return engine

app.py

@@ -1,20 +1,11 @@
 import tempfile
 from typing import Optional
 import gradio as gr
-import numpy as np
-from TTS.api import TTS
-from huggingface_hub import hf_hub_download
+from engine import TextToSpeech
 import subprocess
 
 
 MAX_TXT_LEN = 100
-REPO_NAME="DigitalUmuganda/Kinyarwanda_YourTTS"
-hf_hub_download(repo_id=REPO_NAME,filename="config.json")
-hf_hub_download(repo_id=REPO_NAME,filename="SE_checkpoint.pth.tar")
-hf_hub_download(repo_id=REPO_NAME,filename="config_se.json")
-hf_hub_download(repo_id=REPO_NAME,filename="model.pth")
-hf_hub_download(repo_id=REPO_NAME,filename="speakers.pth")
-hf_hub_download(repo_id=REPO_NAME,filename="conditioning_audio.wav")
 
 def generate_audio(text):
     if len(text) > MAX_TXT_LEN:
@@ -31,13 +22,7 @@ def generate_audio(text):
     # )
     # if synthesizer is None:
     #     raise NameError("model not found")
-    # tts = TTS(model_path="Kinyarwanda_YourTTS/model.pth",
-    #         config_path="Kinyarwanda_YourTTS/config.json",
-    #         tts_speakers_file="Kinyarwanda_YourTTS/speakers.pth",
-    #         encoder_checkpoint="Kinyarwanda_YourTTS/SE_checkpoint.pth.tar",
-    #         encoder_config="Kinyarwanda_YourTTS/config_se.json",)
-    # wav = tts.tts(text, speaker_wav="kinyarwanda_YourTTS/conditioning_audio.wav")
-    # return wav
+    tts_engine= TextToSpeech()
     text1 = subprocess.check_output("ls", shell=True)+ subprocess.check_output("ls", shell=True)
     text2 = text1.decode("utf-8")
     return text2

engine.py

@@ -0,0 +1,24 @@
+import numpy as np
+from TTS.api import TTS
+from huggingface_hub import hf_hub_download
+import subprocess
+
+REPO_NAME="DigitalUmuganda/Kinyarwanda_YourTTS"
+class TextToSpeech():
+
+    def __init__(self) -> None:
+        hf_hub_download(repo_id=REPO_NAME,filename="config.json")
+        hf_hub_download(repo_id=REPO_NAME,filename="SE_checkpoint.pth.tar")
+        hf_hub_download(repo_id=REPO_NAME,filename="config_se.json")
+        hf_hub_download(repo_id=REPO_NAME,filename="model.pth")
+        hf_hub_download(repo_id=REPO_NAME,filename="speakers.pth")
+        hf_hub_download(repo_id=REPO_NAME,filename="conditioning_audio.wav")
+    
+    def run_tts(self,text):
+        tts = TTS(model_path="Kinyarwanda_YourTTS/model.pth",
+            config_path="Kinyarwanda_YourTTS/config.json",
+            tts_speakers_file="Kinyarwanda_YourTTS/speakers.pth",
+            encoder_checkpoint="Kinyarwanda_YourTTS/SE_checkpoint.pth.tar",
+            encoder_config="Kinyarwanda_YourTTS/config_se.json",)
+        wav = tts.tts(text, speaker_wav="kinyarwanda_YourTTS/conditioning_audio.wav")
+        return wav