Update app.py

app.py

@@ -1,38 +1,44 @@
 import os
 import sys
 import gradio as gr
-import random
 import numpy as np
 import torch
 import librosa
 import torchaudio
 from scipy.signal import resample
 import time
-import requests
+from huggingface_hub import snapshot_download, hf_hub_download
 
 ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
 sys.path.append(f'{ROOT_DIR}/third_party/Matcha-TTS')
 
 from cosyvoice.cli.cosyvoice import CosyVoice
 from cosyvoice.utils.file_utils import load_wav
-from huggingface_hub import snapshot_download
+
+def download_audio_from_dataset(filename, save_path="temp_prompt.wav"):
+    return hf_hub_download(
+        repo_id="kanahomaisa/breezyvoice-samples",
+        filename=filename,
+        repo_type="dataset"
+    )
 
 preset_speakers = {
     "6歲": {
-        "audio_path": "data/smile_train.wav",
+        "filename": "smile_train.wav",
         "transcription": "名字是微笑號，只是呢你們看，這一輛微笑號它這裡有寫八百型的喔，它是八百山出來"
     },
     "8歲": {
-        "audio_path": "data/utensils.wav",
+        "filename": "utensils.wav",
         "transcription": "出來了出來了，你知道這個餐具是可以挖的，這個餐具可以用窩課魅一起挖嗎?"
     }
 }
 
 def apply_preset(speaker_key):
     if speaker_key in preset_speakers:
-        path = preset_speakers[speaker_key]["audio_path"]
-        text = preset_speakers[speaker_key]["transcription"]
-        return path, text
+        filename = preset_speakers[speaker_key]["filename"]
+        transcription = preset_speakers[speaker_key]["transcription"]
+        local_path = download_audio_from_dataset(filename)
+        return local_path, transcription
     return None, ""
 
 max_val = 0.8
@@ -97,12 +103,6 @@ def synthesize_speech(speaker_audio_path, content_text, speaker_text, speaker_ke
     torchaudio.save(out_path, torch.tensor(audio_data).unsqueeze(0), sample_rate=target_sr)
     return out_path
 
-def download_audio_from_url(url, save_path="temp_prompt.wav"):
-    response = requests.get(url)
-    with open(save_path, "wb") as f:
-        f.write(response.content)
-    return save_path
-
 with gr.Blocks() as demo:
     gr.Markdown("""
     # 小睿語音合成
@@ -147,4 +147,5 @@ with gr.Blocks() as demo:
     為了加速，已關閉自動語音辨識，務必正確輸入語音樣本的文字轉寫。
     """)
 
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()