import gradio as gr import os import shutil import PIL from PIL import Image import py7zr import tarfile import zipfile import ffmpeg from pathlib import Path import tempfile from PIL import ImageDraw def generate_thumbnail(input_video, time_position=0): if input_video is None: return None temp_dir = tempfile.mkdtemp() output_path = os.path.join(temp_dir, "thumbnail.jpg") stream = ffmpeg.input(input_video.name, ss=time_position) stream = ffmpeg.output(stream, output_path, vframes=1) ffmpeg.run(stream, overwrite_output=True) return output_path def extract_audio(input_video, output_format="mp3"): if input_video is None: return None temp_dir = tempfile.mkdtemp() output_path = os.path.join(temp_dir, f"extracted_audio.{output_format}") stream = ffmpeg.input(input_video.name) stream = ffmpeg.output(stream, output_path, acodec='libmp3lame') ffmpeg.run(stream, overwrite_output=True) return output_path def trim_video(input_video, start_time, end_time): if input_video is None: return None temp_dir = tempfile.mkdtemp() output_path = os.path.join(temp_dir, "trimmed_video.mp4") stream = ffmpeg.input(input_video.name, ss=start_time, t=end_time-start_time) stream = ffmpeg.output(stream, output_path, acodec='copy', vcodec='copy') ffmpeg.run(stream, overwrite_output=True) return output_path def crop_image(input_image, x1, y1, x2, y2): if input_image is None: return None temp_dir = tempfile.mkdtemp() img = Image.open(input_image.name) cropped = img.crop((x1, y1, x2, y2)) output_path = os.path.join(temp_dir, f"cropped_image.{input_image.name.split('.')[-1]}") cropped.save(output_path) return output_path def add_watermark(input_image, watermark_text, position="bottom-right"): if input_image is None: return None temp_dir = tempfile.mkdtemp() img = Image.open(input_image.name) draw = ImageDraw.Draw(img) # Basic font settings - you can enhance this with custom fonts font_size = int(min(img.width, img.height) * 0.05) text_width = len(watermark_text) * font_size * 0.6 text_height = font_size # Position mapping positions = { "top-left": (10, 10), "top-right": (img.width - text_width - 10, 10), "bottom-left": (10, img.height - text_height - 10), "bottom-right": (img.width - text_width - 10, img.height - text_height - 10) } x, y = positions.get(position, positions["bottom-right"]) draw.text((x, y), watermark_text, fill=(255, 255, 255, 128)) output_path = os.path.join(temp_dir, f"watermarked_image.{input_image.name.split('.')[-1]}") img.save(output_path) return output_path def convert_image(input_image, output_format, compression_level): if input_image is None: return None temp_dir = tempfile.mkdtemp() output_path = os.path.join(temp_dir, f"converted_image.{output_format.lower()}") img = Image.open(input_image.name) if output_format.lower() == 'jpeg' and img.mode == 'RGBA': background = Image.new('RGB', img.size, (255, 255, 255)) background.paste(img, mask=img.split()[3]) img = background # Quality ranges from 1 (worst) to 95 (best) quality = max(1, min(95, int(95 - (compression_level * 0.94)))) if output_format.lower() in ['jpeg', 'webp']: img.save(output_path, quality=quality, optimize=True) elif output_format.lower() == 'png': img.save(output_path, optimize=True, compression_level=int(compression_level/10)) else: img.save(output_path) return output_path def compress_video(input_video, compression_level, output_format="mp4"): if input_video is None: return None temp_dir = tempfile.mkdtemp() output_path = os.path.join(temp_dir, f"compressed_video.{output_format}") # CRF scale is 0-51 (lower means better quality, higher means more compression) crf = int(51 * (compression_level / 100)) stream = ffmpeg.input(input_video.name) stream = ffmpeg.output(stream, output_path, vcodec='libx265', crf=crf, acodec='aac') ffmpeg.run(stream, overwrite_output=True) return output_path def compress_files(input_files, compression_format, compression_level): if not input_files: return None temp_dir = tempfile.mkdtemp() timestamp = Path(input_files[0].name).stem if compression_format == "zip": output_path = os.path.join(temp_dir, f"compressed_{timestamp}.zip") compression = zipfile.ZIP_DEFLATED compresslevel = int(9 * compression_level / 100) # 0-9 scale with zipfile.ZipFile(output_path, 'w', compression=compression, compresslevel=compresslevel) as zipf: for file in input_files: zipf.write(file.name, os.path.basename(file.name)) if compression_format == "7z": output_path = os.path.join(temp_dir, f"compressed_{timestamp}.7z") # Map compression level to preset filters compression_filters = [ {'id': py7zr.FILTER_LZMA2, 'preset': int(9 * compression_level / 100)} ] with py7zr.SevenZipFile(output_path, 'w', filters=compression_filters) as szf: for file in input_files: szf.write(file.name, os.path.basename(file.name)) elif compression_format == "tar.gz": output_path = os.path.join(temp_dir, f"compressed_{timestamp}.tar.gz") compresslevel = int(9 * compression_level / 100) # 0-9 scale with tarfile.open(output_path, "w:gz", compresslevel=compresslevel) as tar: for file in input_files: tar.add(file.name, arcname=os.path.basename(file.name)) elif compression_format == "tar.xz": output_path = os.path.join(temp_dir, f"compressed_{timestamp}.tar.xz") preset = int(9 * compression_level / 100) # 0-9 scale with tarfile.open(output_path, "w:xz", preset=preset) as tar: for file in input_files: tar.add(file.name, arcname=os.path.basename(file.name)) return output_path def resize_image(input_image, width, height, maintain_aspect): if input_image is None: return None temp_dir = tempfile.mkdtemp() img = Image.open(input_image.name) if maintain_aspect: # Calculate new dimensions maintaining aspect ratio aspect_ratio = img.width / img.height if width: height = int(width / aspect_ratio) else: width = int(height * aspect_ratio) # Ensure we have valid dimensions width = width if width else img.width height = height if height else img.height resized_img = img.resize((width, height), Image.Resampling.LANCZOS) output_path = os.path.join(temp_dir, f"resized_image.{input_image.name.split('.')[-1]}") resized_img.save(output_path) return output_path # Create the Gradio interface with gr.Blocks(title="Media Toolbox") as app: gr.Markdown("""# Universal Media Toolbox (For all Your Ai Media Needs ✨✨) #### Note: Higher compression values will significantly reduce file quality and size. #### Recommended compression levels: #### - Images: 20-40 for good balance between quality and size #### - Videos: 30-50 for reasonable quality #### - Files: 50-70 for general use #### Only use maximum compression (90-100) when file size is critical and quality loss is acceptable """) with gr.Tab("Image Conversion"): with gr.Row(): with gr.Column(): image_input = gr.File(label="Upload Image") image_format = gr.Dropdown( choices=["PNG", "JPEG", "WEBP", "GIF", "BMP"], value="PNG", label="Convert to Format" ) image_compression = gr.Slider( minimum=0, maximum=100, value=20, label="Compression Level (0: Highest Quality, 100: Maximum Compression)" ) image_convert_btn = gr.Button("Convert Image") with gr.Column(): image_output = gr.File(label="Converted Image") with gr.Tab("Image Resize"): with gr.Row(): with gr.Column(): resize_input = gr.File(label="Upload Image") with gr.Row(): width = gr.Number(label="Width (pixels)", value=800) height = gr.Number(label="Height (pixels)", value=600) maintain_aspect = gr.Checkbox(label="Maintain Aspect Ratio", value=True) resize_btn = gr.Button("Resize Image") with gr.Column(): resize_output = gr.File(label="Resized Image") with gr.Tab("Image Crop"): with gr.Row(): with gr.Column(): crop_input = gr.File(label="Upload Image") with gr.Row(): x1 = gr.Number(label="Left", value=0) y1 = gr.Number(label="Top", value=0) x2 = gr.Number(label="Right", value=100) y2 = gr.Number(label="Bottom", value=100) crop_btn = gr.Button("Crop Image") with gr.Column(): crop_output = gr.File(label="Cropped Image") with gr.Tab("Image Watermark"): with gr.Row(): with gr.Column(): watermark_input = gr.File(label="Upload Image") watermark_text = gr.Textbox(label="Watermark Text") position = gr.Dropdown( choices=["top-left", "top-right", "bottom-left", "bottom-right"], value="bottom-right", label="Position" ) watermark_btn = gr.Button("Add Watermark") with gr.Column(): watermark_output = gr.File(label="Watermarked Image") with gr.Tab("Video Compression"): with gr.Row(): with gr.Column(): video_input = gr.File(label="Upload Video") compression_level = gr.Slider( minimum=0, maximum=100, value=50, label="Compression Level (0: Highest Quality, 100: Highest Compression)" ) video_compress_btn = gr.Button("Compress Video") with gr.Column(): video_output = gr.File(label="Compressed Video") with gr.Tab("Video Trim"): with gr.Row(): with gr.Column(): trim_input = gr.File(label="Upload Video") with gr.Row(): start_time = gr.Number(label="Start Time (seconds)", value=0) end_time = gr.Number(label="End Time (seconds)", value=10) trim_btn = gr.Button("Trim Video") with gr.Column(): trim_output = gr.File(label="Trimmed Video") with gr.Tab("Audio Extraction"): with gr.Row(): with gr.Column(): audio_input = gr.File(label="Upload Video") audio_format = gr.Dropdown( choices=["mp3", "wav", "aac"], value="mp3", label="Audio Format" ) extract_btn = gr.Button("Extract Audio") with gr.Column(): audio_output = gr.File(label="Extracted Audio") with gr.Tab("Video Thumbnail"): with gr.Row(): with gr.Column(): thumbnail_input = gr.File(label="Upload Video") time_pos = gr.Number(label="Time Position (seconds)", value=0) thumbnail_btn = gr.Button("Generate Thumbnail") with gr.Column(): thumbnail_output = gr.File(label="Video Thumbnail") with gr.Tab("File Compression"): with gr.Row(): with gr.Column(): files_input = gr.File(label="Upload Files", file_count="multiple") compression_format = gr.Dropdown( choices=["zip", "7z", "tar.gz", "tar.xz"], value="zip", label="Compression Format" ) file_compression = gr.Slider( minimum=0, maximum=100, value=50, label="Compression Level (0: Fastest/Largest, 100: Slowest/Smallest)" ) files_compress_btn = gr.Button("Compress Files") with gr.Column(): files_output = gr.File(label="Compressed Files") # Connect all interface components image_convert_btn.click(convert_image, inputs=[image_input, image_format, image_compression], outputs=image_output) resize_btn.click(resize_image, inputs=[resize_input, width, height, maintain_aspect], outputs=resize_output) crop_btn.click(crop_image, inputs=[crop_input, x1, y1, x2, y2], outputs=crop_output) watermark_btn.click(add_watermark, inputs=[watermark_input, watermark_text, position], outputs=watermark_output) video_compress_btn.click(compress_video, inputs=[video_input, compression_level], outputs=video_output) trim_btn.click(trim_video, inputs=[trim_input, start_time, end_time], outputs=trim_output) extract_btn.click(extract_audio, inputs=[audio_input, audio_format], outputs=audio_output) thumbnail_btn.click(generate_thumbnail, inputs=[thumbnail_input, time_pos], outputs=thumbnail_output) files_compress_btn.click(compress_files, inputs=[files_input, compression_format, file_compression], outputs=files_output) app.launch(ssr_mode=False,server_port=7860,)