Update app.py

app.py

@@ -1,12 +1,12 @@
-
-import tempfile ,os
-from TTS.config import load_config
+import tempfile
+import os
 import gradio as gr
-
+from TTS.config import load_config
 from TTS.utils.manage import ModelManager
 from TTS.utils.synthesizer import Synthesizer
+from TTS.utils.download import download_url
 
-MODEL_NAMES=[
+MODEL_NAMES = [
     "vits male1 (best)",
     "vits female (best)",
     "vits-male",
@@ -15,85 +15,85 @@ MODEL_NAMES=[
     "glowtts-female",
     "female tacotron2"
 ]
-MAX_TXT_LEN = 800
-model_path = os.getcwd() + "/best_model.pth"
-config_path = os.getcwd() + "/config.json"
-   
 
+MAX_TXT_LEN = 800
 
-from TTS.utils.download import download_url
-modelInfo=[
-    ["vits-male","best_model_65633.pth","config-0.json","https://huggingface.co/Kamtera/persian-tts-male-vits/resolve/main/"],
-    ["vits female (best)","checkpoint_48000.pth","config-2.json","https://huggingface.co/Kamtera/persian-tts-female-vits/resolve/main/"],
-    ["glowtts-male","best_model_77797.pth","config-1.json","https://huggingface.co/Kamtera/persian-tts-male-glow_tts/resolve/main/"],
-    ["glowtts-female","best_model.pth","config.json","https://huggingface.co/Kamtera/persian-tts-female-glow_tts/resolve/main/"],
-    ["vits male1 (best)","checkpoint_88000.pth","config.json","https://huggingface.co/Kamtera/persian-tts-male1-vits/resolve/main/"],
-    ["vits female1","checkpoint_50000.pth","config.json","https://huggingface.co/Kamtera/persian-tts-female1-vits/resolve/main/"],
-    ["female tacotron2","checkpoint_313000.pth","config-2.json","https://huggingface.co/Kamtera/persian-tts-female-tacotron2/resolve/main/"]
+# Define model information
+modelInfo = [
+    ["vits-male", "best_model_65633.pth", "config-0.json", "https://huggingface.co/Kamtera/persian-tts-male-vits/resolve/main/"],
+    ["vits female (best)", "checkpoint_48000.pth", "config-2.json", "https://huggingface.co/Kamtera/persian-tts-female-vits/resolve/main/"],
+    ["glowtts-male", "best_model_77797.pth", "config-1.json", "https://huggingface.co/Kamtera/persian-tts-male-glow_tts/resolve/main/"],
+    ["glowtts-female", "best_model.pth", "config.json", "https://huggingface.co/Kamtera/persian-tts-female-glow_tts/resolve/main/"],
+    ["vits male1 (best)", "checkpoint_88000.pth", "config.json", "https://huggingface.co/Kamtera/persian-tts-male1-vits/resolve/main/"],
+    ["vits female1", "checkpoint_50000.pth", "config.json", "https://huggingface.co/Kamtera/persian-tts-female1-vits/resolve/main/"],
+    ["female tacotron2", "checkpoint_313000.pth", "config-2.json", "https://huggingface.co/Kamtera/persian-tts-female-tacotron2/resolve/main/"]
 ]
 
-for d in modelInfo:
-    directory=d[0]
+# Download model files if not exist
+for model in modelInfo:
+    directory = model[0]
     if not os.path.exists(directory):
         os.makedirs(directory)
-    print("|> Downloading: ",directory)
-    download_url(
-        d[3]+d[1],directory,"best_model.pth"
-    )
-    download_url(
-        d[3]+d[2],directory,"config.json"
-    )
-def tts(text: str,model_name: str):
+    if not os.path.exists(os.path.join(directory, "best_model.pth")):
+        print(f"|> Downloading model: {directory}")
+        download_url(model[3] + model[1], directory, "best_model.pth")
+    if not os.path.exists(os.path.join(directory, "config.json")):
+        print(f"|> Downloading config: {directory}")
+        download_url(model[3] + model[2], directory, "config.json")
+
+def tts(text: str, model_name: str):
     if len(text) > MAX_TXT_LEN:
         text = text[:MAX_TXT_LEN]
         print(f"Input text was cutoff since it went over the {MAX_TXT_LEN} character limit.")
     print(text)
-
     
-    # synthesize
-    synthesizer = Synthesizer(
-        model_name+"/best_model.pth", model_name+"/config.json"
-    )
-    if synthesizer is None:
-        raise NameError("model not found")
+    model_dir = model_name.replace(" ", "_")
+    model_path = os.path.join(model_dir, "best_model.pth")
+    config_path = os.path.join(model_dir, "config.json")
+    
+    if not os.path.exists(model_path) or not os.path.exists(config_path):
+        raise FileNotFoundError(f"Model or config not found for {model_name}")
+    
+    # Load config
+    config = load_config(config_path)
+    
+    # Create the synthesizer
+    synthesizer = Synthesizer(model_path, config)
+    
+    # Synthesize text
     wavs = synthesizer.tts(text)
-    # return output
+    
+    # Save the output to a temporary file
     with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
-        synthesizer.save_wav(wavs, fp)
+        synthesizer.save_wav(wavs, fp.name)
         return fp.name
 
-
-description="""
-
+description = """
+A Persian Text-to-Speech (TTS) demo using various models.
 """
-article= ""
-examples=[
-    ["و خداوند شما را با ارسال روح در جسم زندگانی و حیات بخشید","vits-male"],
-    ["تاجر تو چه تجارت می کنی ، تو را چه که چه تجارت می کنم؟","vits female (best)"],
-    ["شیش سیخ جیگر سیخی شیش هزار","vits female (best)"],
-    ["سه شیشه شیر ، سه سیر سرشیر","vits female (best)"],
-    ["دزدی دزدید ز بز دزدی بزی ، عجب دزدی که دزدید ز بز دزدی بزی","vits male1 (best)"],
-    ["مثنوی یکی از قالب های شعری است ک هر بیت قافیه ی جداگانه دارد","vits female1"],
-    ["در گلو ماند خس او سالها، چیست آن خس مهر جاه و مالها","vits male1 (best)"],
+
+examples = [
+    ["و خداوند شما را با ارسال روح در جسم زندگانی و حیات بخشید", "vits male1 (best)"],
+    ["تاجر تو چه تجارت می کنی ، تو را چه که چه تجارت می کنم؟", "vits female (best)"],
+    ["شیش سیخ جیگر سیخی شیش هزار", "vits female (best)"],
+    ["سه شیشه شیر ، سه سیر سرشیر", "vits female (best)"],
+    ["دزدی دزدید ز بز دزدی بزی ، عجب دزدی که دزدید ز بز دزدی بزی", "vits male1 (best)"],
+    ["مثنوی یکی از قالب های شعری است ک هر بیت قافیه ی جداگانه دارد", "vits female1"],
+    ["در گلو ماند خس او سالها، چیست آن خس مهر جاه و مالها", "vits male1 (best)"],
 ]
+
 iface = gr.Interface(
     fn=tts,
     inputs=[
-        gr.Textbox(
-            label="Text",
-            value="زندگی فقط یک بار است؛ از آن به خوبی استفاده کن",
-        ),
-        gr.Radio(
-            label="Pick a TTS Model ",
-            choices=MODEL_NAMES,
-            value="vits-female",
-        ),
+        gr.Textbox(label="Text", value="زندگی فقط یک بار است؛ از آن به خوبی استفاده کن"),
+        gr.Radio(label="Pick a TTS Model", choices=MODEL_NAMES, value="vits female (best)")
     ],
-    outputs=gr.Audio(label="Output",type='filepath'),
+    outputs=gr.Audio(label="Output", type='filepath'),
     examples=examples,
-    title="🗣️ Persian tts 🗣️",
+    title="🗣️ Persian TTS 🗣️",
     description=description,
-    article=article,
+    article="",
     live=False
 )
-iface.launch(share=False)
+
+iface.launch(share=False)