import cv2
import numpy as np
from PIL import Image, ImageDraw, ImageFont
from typing import List, Tuple, Dict

class BrandVisualizationManager:
    """Visualize detected brands with bounding boxes and labels (like YOLO)"""

    def __init__(self):
        """Initialize visualization manager"""
        # Color palette for different brand categories
        self.colors = {
            'luxury': (218, 165, 32),      # Gold
            'sportswear': (0, 191, 255),   # Deep Sky Blue
            'tech': (169, 169, 169),       # Dark Gray
            'automotive': (220, 20, 60),   # Crimson
            'fashion': (186, 85, 211),     # Medium Orchid
            'watches': (184, 134, 11),     # Dark Goldenrod
            'default': (0, 255, 0)         # Green
        }

        print("✓ Brand Visualization Manager initialized")

    def draw_brand_detections(self, image: Image.Image, brand_detections: List[Dict],
                              show_confidence: bool = True) -> Image.Image:
        """Draw bounding boxes and labels for detected brands

        Args:
            image: PIL Image
            brand_detections: List of dicts with keys: 'name', 'confidence', 'bbox', 'category'
            show_confidence: Whether to show confidence scores

        Returns:
            Image with drawn bounding boxes
        """
        if not brand_detections:
            return image

        # Convert PIL to OpenCV format
        img_cv = cv2.cvtColor(np.array(image), cv2.COLOR_RGB2BGR)

        for detection in brand_detections:
            brand_name = detection.get('name', 'Unknown')
            confidence = detection.get('confidence', 0.0)
            bbox = detection.get('bbox')
            category = detection.get('category', 'default')

            if bbox is None:
                continue

            x1, y1, x2, y2 = bbox
            color = self.colors.get(category, self.colors['default'])

            # Draw bounding box
            cv2.rectangle(img_cv, (int(x1), int(y1)), (int(x2), int(y2)), color, 3)

            # Prepare label text
            if show_confidence:
                label = f"{brand_name} {confidence:.2f}"
            else:
                label = brand_name

            # Calculate text size
            font = cv2.FONT_HERSHEY_SIMPLEX
            font_scale = 0.7
            thickness = 2
            (text_width, text_height), baseline = cv2.getTextSize(label, font, font_scale, thickness)

            # Draw label background
            cv2.rectangle(img_cv,
                         (int(x1), int(y1) - text_height - 10),
                         (int(x1) + text_width + 10, int(y1)),
                         color, -1)

            # Draw label text
            cv2.putText(img_cv, label,
                       (int(x1) + 5, int(y1) - 5),
                       font, font_scale, (255, 255, 255), thickness, cv2.LINE_AA)

        # Convert back to PIL
        img_pil = Image.fromarray(cv2.cvtColor(img_cv, cv2.COLOR_BGR2RGB))
        return img_pil

    def format_brand_list(self, brand_detections: List[Dict]) -> str:
        """Format brand detections as readable text

        Args:
            brand_detections: List of brand detection dicts

        Returns:
            Formatted string with brand names and confidences
        """
        if not brand_detections:
            return "No brands identified"

        formatted = []
        for detection in brand_detections:
            brand_name = detection.get('name', 'Unknown')
            confidence = detection.get('confidence', 0.0)
            # 移除 category 標籤顯示，保持簡潔

            formatted.append(f"{brand_name} ({confidence:.2f})")

        return ", ".join(formatted)

print("✓ BrandVisualizationManager defined")