app.py

import gradio as gr
import tempfile
import os
import re
from moviepy.editor import VideoFileClip, concatenate_videoclips, AudioFileClip, CompositeAudioClip, concatenate_audioclips
import numpy as np
import logging
import sys
import traceback
import socket
import shutil

# Set up logging to debug issues
logging.basicConfig(level=logging.INFO, handlers=[logging.StreamHandler(sys.stdout)])
logger = logging.getLogger(__name__)

# --- Functions ---

def check_port(port):
    """
    Check if a port is available.
    Returns True if the port is free, False if it's in use.
    """
    with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
        try:
            s.bind(("0.0.0.0", port))
            return True
        except socket.error:
            return False

def sort_files_by_index(file_list, prefix_pattern):
    """
    Sort files based on their numerical index in the filename.
    Args:
        file_list: List of file paths (e.g., ['file2.mp4', 'file1.mp4', 'file3.mp4'])
        prefix_pattern: Regex pattern to match the prefix and number (e.g., r'file(\d+)\.mp4')
    Returns:
        Sorted list of file paths
    """
    if not file_list:
        return []
    
    def extract_index(filename):
        match = re.match(prefix_pattern, os.path.basename(filename))
        if match:
            return int(match.group(1))
        return float('inf')  # Invalid filenames go to the end
    
    sorted_files = sorted(file_list, key=extract_index)
    logger.info(f"Sorted files: {sorted_files}")
    return sorted_files

def get_indices_string(file_list, prefix_pattern):
    """
    Extract the numerical indices from filenames and return as a string.
    Args:
        file_list: List of file paths
        prefix_pattern: Regex pattern to match the prefix and number
    Returns:
        String of indices (e.g., '123' for ['file1.mp4', 'file2.mp4', 'file3.mp4'])
    """
    if not file_list:
        return ""
    indices = []
    for filename in file_list:
        match = re.match(prefix_pattern, os.path.basename(filename))
        if match:
            indices.append(match.group(1))
    return "".join(indices)

def split_files_by_extension(all_files):
    """
    Split a list of files into videos (.mp4) and audios (.mp3, .wav) based on extensions.
    Args:
        all_files: List of file paths (e.g., ['file1.mp4', 'audio1.mp3', 'file2.mp4', 'audio2.mp3'])
    Returns:
        Tuple of (video_files, audio_files)
    """
    video_files = []
    audio_files = []
    for file_path in all_files:
        filename = os.path.basename(file_path).lower()
        if filename.endswith('.mp4'):
            video_files.append(file_path)
        elif filename.endswith(('.mp3', '.wav')):
            audio_files.append(file_path)
        else:
            logger.warning(f"Unsupported file extension for {file_path}; skipping")
    return video_files, audio_files

def trim_silence(audio_clip, threshold=0.005):
    """
    Trim silence from the start and end of an audio clip.
    Args:
        audio_clip: AudioFileClip object
        threshold: Amplitude threshold below which audio is considered silent
    Returns:
        Trimmed AudioFileClip
    """
    try:
        # Get audio data as numpy array
        samples = audio_clip.to_soundarray(fps=44100)
        # Compute amplitude (RMS)
        if len(samples.shape) > 1:  # Stereo audio
            amplitudes = np.sqrt(np.mean(samples**2, axis=1))
        else:  # Mono audio
            amplitudes = np.sqrt(samples**2)
        
        # Find non-silent regions
        non_silent = amplitudes > threshold
        if not np.any(non_silent):
            logger.warning("Audio clip is completely silent; returning original clip")
            return audio_clip
        
        # Find start and end indices
        start_idx = np.argmax(non_silent)
        end_idx = len(non_silent) - np.argmax(non_silent[::-1])
        
        # Convert indices to time (seconds)
        start_time = start_idx / 44100
        end_time = end_idx / 44100
        
        # Ensure the trimmed duration is reasonable
        if end_time <= start_time:
            logger.warning("Trimmed duration is zero or negative; returning original clip")
            return audio_clip
        
        # Trim the audio
        trimmed_audio = audio_clip.subclip(start_time, end_time)
        logger.info(f"Trimmed audio from {start_time:.2f}s to {end_time:.2f}s (original duration: {audio_clip.duration:.2f}s)")
        return trimmed_audio
    except Exception as e:
        logger.error(f"Error trimming silence: {str(e)}")
        return audio_clip

def merge_videos_and_audios(video_files=None, audio_files=None, orig_vol=1.0, music_vol=0.5, temp_dir=None):
    """
    Merge multiple video clips and/or audio clips based on inputs provided.
    - If only video_files: Merge videos, retaining their original audio.
    - If only audio_files: Merge audio files into a single audio file.
    - If both: Merge videos and overlay the concatenated audio.
    Files are sorted by numerical index in their filenames (e.g., file1.mp4, file2.mp4).
    Args:
        video_files: List of video file paths (optional)
        audio_files: List of audio file paths (optional)
        orig_vol: Volume for original video audio (0.0 to 1.0)
        music_vol: Volume for background audio (0.0 to 1.0)
        temp_dir: Temporary directory to clean up (optional)
    Returns:
        Path to the merged file (video or audio) or error message.
    """
    try:
        # Sort files by numerical index
        video_files = sort_files_by_index(video_files, r'file(\d+)\.mp4')
        audio_files = sort_files_by_index(audio_files, r'audio(\d+)\.(mp3|wav)')
        
        # Get indices for output naming
        video_indices = get_indices_string(video_files, r'file(\d+)\.mp4')
        audio_indices = get_indices_string(audio_files, r'audio(\d+)\.(mp3|wav)')

        # Ensure at least two files are provided (videos, audios, or combination)
        video_count = len(video_files) if video_files else 0
        audio_count = len(audio_files) if audio_files else 0
        total_files = video_count + audio_count

        logger.info(f"Starting merge with {video_count} video files and {audio_count} audio files")

        if total_files < 2:
            error_msg = "Error: Please upload at least 2 files total (videos, audios, or a combination)."
            logger.error(error_msg)
            return error_msg

        # Create a temporary output path (use provided temp_dir if available)
        if temp_dir is None:
            temp_dir = tempfile.mkdtemp()
        output_dir = temp_dir

        # Case 1: Audio only
        if audio_count >= 2 and video_count == 0:
            output_filename = f"combined_audio_{audio_indices}.mp3"
            output_path = os.path.join(output_dir, output_filename)
            logger.info("Merging audio files only")
            
            # Load, normalize, and trim audio clips
            audio_clips = []
            for audio in audio_files:
                clip = AudioFileClip(audio).set_fps(44100)  # Normalize sample rate
                clip = clip.to_stereo() if clip.nchannels == 1 else clip  # Convert mono to stereo
                logger.info(f"Original audio duration for {audio}: {clip.duration:.2f}s, channels: {clip.nchannels}")
                trimmed_clip = trim_silence(clip)
                if trimmed_clip.duration > 0:
                    audio_clips.append(trimmed_clip)
                else:
                    logger.warning(f"Skipping audio file {audio} as it has zero duration after trimming")
            
            # Check if we have enough clips to concatenate
            if len(audio_clips) < 2:
                error_msg = "Error: Fewer than 2 audio clips available after trimming (clips may be silent or too short)."
                logger.error(error_msg)
                for clip in audio_clips:
                    clip.close()
                return error_msg
            
            # Log durations after trimming
            for i, clip in enumerate(audio_clips):
                logger.info(f"Trimmed audio {i+1} duration: {clip.duration:.2f}s, channels: {clip.nchannels}")
            
            # Concatenate audio clips
            logger.info(f"Attempting to concatenate {len(audio_clips)} audio clips")
            final_audio_clip = concatenate_audioclips(audio_clips)
            logger.info(f"Concatenated audio duration: {final_audio_clip.duration:.2f}s, channels: {final_audio_clip.nchannels}")
            
            # Verify concatenated duration
            expected_duration = sum(clip.duration for clip in audio_clips)
            if abs(final_audio_clip.duration - expected_duration) > 0.1:
                logger.warning(f"Concatenated duration ({final_audio_clip.duration:.2f}s) does not match expected duration ({expected_duration:.2f}s)")
            
            # Write the final audio
            logger.info(f"Writing output audio to {output_path}")
            final_audio_clip.write_audiofile(output_path, codec="mp3")
            
            # Close resources
            final_audio_clip.close()
            for clip in audio_clips:
                clip.close()
            
            logger.info("Audio merge completed successfully")
            return output_path

        # Case 2: Video only or Video with Audio
        if audio_indices:
            output_filename = f"combined_video_{video_indices}_with_audio_{audio_indices}.mp4"
        else:
            output_filename = f"combined_video_{video_indices}.mp4"
        output_path = os.path.join(output_dir, output_filename)
        
        # Load and concatenate video clips
        video_clips = [VideoFileClip(video) for video in video_files]
        final_video_clip = concatenate_videoclips(video_clips, method='compose')
        
        # Determine final video duration
        video_duration = final_video_clip.duration or sum(clip.duration for clip in video_clips)
        logger.info(f"Total video duration: {video_duration:.2f}s")
        
        # Handle audio (if provided)
        if audio_files:
            logger.info("Processing audio files")
            # Load, normalize, and trim audio clips
            audio_clips = []
            for audio in audio_files:
                clip = AudioFileClip(audio).set_fps(44100)  # Normalize sample rate
                clip = clip.to_stereo() if clip.nchannels == 1 else clip  # Convert mono to stereo
                logger.info(f"Original audio duration for {audio}: {clip.duration:.2f}s, channels: {clip.nchannels}")
                trimmed_clip = trim_silence(clip)
                if trimmed_clip.duration > 0:
                    audio_clips.append(trimmed_clip)
                else:
                    logger.warning(f"Skipping audio file {audio} as it has zero duration after trimming")
            
            # Log durations after trimming
            for i, clip in enumerate(audio_clips):
                logger.info(f"Trimmed audio {i+1} duration: {clip.duration:.2f}s, channels: {clip.nchannels}")
            
            if not audio_clips:
                logger.warning("No valid audio clips after trimming; using original video audio only")
                final_audio = final_video_clip.audio.volumex(orig_vol) if final_video_clip.audio else None
            else:
                # Concatenate audio clips
                logger.info(f"Attempting to concatenate {len(audio_clips)} audio clips")
                concatenated_audio = concatenate_audioclips(audio_clips)
                logger.info(f"Concatenated audio duration: {concatenated_audio.duration:.2f}s, channels: {concatenated_audio.nchannels}")
                
                # Verify concatenated duration
                expected_duration = sum(clip.duration for clip in audio_clips)
                if abs(concatenated_audio.duration - expected_duration) > 0.1:
                    logger.warning(f"Concatenated duration ({concatenated_audio.duration:.2f}s) does not match expected duration ({expected_duration:.2f}s)")
                
                # Adjust concatenated audio duration to match video duration (trim or loop)
                if concatenated_audio.duration > video_duration:
                    concatenated_audio = concatenated_audio.subclip(0, video_duration)
                    logger.info(f"Trimmed concatenated audio to match video duration: {concatenated_audio.duration:.2f}s")
                elif concatenated_audio.duration < video_duration:
                    # Loop the audio to match video duration
                    concatenated_audio = concatenated_audio.fx(lambda clip: clip.loop(duration=video_duration))
                    logger.info(f"Looped concatenated audio to match video duration: {concatenated_audio.duration:.2f}s")
                
                # Apply volume to concatenated audio
                concatenated_audio = concatenated_audio.volumex(music_vol)
                
                # Get original video audio (if any) and apply volume
                original_audio = final_video_clip.audio.volumex(orig_vol) if final_video_clip.audio else None
                
                # Composite the audio tracks
                if original_audio:
                    final_audio = CompositeAudioClip([original_audio, concatenated_audio])
                else:
                    final_audio = concatenated_audio
        else:
            logger.info("No audio files provided; using original video audio if available")
            # If no audio files provided, retain original video audio (if any)
            final_audio = final_video_clip.audio.volumex(orig_vol) if final_video_clip.audio else None
        
        # Set the audio to the final video
        final_video_clip = final_video_clip.set_audio(final_audio)
        
        # Write the final video
        logger.info(f"Writing output video to {output_path}")
        final_video_clip.write_videofile(output_path, codec="libx264", fps=30, audio_codec="aac", ffmpeg_params=["-preset", "fast"])
        
        # Close resources
        final_video_clip.close()
        for clip in video_clips:
            clip.close()
        if audio_files and audio_clips:
            for clip in audio_clips:
                clip.close()
            if 'concatenated_audio' in locals():
                concatenated_audio.close()
        
        logger.info("Video merge completed successfully")
        return output_path
    except Exception as e:
        error_msg = f"Error during merging: {str(e)}\n{traceback.format_exc()}"
        logger.error(error_msg)
        return error_msg
    finally:
        # Clean up temporary directory if it was created in this function
        if temp_dir and os.path.exists(temp_dir):
            try:
                shutil.rmtree(temp_dir)
                logger.info(f"Cleaned up temporary directory: {temp_dir}")
            except Exception as e:
                logger.warning(f"Failed to clean up temporary directory {temp_dir}: {str(e)}")

# --- Gradio App Using Blocks ---

def gradio_merge_files(file_binaries, orig_vol, music_vol, file_names=None):
    """
    Gradio endpoint to merge videos and/or audio from binary file uploads.
    Args:
        file_binaries: List of binary data (bytes objects)
        orig_vol: Volume for original video audio (0.0 to 1.0)
        music_vol: Volume for background audio (0.0 to 1.0)
        file_names: List of original filenames (passed separately for API calls)
    Returns:
        Path to the merged file (video or audio) or error message
    """
    try:
        logger.info(f"Received {len(file_binaries) if file_binaries else 0} binary files")
        
        if not file_binaries or len(file_binaries) < 2:
            error_msg = "Error: Please upload at least 2 files."
            logger.error(error_msg)
            return error_msg, None

        # Create a temporary directory to store uploaded files
        temp_dir = tempfile.mkdtemp()
        all_files = []

        # When called via the UI, Gradio provides binary data but not filenames directly.
        # When called via API (e.g., n8n), we need to get filenames from the request.
        # For UI testing, infer filenames based on index; for API, use provided file_names.
        if file_names is None:
            # Fallback for UI: assign temporary filenames (not ideal for production)
            file_names = []
            for idx, binary in enumerate(file_binaries):
                ext = ".mp4" if idx % 2 == 0 else ".mp3"  # Dummy assignment for UI testing
                file_names.append(f"temp_file_{idx}{ext}")
            logger.warning("No file names provided; using temporary names for UI testing. For API calls, pass file_names.")

        if len(file_names) != len(file_binaries):
            error_msg = f"Error: Mismatch between file binaries ({len(file_binaries)}) and file names ({len(file_names)})."
            logger.error(error_msg)
            return error_msg, None

        # Save each binary file to the temporary directory
        for binary, filename in zip(file_binaries, file_names):
            if binary is None:
                logger.warning(f"Skipping None binary data for file {filename}")
                continue
            # Validate filename
            original_filename = os.path.basename(filename)
            if not re.match(r'file\d+\.mp4', original_filename, re.IGNORECASE) and \
               not re.match(r'audio\d+\.(mp3|wav)', original_filename, re.IGNORECASE):
                logger.warning(f"Filename {original_filename} does not match expected pattern; skipping")
                continue
            # Create a temporary file path
            temp_file_path = os.path.join(temp_dir, original_filename)
            # Write the binary data to the temporary file
            with open(temp_file_path, 'wb') as temp_file:
                temp_file.write(binary)
            all_files.append(temp_file_path)
            logger.info(f"Saved uploaded file to {temp_file_path}")

        if len(all_files) < 2:
            error_msg = "Error: Fewer than 2 valid files after filtering."
            logger.error(error_msg)
            return error_msg, None

        # Split files into videos and audios based on extensions
        video_files, audio_files = split_files_by_extension(all_files)
        logger.info(f"Identified {len(video_files)} video files: {video_files}")
        logger.info(f"Identified {len(audio_files)} audio files: {audio_files}")
        
        result = merge_videos_and_audios(
            video_files=video_files,
            audio_files=audio_files,
            orig_vol=orig_vol,
            music_vol=music_vol,
            temp_dir=temp_dir
        )
        
        if isinstance(result, str) and result.startswith("Error"):
            logger.error(result)
            return result, None
        else:
            logger.info(f"Merge successful. Output saved at: {result}")
            # Return appropriate output based on file type
            if result.endswith(".mp3"):
                return None, result  # Audio output
            else:
                return result, None  # Video output
    except Exception as e:
        error_msg = f"Error processing files: {str(e)}\n{traceback.format_exc()}"
        logger.error(error_msg)
        return error_msg, None
    finally:
        # Cleanup is handled in merge_videos_and_audios
        pass

# --- Main Execution ---

if __name__ == "__main__":
    logger.info(f"Environment: {os.environ.get('HUGGINGFACE_SPACES', 'Not in HF Spaces')}")
    logger.info(f"Arguments: {sys.argv}")
    
    # Check ports in a wider range
    default_port = 7860
    ports_to_try = list(range(default_port, default_port + 11))  # 7860 to 7870
    
    selected_port = None
    for port in ports_to_try:
        logger.info(f"Checking if port {port} is available")
        if check_port(port):
            logger.info(f"Port {port} is available")
            selected_port = port
            break
        else:
            logger.warning(f"Port {port} is already in use")
    
    if selected_port is None:
        logger.error("No available ports found in range 7860-7870")
        sys.exit(1)

    logger.info("Launching Gradio Blocks interface")

    with gr.Blocks(title="Video and Audio Merger API") as app:
        gr.Markdown("## Video and Audio Merger API")
        gr.Markdown("Upload at least 2 files total (videos, audios, or a combination) to merge them.")
        gr.Markdown("For API usage, send binary files via multipart/form-data. Name videos as file1.mp4, file2.mp4, etc., and audios as audio1.mp3, audio2.mp3, etc.")
        
        with gr.Row():
            file_input = gr.File(label="Upload Files (Videos: .mp4, Audios: .mp3/.wav)", type="binary", file_count="multiple")
        
        with gr.Row():
            orig_vol_input = gr.Slider(minimum=0.0, maximum=1.0, value=1.0, step=0.05, label="Original Video Audio Volume")
            music_vol_input = gr.Slider(minimum=0.0, maximum=1.0, value=0.5, step=0.05, label="Background Audio Volume")
        
        output_video = gr.Video(label="Merged Video (if videos provided)")
        output_audio = gr.Audio(label="Merged Audio (if only audios provided)")
        merge_button = gr.Button("Merge Files")
        
        merge_button.click(
            fn=gradio_merge_files,
            inputs=[file_input, orig_vol_input, music_vol_input],
            outputs=[output_video, output_audio]
        )

    try:
        logger.info(f"Attempting to launch Gradio app on port {selected_port}")
        app.queue(api_open=True)
        app.launch(server_port=selected_port, share=True)
        logger.info(f"Gradio app launched successfully on port {app.server_port}")
    except Exception as e:
        error_msg = f"Failed to launch Gradio interface: {str(e)}\n{traceback.format_exc()}"
        logger.error(error_msg)
        raise