Update app.py

app.py

@@ -2,23 +2,17 @@ import gradio as gr
 import edge_tts
 import asyncio
 import tempfile
-import os
 import nltk
+import os
 import srt
 from pydub import AudioSegment, silence
 import datetime
 import nest_asyncio
-import threading
 
-# Enable nested event loop (required for Spaces)
+nltk.download("punkt")
 nest_asyncio.apply()
 
-# 📦 Safe punkt download
-nltk_data_path = os.path.join(os.path.expanduser("~"), "nltk_data")
-nltk.download("punkt", download_dir=nltk_data_path)
-nltk.data.path.append(nltk_data_path)
 
-# 🗣️ Text-to-Speech and SRT generator
 async def text_to_speech(text, voice, rate, pitch):
     if not text.strip():
         return None, None, "Please enter some text."
@@ -34,11 +28,10 @@ async def text_to_speech(text, voice, rate, pitch):
         tmp_path = tmp_file.name
         await communicate.save(tmp_path)
 
-    # SRT generation
     srt_path = generate_srt(tmp_path, text)
     return tmp_path, srt_path, ""
 
-# 🧠 Generate subtitles
+
 def generate_srt(audio_path, text):
     audio = AudioSegment.from_file(audio_path)
     silences = silence.detect_silence(audio, min_silence_len=400, silence_thresh=audio.dBFS - 16)
@@ -68,48 +61,33 @@ def generate_srt(audio_path, text):
         srt_file.write(srt_data)
         return srt_file.name
 
-# Interface wrapper
+
 async def tts_interface(text, voice, rate, pitch):
-    audio_path, srt_path, message = await text_to_speech(text, voice, rate, pitch)
-    return audio_path, srt_path, message
-
-# UI setup
-async def create_demo():
-    voices = await edge_tts.list_voices()
-    voice_dict = {f"{v['ShortName']} - {v['Locale']} ({v['Gender']})": v['ShortName'] for v in voices}
-
-    with gr.Blocks() as demo:
-        gr.Markdown("# 🎙️ Text-to-Speech + Subtitle Generator")
-
-        with gr.Row():
-            with gr.Column():
-                text_input = gr.Textbox(label="Input Text", lines=5)
-                voice_dropdown = gr.Dropdown(choices=[""] + list(voice_dict.keys()), label="Select Voice")
-                rate_slider = gr.Slider(minimum=-50, maximum=50, value=0, label="Speech Rate (%)")
-                pitch_slider = gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch (Hz)")
-                generate_btn = gr.Button("🎧 Generate Audio + SRT")
-
-            with gr.Column():
-                audio_output = gr.Audio(label="Generated Audio", type="filepath")
-                srt_output = gr.File(label="Download Subtitle (.srt)")
-                message_output = gr.Textbox(label="Status", interactive=False)
-
-        generate_btn.click(
-            fn=tts_interface,
-            inputs=[text_input, voice_dropdown, rate_slider, pitch_slider],
-            outputs=[audio_output, srt_output, message_output]
-        )
-
-    return demo
-
-# Global demo instance (for Hugging Face Spaces compatibility)
-demo = None
-
-def launch_app():
-    global demo
-    demo = asyncio.run(create_demo())
-    demo.queue()
-    demo.launch()
-
-# Start the app in a thread (non-blocking, Space can detect `demo`)
-threading.Thread(target=launch_app).start()
+    return await text_to_speech(text, voice, rate, pitch)
+
+
+# ⬇️ Create demo synchronously (run async functions in loop)
+voices = asyncio.run(edge_tts.list_voices())
+voice_dict = {f"{v['ShortName']} - {v['Locale']} ({v['Gender']})": v['ShortName'] for v in voices}
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 🎙️ Text-to-Speech + Subtitle Generator")
+
+    with gr.Row():
+        with gr.Column():
+            text_input = gr.Textbox(label="Input Text", lines=5)
+            voice_dropdown = gr.Dropdown(choices=[""] + list(voice_dict.keys()), label="Select Voice")
+            rate_slider = gr.Slider(minimum=-50, maximum=50, value=0, label="Speech Rate (%)")
+            pitch_slider = gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch (Hz)")
+            generate_btn = gr.Button("🎧 Generate Audio + SRT")
+
+        with gr.Column():
+            audio_output = gr.Audio(label="Generated Audio", type="filepath")
+            srt_output = gr.File(label="Download Subtitle (.srt)")
+            message_output = gr.Textbox(label="Status", interactive=False)
+
+    generate_btn.click(
+        fn=lambda text, voice, rate, pitch: asyncio.run(tts_interface(text, voice, rate, pitch)),
+        inputs=[text_input, voice_dropdown, rate_slider, pitch_slider],
+        outputs=[audio_output, srt_output, message_output]
+    )