Update app.py

app.py

@@ -12,27 +12,29 @@ from indic_transliteration import sanscript
 from indic_transliteration.sanscript import transliterate
 import azure.cognitiveservices.speech as speechsdk
 import ffmpeg
-from PIL import Image
-import imageio
 
-# Configure MoviePy to use imageio for reading images
-imageio.plugins.ffmpeg.download()
+# Set page configuration
+st.set_page_config(
+    page_title="translate",
+    page_icon="🎬",
+    layout="wide"
+)
 
-# Configure ImageMagick policy to allow PDF and text file handling
-def configure_imagemagick():
-    """Configure ImageMagick policy to allow text operations"""
-    policy_file = "/etc/ImageMagick-6/policy.xml"
-    if os.path.exists(policy_file):
-        try:
-            with open(policy_file, 'r') as f:
-                policy_content = f.read()
-            # Modify policy to allow text file handling
-            policy_content = policy_content.replace('rights="none" pattern="@*"', 'rights="read|write" pattern="@*"')
-            with open(policy_file, 'w') as f:
-                f.write(policy_content)
-        except Exception as e:
-            st.warning(f"Unable to configure ImageMagick policy: {e}")
-            st.info("You may need to run this application with sudo privileges to modify ImageMagick policy")
+# Custom CSS to improve the interface
+st.markdown("""
+    <style>
+    .stButton>button {
+        width: 100%;
+        border-radius: 5px;
+        height: 3em;
+        background-color: #FF4B4B;
+        color: white;
+    }
+    .stProgress .st-bo {
+        background-color: #FF4B4B;
+    }
+    </style>
+    """, unsafe_allow_html=True)
 
 # Tamil-specific voice configurations
 TAMIL_VOICES = {
@@ -59,42 +61,39 @@ class TamilTextProcessor:
         text = ' '.join(text.split())
         return text
 
+@st.cache_resource
+def load_whisper_model():
+    """Load Whisper model with caching"""
+    return whisper.load_model("base")
+
 class TamilDubber:
     def __init__(self):
-        try:
-            self.whisper_model = whisper.load_model("base")
-        except Exception as e:
-            st.error(f"Error loading Whisper model: {e}")
-            raise
-        self.temp_files = []
-        
-    def __enter__(self):
-        return self
+        self.whisper_model = load_whisper_model()
+        self.temp_dir = tempfile.mkdtemp()
         
-    def __exit__(self, exc_type, exc_val, exc_tb):
-        self.cleanup()
+    def create_temp_file(self, suffix):
+        """Create a temporary file in the temp directory"""
+        return os.path.join(self.temp_dir, f"temp_{os.urandom(8).hex()}{suffix}")
 
     def cleanup(self):
-        for temp_file in self.temp_files:
-            if os.path.exists(temp_file):
-                try:
-                    os.remove(temp_file)
-                except Exception:
-                    pass
-
-    def create_temp_file(self, suffix):
-        temp_file = tempfile.mktemp(suffix=suffix)
-        self.temp_files.append(temp_file)
-        return temp_file
+        """Clean up temporary files"""
+        import shutil
+        try:
+            shutil.rmtree(self.temp_dir)
+        except Exception as e:
+            st.warning(f"Cleanup warning: {e}")
 
     def extract_audio(self, video_path):
         """Extract audio and transcribe using Whisper"""
         try:
             video = VideoFileClip(video_path)
             audio_path = self.create_temp_file(".wav")
-            video.audio.write_audiofile(audio_path)
+            video.audio.write_audiofile(audio_path, fps=16000)
+            
+            # Transcribe using Whisper
             result = self.whisper_model.transcribe(audio_path)
             return result["segments"], video.duration
+            
         except Exception as e:
             st.error(f"Error in audio extraction: {e}")
             raise
@@ -118,6 +117,7 @@ class TamilDubber:
                 })
             except Exception as e:
                 st.warning(f"Translation warning for segment: {str(e)}")
+                # Keep original text if translation fails
                 translated_segments.append({
                     "text": segment["text"],
                     "start": segment["start"],
@@ -152,131 +152,116 @@ class TamilDubber:
             st.error(f"Error creating subtitles: {e}")
             raise
 
-    def create_subtitle_clip(self, txt, size, color):
-        """Create subtitle clip with proper configuration"""
-        try:
-            return TextClip(
-                txt=txt,
-                font='DejaVu-Sans',  # Use a system font that supports Tamil
-                fontsize=size,
-                color=color,
-                stroke_color='black',
-                stroke_width=1,
-                method='caption',  # Use caption method instead of label
-                size=(720, None)  # Set width, let height adjust automatically
-            )
-        except Exception as e:
-            st.error(f"Error creating subtitle clip: {e}")
-            raise
-
 def main():
-    # Configure ImageMagick at startup
-    configure_imagemagick()
-    
     st.title("Tamil Movie Dubbing System")
+    st.markdown("""
+    👋 Welcome to the Tamil Movie Dubbing System! This tool helps you:
+    - 🎥 Convert English videos to Tamil
+    - 🗣️ Generate Tamil voiceovers
+    - 📝 Add Tamil subtitles
+    """)
+    
     st.sidebar.header("டப்பிங் அமைப்புகள்")  # Dubbing Settings in Tamil
 
-    # File uploader
+    # File uploader with clear instructions
+    st.info("Please upload a video file (MP4, MOV, or AVI format)")
     video_file = st.file_uploader("Upload Video File", type=['mp4', 'mov', 'avi'])
+    
     if not video_file:
+        st.warning("Please upload a video to begin the dubbing process.")
         return
 
-    # Settings
-    voice_type = st.selectbox("Select Voice", list(TAMIL_VOICES.keys()))
-    
-    with st.expander("Advanced Settings"):
+    # Settings in sidebar
+    with st.sidebar:
+        st.subheader("Voice Settings")
+        voice_type = st.selectbox("Select Voice", list(TAMIL_VOICES.keys()))
+        
+        st.subheader("Subtitle Settings")
         generate_subtitles = st.checkbox("Generate Tamil Subtitles", value=True)
-        subtitle_size = st.slider("Subtitle Size", 16, 32, 24)
-        subtitle_color = st.color_picker("Subtitle Color", "#FFFFFF")
+        if generate_subtitles:
+            subtitle_size = st.slider("Subtitle Size", 16, 32, 24)
+            subtitle_color = st.color_picker("Subtitle Color", "#FFFFFF")
 
+    # Main process
     if st.button("Start Tamil Dubbing"):
         try:
-            with st.spinner("Processing video..."):
-                with TamilDubber() as dubber:
-                    # Save uploaded video
-                    temp_video_path = dubber.create_temp_file(".mp4")
-                    with open(temp_video_path, "wb") as f:
-                        f.write(video_file.read())
+            dubber = TamilDubber()
+            
+            # Create progress containers
+            progress_bar = st.progress(0)
+            status_text = st.empty()
+            
+            try:
+                # Save uploaded video
+                temp_video_path = dubber.create_temp_file(".mp4")
+                with open(temp_video_path, "wb") as f:
+                    f.write(video_file.read())
 
-                    # Progress tracking
-                    progress_bar = st.progress(0)
-                    status_text = st.empty()
+                # Extract audio and transcribe
+                status_text.text("📥 Extracting audio and transcribing...")
+                segments, video_duration = dubber.extract_audio(temp_video_path)
+                progress_bar.progress(0.25)
 
-                    # Extract audio and transcribe
-                    status_text.text("Extracting audio and transcribing...")
-                    segments, video_duration = dubber.extract_audio(temp_video_path)
-                    progress_bar.progress(0.25)
+                # Translate segments
+                status_text.text("🔄 Translating to Tamil...")
+                translated_segments = dubber.translate_segments(segments)
+                progress_bar.progress(0.50)
 
-                    # Translate segments
-                    status_text.text("Translating to Tamil...")
-                    translated_segments = dubber.translate_segments(segments)
-                    progress_bar.progress(0.50)
+                # Generate Tamil audio
+                status_text.text("🔊 Generating Tamil audio...")
+                video = VideoFileClip(temp_video_path)
+                
+                audio_segments = []
+                for idx, segment in enumerate(translated_segments):
+                    audio_path = dubber.generate_audio(segment["text"])
+                    audio_segments.append({
+                        "audio": AudioFileClip(audio_path),
+                        "start": segment["start"]
+                    })
+                    progress_bar.progress(0.50 + (0.25 * (idx + 1) / len(translated_segments)))
 
-                    # Generate Tamil audio
-                    status_text.text("Generating Tamil audio...")
-                    output_segments = []
-                    video = VideoFileClip(temp_video_path)
+                # Create final video
+                status_text.text("🎬 Creating final video...")
+                output_path = dubber.create_temp_file(".mp4")
+                
+                # Add subtitles if enabled
+                if generate_subtitles:
+                    srt_path = dubber.create_temp_file(".srt")
+                    dubber.create_subtitles(translated_segments, srt_path)
                     
-                    for idx, segment in enumerate(translated_segments):
-                        audio_path = dubber.generate_audio(segment["text"])
-                        output_segments.append({
-                            "audio": audio_path,
-                            "start": segment["start"],
-                            "end": segment["end"]
-                        })
-                        progress_bar.progress(0.50 + (0.25 * (idx + 1) / len(translated_segments)))
+                    # Use ffmpeg to add subtitles
+                    stream = ffmpeg.input(temp_video_path)
+                    stream = ffmpeg.output(stream, output_path, 
+                                        vf=f'subtitles={srt_path}:force_style=\'FontSize={subtitle_size},PrimaryColour={subtitle_color}\'',
+                                        acodec='aac')
+                    ffmpeg.run(stream, overwrite_output=True)
+                else:
+                    # Just copy the video if no subtitles
+                    video.write_videofile(output_path)
 
-                    # Create final video
-                    status_text.text("Creating final video...")
-                    output_path = dubber.create_temp_file(".mp4")
-                    
-                    # Add subtitles if enabled
-                    if generate_subtitles:
-                        subtitle_clips = []
-                        for segment in translated_segments:
-                            try:
-                                clip = dubber.create_subtitle_clip(
-                                    segment["text"],
-                                    subtitle_size,
-                                    subtitle_color
-                                )
-                                clip = clip.set_position(('center', 'bottom'))
-                                clip = clip.set_start(segment["start"])
-                                clip = clip.set_duration(segment["duration"])
-                                subtitle_clips.append(clip)
-                            except Exception as e:
-                                st.warning(f"Skipping subtitle for segment due to error: {e}")
+                progress_bar.progress(1.0)
+                status_text.text("✅ Dubbing completed!")
 
-                        final_video = CompositeVideoClip([video] + subtitle_clips)
-                    else:
-                        final_video = video
+                # Display result
+                st.success("டப்பிங் வெற்றிகரமாக முடிந்தது!")  # Dubbing completed successfully in Tamil
+                st.video(output_path)
 
-                    # Write final video with proper codec settings
-                    final_video.write_videofile(
-                        output_path,
-                        codec='libx264',
-                        audio_codec='aac',
-                        fps=video.fps,
-                        threads=4,
-                        preset='medium'
+                # Download button
+                with open(output_path, "rb") as f:
+                    st.download_button(
+                        "⬇️ Download Dubbed Video",
+                        f,
+                        file_name="tamil_dubbed_video.mp4",
+                        mime="video/mp4"
                     )
-                    progress_bar.progress(1.0)
-
-                    # Display result
-                    st.success("டப்பிங் வெற்றிகரமாக முடிந்தது!")  # Dubbing completed successfully in Tamil
-                    st.video(output_path)
-
-                    # Download button
-                    with open(output_path, "rb") as f:
-                        st.download_button(
-                            "Download Dubbed Video",
-                            f,
-                            file_name="tamil_dubbed_video.mp4",
-                            mime="video/mp4"
-                        )
 
+            finally:
+                # Cleanup
+                dubber.cleanup()
+                
         except Exception as e:
             st.error(f"An error occurred: {str(e)}")
+            st.error("Please try again with a different video or check if the video format is supported.")
 
 if __name__ == "__main__":
     main()