MCP Image Recognition Server

from io import BytesIO from PIL import Image import base64 def process_image_data(image_bytes: bytes, max_size: int = 1536, max_file_size_mb: int = 4) -> str: """ Process raw image bytes: 1. Open with PIL. 2. Resize if larger than max_size (longest edge). 3. Convert/Compress to JPEG to ensure it's within size limits and widely compatible. 4. Return base64 string. """ try: with Image.open(BytesIO(image_bytes)) as img: # Convert to RGB if necessary (handling PNG alpha channel, etc.) if img.mode in ('RGBA', 'LA') or (img.mode == 'P' and 'transparency' in img.info): img = img.convert('RGB') # Resize logic width, height = img.size if width > max_size or height > max_size: ratio = min(max_size / width, max_size / height) new_size = (int(width * ratio), int(height * ratio)) img = img.resize(new_size, Image.Resampling.LANCZOS) # Save to buffer as JPEG buffer = BytesIO() # Start with high quality, lower it if needed (simple logic for now: just use 85) img.save(buffer, format="JPEG", quality=85) # Check size. If > max_file_size_mb, could compress further, # but usually 1536px side at q85 is < 1MB. return base64.b64encode(buffer.getvalue()).decode('utf-8') except Exception as e: # If image processing fails (e.g. invalid image format), raise it up raise ValueError(f"Failed to process image data: {str(e)}")