Fix language flags generated by espeak-ng phonemizer (#1801)

* fix language flags generated by espeak-ng phonemizer

* Style

* Updated language flag regex to consider all language codes alike
This commit is contained in:
rbaraglia
2022-08-07 13:57:40 +02:00
committed by GitHub
parent 8c645080ac
commit 75ac9e3f0c
@@ -1,4 +1,5 @@
import logging
import re
import subprocess
from typing import Dict, List
@@ -163,6 +164,13 @@ class ESpeak(BasePhonemizer):
# dealing with the conditions descrived above
ph_decoded = ph_decoded[:1].replace("_", "") + ph_decoded[1:]
# espeak-ng backend can add language flags that need to be removed:
# "sɛʁtˈɛ̃ mˈo kɔm (en)fˈʊtbɔːl(fr) ʒenˈɛʁ de- flˈaɡ də- lˈɑ̃ɡ."
# phonemize needs to remove the language flags of the returned text:
# "sɛʁtˈɛ̃ mˈo kɔm fˈʊtbɔːl ʒenˈɛʁ de- flˈaɡ də- lˈɑ̃ɡ."
ph_decoded = re.sub(r"\(.+?\)", "", ph_decoded)
phonemes += ph_decoded.strip()
return phonemes.replace("_", separator)