update app to include audio playback

app.py

@@ -3,7 +3,6 @@ Main entry point for the Audio Translation Web Application
 Handles file upload, processing pipeline, and UI rendering
 """
 
-# Configure logging first
 import logging
 logging.basicConfig(
     level=logging.INFO,
@@ -21,42 +20,7 @@ import time
 import subprocess
 from utils.stt import transcribe_audio
 from utils.translation import translate_text
-from utils.tts import generate_speech
-
-# Hugging Face Spaces Setup Automation
-def setup_huggingface_space():
-    """Automatically configure Hugging Face Space requirements"""
-    logger.info("Running Hugging Face space setup")
-    st.sidebar.header("Space Configuration")
-    
-    try:
-        subprocess.run(["espeak-ng", "--version"], check=True, capture_output=True)
-        logger.info("espeak-ng verification successful")
-    except (FileNotFoundError, subprocess.CalledProcessError):
-        logger.error("Missing espeak-ng dependency")
-        st.sidebar.error("""
-        **Missing System Dependencies!** Add this to your Space settings:
-        ```txt
-        apt-get update && apt-get install -y espeak-ng
-        ```
-        """)
-        st.stop()
-
-    model_dir = "./kokoro"
-    required_files = [
-        f"{model_dir}/kokoro-v0_19.pth",
-        f"{model_dir}/voices/af_bella.pt"
-    ]
-    
-    if not all(os.path.exists(f) for f in required_files):
-        logger.error("Missing model files in %s", model_dir)
-        st.sidebar.warning("""
-        **Missing Model Files!** Add this to your Space settings:
-        ```txt
-        git clone https://huggingface.co/hexgrad/Kokoro-82M ./kokoro
-        ```
-        """)
-        st.stop()
+from utils.tts import get_tts_engine, generate_speech
 
 # Initialize environment configurations
 os.makedirs("temp/uploads", exist_ok=True)
@@ -111,15 +75,18 @@ def handle_file_processing(upload_path):
         # TTS Phase
         logger.info("Beginning TTS generation")
         status_text.markdown("🎵 **Generating Chinese Speech...**")
-        with st.spinner("Initializing TTS engine..."):
-            output_path = generate_speech(chinese_text, voice="zf_xiaobei")
+        
+        # Initialize TTS engine with appropriate language code for Chinese
+        engine = get_tts_engine(lang_code='z')  # 'z' for Mandarin Chinese
+        
+        # Generate speech and get the file path
+        output_path = engine.generate_speech(chinese_text, voice="zf_xiaobei")
         progress_bar.progress(100)
         logger.info(f"TTS completed. Output file: {output_path}")
         
+        # Store the text for streaming playback
+        st.session_state.current_text = chinese_text
         
-        # Display results
-
-        # Display results
         status_text.success("✅ Processing Complete!")
         return english_text, chinese_text, output_path
         
@@ -144,7 +111,10 @@ def render_results(english_text, chinese_text, output_path):
 
     with col2:
         st.subheader("Audio Output")
+        # Standard audio player for the full file
         st.audio(output_path)
+        
+        # Download button
         with open(output_path, "rb") as f:
             st.download_button(
                 label="Download Audio",
@@ -152,15 +122,57 @@ def render_results(english_text, chinese_text, output_path):
                 file_name="translated_audio.wav",
                 mime="audio/wav"
             )
+        
+        # Streaming playback controls
+        st.subheader("Streaming Playback")
+        if st.button("Stream Audio"):
+            engine = get_tts_engine(lang_code='z')
+            streaming_placeholder = st.empty()
+            
+            # Stream the audio in chunks
+            for sample_rate, audio_chunk in engine.generate_speech_stream(
+                chinese_text, 
+                voice="zf_xiaobei"
+            ):
+                # Create a temporary file for each chunk
+                temp_chunk_path = f"temp/outputs/chunk_{time.time()}.wav"
+                import soundfile as sf
+                sf.write(temp_chunk_path, audio_chunk, sample_rate)
+                
+                # Play the chunk
+                with streaming_placeholder:
+                    st.audio(temp_chunk_path, sample_rate=sample_rate)
+                
+                # Clean up the temporary chunk file
+                os.remove(temp_chunk_path)
+
+def initialize_session_state():
+    """Initialize session state variables"""
+    if 'current_text' not in st.session_state:
+        st.session_state.current_text = None
 
 def main():
     """Main application workflow"""
     logger.info("Starting application")
-    # setup_huggingface_space()  # First-run configuration checks
     configure_page()
+    initialize_session_state()
+    
     st.title("🎧 High-Quality Audio Translation System")
     st.markdown("Upload English Audio → Get Chinese Speech Output")
 
+    # Voice selection in sidebar
+    st.sidebar.header("TTS Settings")
+    voice_options = {
+        "Xiaobei (Female)": "zf_xiaobei",
+        "Yunjian (Male)": "zm_yunjian",
+    }
+    selected_voice = st.sidebar.selectbox(
+        "Select Voice",
+        list(voice_options.keys()),
+        format_func=lambda x: x
+    )
+    speed = st.sidebar.slider("Speech Speed", 0.5, 2.0, 1.0, 0.1)
+
     uploaded_file = st.file_uploader(
         "Select Audio File (MP3/WAV)",
         type=["mp3", "wav"],

requirements.txt

@@ -13,4 +13,4 @@ scipy>=1.11
 munch>=2.5
 accelerate>=1.2.0
 soundfile>=0.13.0
-# git+https://github.com/hexgrad/Kokoro-82M
+kokoro>=0.7.9