Update app.py

app.py

@@ -8,17 +8,18 @@ from huggingface_hub import HfApi
 
 dataset = load_dataset("intersteller2887/Turing-test-dataset", split="train")
 
-print(dataset[0])  # 查看第一条数据结构
-print(type(dataset[0]['audio']))  # audio字段的类型
-print(dataset[0]['audio'])  # audio字段的具体值
-
-all_data_audio_paths = [
-    item['audio']['path'] for item in dataset
-    if item['audio']['path'].endswith(".wav") and "/data/" in item['audio']['path'].replace("\\", "/")
+all_audio_paths = [
+    item["audio"]["path"] 
+    for item in dataset 
+    if item["audio"] and "path" in item["audio"]
 ]
 
-for p in all_data_audio_paths:
-    print(p)
+print("Found audio files:")
+for path in all_audio_paths[:5]:
+    print(path)
+
+valid_audio_paths = [path for path in all_audio_paths if os.path.exists(path)]
+print(f"Total valid audio files: {len(valid_audio_paths)}")
 
 sample1_audio_path = next((p for p in all_data_audio_paths if p.endswith("sample1.wav")), None)