khmer-g2p-ipa / app.py
seanghay's picture
update
d1c3c7d verified
raw
history blame
1.15 kB
import gradio as gr
import phonetisaurus
from collections import namedtuple
import re
import shlex
import tempfile
import subprocess
def phonemizer(text: str):
if not text:
return ""
env = phonetisaurus.guess_environment()
model_path = "./model.fst"
with tempfile.NamedTemporaryFile(suffix=".txt", mode="w+") as temp_file:
print(text, file=temp_file)
temp_file.seek(0)
phonetisaurus_cmd = [
"phonetisaurus-apply",
"--model",
shlex.quote(str(model_path)),
"--word_list",
shlex.quote(str(temp_file.name)),
"--nbest",
str(1),
]
result_str = subprocess.check_output(phonetisaurus_cmd, env=env, universal_newlines=True)
phoneme = result_str.split('\t')[1].strip()
phoneme = re.sub(r'\s', "", phoneme)
return f"/{phoneme}/"
iface = gr.Interface(
title="Khmer Phonemizer",
description="Convert Khmer & English into Phonetic Symbols",
fn=phonemizer, inputs="text",
outputs="text",
allow_flagging="never",
)
iface.launch(show_api=False)