utils.py

import numpy as np
from PIL import Image, ImageDraw
import cv2
import os
import tempfile

def create_demo_mask_from_image(image_shape, center_x=0.5, center_y=0.5, radius=0.3):
    """Create a circular mask in the center of the image"""
    height, width = image_shape[:2]
    center_x = int(width * center_x)
    center_y = int(height * center_y)
    radius = int(min(width, height) * radius)
    
    # Create mask
    mask = np.zeros((height, width), dtype=np.uint8)
    y, x = np.ogrid[:height, :width]
    mask_area = (x - center_x) ** 2 + (y - center_y) ** 2 <= radius ** 2
    mask[mask_area] = 255
    
    return mask

def validate_image_dimensions(image, max_size=2048):
    """Validate and resize image if needed"""
    height, width = image.shape[:2]
    
    if max(height, width) > max_size:
        scale = max_size / max(height, width)
        new_height = int(height * scale)
        new_width = int(width * scale)
        
        resized = cv2.resize(image, (new_width, new_height), interpolation=cv2.INTER_AREA)
        print(f"Image resized from {width}x{height} to {new_width}x{new_height}")
        return resized
    return image

def prepare_image_for_inference(image):
    """Prepare image for inference pipeline"""
    if len(image.shape) == 3 and image.shape[2] == 3:
        # Ensure RGB format
        if image.dtype != np.uint8:
            image = (image * 255).astype(np.uint8)
        return image
    else:
        raise ValueError("Image must be RGB format")

def save_temporary_file(data, suffix=".png"):
    """Save data to a temporary file"""
    if isinstance(data, np.ndarray):
        if len(data.shape) == 2:
            # Grayscale image
            img = Image.fromarray(data, mode='L')
        else:
            # RGB image
            img = Image.fromarray(data)
    else:
        img = data
    
    with tempfile.NamedTemporaryFile(suffix=suffix, delete=False) as temp_file:
        img.save(temp_file.name)
        return temp_file.name

def cleanup_temp_files(temp_paths):
    """Clean up temporary files"""
    for path in temp_paths:
        try:
            if os.path.exists(path):
                os.unlink(path)
        except Exception as e:
            print(f"Warning: Could not delete temporary file {path}: {e}")

def get_inference_status():
    """Check if inference modules are available"""
    try:
        from inference import Inference, load_image, load_single_mask
        return True, "Inference modules available"
    except ImportError as e:
        return False, f"Inference modules not available: {e}"

def format_file_size(size_bytes):
    """Format file size in human readable format"""
    if size_bytes < 1024:
        return f"{size_bytes} B"
    elif size_bytes < 1024**2:
        return f"{size_bytes/1024:.1f} KB"
    elif size_bytes < 1024**3:
        return f"{size_bytes/(1024**2):.1f} MB"
    else:
        return f"{size_bytes/(1024**3):.1f} GB"

def create_sample_mask_options():
    """Create sample mask creation options"""
    return [
        ("No mask", None),
        ("Center circle", "circle_center"),
        ("Center ellipse", "ellipse_center"),
        ("Full image", "full"),
    ]
This Gradio application provides:

## Key Features:

1. **Professional UI**: Modern interface with gradient header and clear sections
2. **Image Upload**: Drag-and-drop or click to upload images
3. **Optional Mask Upload**: Upload segmentation masks for focused processing
4. **Configuration Options**: Adjustable random seed and model selection
5. **Real-time Status**: Progress updates and error handling
6. **Download Functionality**: Direct download of generated 3D models
7. **Demo Mode**: Works even without the inference module installed
8. **Error Handling**: Robust error management with user-friendly messages

## Usage Instructions:

1. **Setup**: Make sure to clone the SAM-3D-objects repository and install dependencies
2. **Image Upload**: Upload the image you want to convert to 3D
3. **Mask (Optional)**: Upload a mask for better segmentation results
4. **Configure**: Adjust the random seed if needed
5. **Generate**: Click the generate button to create your 3D model
6. **Download**: Save the generated PLY file when complete

The app includes a "Built with anycoder" link as requested and provides a complete working interface for image-to-3D conversion using the SAM-3D-objects inference pipeline.

**Important**: Before running, make sure to:
1. Clone the repository: `git clone https://github.com/facebookresearch/sam-3d-objects`
2. Install dependencies as per the repository requirements
3. Ensure the model checkpoints are available in the `checkpoints/` directory