Update app.py

app.py

@@ -1,30 +1,97 @@
-# app.py
+# # app.py
+# import os
+# import gradio as gr
+# from huggingface_hub import HfApi
+
+# api = HfApi()
+
+# def list_dataset_files(repo_id: str):
+#     # repo_id example: "username/dataset_name"
+#     try:
+#         files = api.list_repo_files(
+#             repo_id=repo_id,
+#             repo_type="dataset",
+#             token=os.getenv("HF_TOKEN")  # None for public datasets
+#         )
+#         if not files:
+#             return "No files found."
+#         return "\n".join(files)
+#     except Exception as e:
+#         return f"Error: {e}"
+
+# with gr.Blocks() as demo:
+#     gr.Markdown("# List files in a Hugging Face dataset repo")
+#     repo_input = gr.Textbox(label="Dataset repo_id", value="Juna190825/ZomiAudioDataset")
+#     output = gr.Textbox(label="Files", lines=20)
+#     btn = gr.Button("List files")
+
+#     btn.click(list_dataset_files, inputs=repo_input, outputs=output)
+
+# demo.launch()
+
+
 import os
 import gradio as gr
-from huggingface_hub import HfApi
+from huggingface_hub import HfApi, hf_hub_download
 
 api = HfApi()
 
+AUDIO_EXTS = {".wav", ".mp3", ".flac", ".ogg", ".m4a"}
+
 def list_dataset_files(repo_id: str):
-    # repo_id example: "username/dataset_name"
     try:
         files = api.list_repo_files(
             repo_id=repo_id,
             repo_type="dataset",
-            token=os.getenv("HF_TOKEN")  # None for public datasets
+            token=os.getenv("HF_TOKEN")
         )
         if not files:
-            return "No files found."
-        return "\n".join(files)
+            return [], "No files found."
+
+        # Filter audio files
+        audio_files = [f for f in files if os.path.splitext(f)[1].lower() in AUDIO_EXTS]
+
+        return audio_files, "\n".join(files)
+
     except Exception as e:
-        return f"Error: {e}"
+        return [], f"Error: {e}"
+
+def load_audio(repo_id, file_path):
+    try:
+        local_path = hf_hub_download(
+            repo_id=repo_id,
+            filename=file_path,
+            repo_type="dataset",
+            token=os.getenv("HF_TOKEN")
+        )
+        return local_path
+    except Exception as e:
+        return None
 
 with gr.Blocks() as demo:
-    gr.Markdown("# List files in a Hugging Face dataset repo")
-    repo_input = gr.Textbox(label="Dataset repo_id", value="Juna190825/ZomiAudioDataset")
-    output = gr.Textbox(label="Files", lines=20)
+    gr.Markdown("# 🎧 List & Play Audio Files from a HF Dataset Repo")
+
+    repo_input = gr.Textbox(label="Dataset repo_id", value="username/dataset_name")
+
+    with gr.Row():
+        file_list = gr.Dropdown(label="Audio files", choices=[])
+        play_audio = gr.Audio(label="Audio Player")
+
+    output = gr.Textbox(label="All Files", lines=20)
     btn = gr.Button("List files")
 
-    btn.click(list_dataset_files, inputs=repo_input, outputs=output)
+    # When clicking "List files"
+    def update_files(repo_id):
+        audio_files, all_files_text = list_dataset_files(repo_id)
+        return gr.Dropdown(choices=audio_files, value=None), all_files_text
+
+    btn.click(update_files, inputs=repo_input, outputs=[file_list, output])
+
+    # When selecting an audio file
+    file_list.change(
+        load_audio,
+        inputs=[repo_input, file_list],
+        outputs=play_audio
+    )
 
 demo.launch()