import write
 import nltk
+from VoPho.engine import Phonemizer
 
 nltk.download('punkt')
 from nltk.tokenize import word_tokenize
 
 import torch
 
-import phonemizer  # en-us
-
 INTRO = """
 <style>
 
@@ -94,13 +93,6 @@ theme = gr.themes.Soft(
     block_background_fill='*neutral_50'
 )
 
-# eventually swap to something else
-global_phonemizer = phonemizer.backend.EspeakBackend(language='en-us',
-                                                     preserve_punctuation=True,
-                                                     with_stress=True,
-                                                     language_switch="remove-flags",
-                                                     tie=False)
-
 
 def split_and_recombine_text(text, desired_length=200, max_length=300):
     """Split text it into chunks of a desired length trying to keep sentences intact."""
@@ -173,11 +165,12 @@ def split_and_recombine_text(text, desired_length=200, max_length=300):
 
     return rv
 
+engine = Phonemizer()
 
 def text_to_phonemes(text):
     text = text.strip()
     print("Text before phonemization: ", text)
-    ps = global_phonemizer.phonemize([text])
+    ps = engine.phonemize(text)
     print("Text after phonemization: ", ps)
     ps = word_tokenize(ps[0])
     ps = ' '.join(ps)

requirements.txt

@@ -23,4 +23,5 @@ gradio
 spaces
 gruut
 txtsplit
-scipy
+scipy
+VoPho