tag_mosaic.py

"""
Progressive Image Reveal Mosaic System for Tag Collector Game

This module provides a completely different approach to the tag mosaic visualization,
treating it as a progressive image reveal where each tag discovery unveils portions
of a template image, with rarer tags revealing more pixels.
"""

import os
import hashlib
import streamlit as st
import numpy as np
import math
import io
import time
import random
from PIL import Image, ImageDraw, ImageFilter
from game_constants import RARITY_LEVELS, ENKEPHALIN_CURRENCY_NAME, ENKEPHALIN_ICON

# Default paths
DEFAULT_TEMPLATES_DIR = "mosaics/templates"
DEFAULT_MOSAICS_DIR = "mosaics"

def ensure_directories():
    """Ensure all required directories exist"""
    # Create mosaics directory if it doesn't exist
    if not os.path.exists(DEFAULT_MOSAICS_DIR):
        os.makedirs(DEFAULT_MOSAICS_DIR)
    
    # Create templates directory if it doesn't exist
    if not os.path.exists(DEFAULT_TEMPLATES_DIR):
        os.makedirs(DEFAULT_TEMPLATES_DIR)

def initialize_mosaic(mosaic_name="main", total_tags=100):
    """
    Initialize the appropriate mosaic type based on the template file.
    
    Args:
        mosaic_name: Name of the mosaic (used for file paths)
        total_tags: Total number of tags expected for this mosaic
        
    Returns:
        An instance of RevealMosaic, AnimatedRevealMosaic, or VideoRevealMosaic
    """

    # Check for GIF template
    gif_path = os.path.join(DEFAULT_TEMPLATES_DIR, f"{mosaic_name}_template.gif")
    if os.path.exists(gif_path):
        # Try to open as GIF and check if animated
        try:
            img = Image.open(gif_path)
            is_animated = hasattr(img, 'n_frames') and img.n_frames > 1
            if is_animated:
                print(f"Found animated GIF template: {gif_path}")
                return AnimatedRevealMosaic(
                    total_tags=total_tags,
                    template_path=gif_path,
                    mosaic_name=mosaic_name
                )
        except Exception as e:
            print(f"Error checking animation: {e}")
    
    # Fallback to standard reveal mosaic
    return RevealMosaic(
        total_tags=total_tags,
        mosaic_name=mosaic_name
    )

class RevealMosaic:
    """Manages the progressive revealing of an image as tags are discovered"""
    
    def __init__(self, 
                total_tags=100, 
                template_path=None, 
                mosaic_name="main",
                save_path=None,
                mask_color=(0, 0, 0)):
        """
        Initialize the reveal mosaic
        
        Args:
            total_tags: Total number of tags expected for this mosaic
            template_path: Path to the template image. If None, generated from mosaic_name
            mosaic_name: Name of the mosaic for storage
            save_path: Path to save the mosaic mask. If None, generated from mosaic_name
            mask_color: Color to use for the mask (default: black)
        """
        # Store basic parameters
        self.total_tags = max(1, total_tags)  # Ensure at least 1 tag
        self.mosaic_name = mosaic_name
        self.mask_color = mask_color
        
        # Ensure directories exist
        ensure_directories()
        
        # Generate paths if not provided
        if save_path is None:
            self.save_path = os.path.join(DEFAULT_MOSAICS_DIR, f"{mosaic_name}_mosaic_mask.png")
        else:
            self.save_path = save_path
            
        if template_path is None:
            self.template_path = os.path.join(DEFAULT_TEMPLATES_DIR, f"{mosaic_name}_template.png")
        else:
            self.template_path = template_path
        
        # Initialize tracking sets
        self.processed_tags = set()  # Tags we've already processed
        self.revealed_pixels = set()  # Pixels that have been revealed
        self.highlighted_tags = []   # Recently added tags
        
        # Load or create template image
        self.template_image = self.load_template_image()
        
        # Get image dimensions
        self.width, self.height = self.template_image.size
        self.total_pixels = self.width * self.height
        
        # Calculate how many pixels each tag should reveal on average
        self.base_pixels_per_tag = self.total_pixels / self.total_tags
        print(f"Base pixels per tag: {self.base_pixels_per_tag:.1f} (Total: {self.total_pixels} pixels, {self.total_tags} tags)")
        
        # Create or load the mask
        self.mask = self.load_or_create_mask()
        
        # Create an image-wide priority map once (for consistent reveals)
        # Only create this when needed to save memory and startup time
        self._priority_map = None
        
        # Track last update time
        self.last_update_time = time.time()
        
        # Flag to indicate if an update is needed
        self.needs_update = False
        
        # Cache for the final image
        self._cached_image = None
        # Flag for whether the cache is valid
        self._cache_valid = False
    
    @property
    def priority_map(self):
        """Lazy-load the priority map only when needed"""
        if self._priority_map is None:
            print(f"Generating priority map for {self.mosaic_name}...")
            self._priority_map = self.create_priority_map()
        return self._priority_map
    
    def load_template_image(self):
        """Load the template image or create a default one"""
        if os.path.exists(self.template_path):
            try:
                img = Image.open(self.template_path).convert('RGB')
                print(f"Loaded template image from {self.template_path}")
                return img
            except Exception as e:
                print(f"Error loading template image: {e}")
        
        # If no image exists or there was an error, create a default
        return self.create_default_template()
    
    def load_or_create_mask(self):
        """Load an existing mask or create a new one"""
        if os.path.exists(self.save_path):
            try:
                mask = Image.open(self.save_path).convert('L')
                
                # Ensure mask size matches template
                if mask.size != self.template_image.size:
                    mask = mask.resize(self.template_image.size)
                    
                print(f"Loaded mask from {self.save_path}")
                
                # Count revealed pixels (where mask is 0/transparent)
                revealed_count = 0
                mask_data = mask.getdata()
                for i, pixel in enumerate(mask_data):
                    if pixel == 0:
                        x = i % self.width
                        y = i // self.width
                        self.revealed_pixels.add((x, y))
                        revealed_count += 1
                        
                print(f"Mask has {revealed_count} revealed pixels out of {self.total_pixels}")
                return mask
            except Exception as e:
                print(f"Error loading mask: {e}")
        
        # Create a new fully opaque mask (nothing revealed)
        mask = Image.new('L', self.template_image.size, 255)
        return mask
    
    def has_new_tags(self, collected_tags):
        """
        Check if there are new tags that haven't been processed yet.
        
        Args:
            collected_tags: Dictionary of {tag_name: info} of collected tags
            
        Returns:
            Tuple of (has_new_tags, count_of_new_tags)
        """
        new_tag_count = 0
        
        # Check for tags that aren't in processed_tags
        for tag, info in collected_tags.items():
            # Only include tags with count > 0
            if info.get("count", 0) <= 0:
                continue
                
            # Count tags we haven't processed yet
            if tag not in self.processed_tags:
                new_tag_count += 1
        
        return (new_tag_count > 0, new_tag_count)
    
    def update_with_tags(self, collected_tags, metadata=None, force_update=False):
        """
        Update the mosaic with newly collected tags.
        
        Args:
            collected_tags: Dictionary of {tag_name: info} of collected tags
            metadata: Optional metadata (not used in this implementation)
            force_update: Force update even if there are no new tags
            
        Returns:
            Number of newly revealed pixels
        """
        # If no force update, quick check for new tags
        has_new, new_count = self.has_new_tags(collected_tags)
        if not has_new and not force_update:
            # No new tags to process
            return 0
        
        # Get start time for performance tracking
        start_time = time.time()
            
        # Clear previous highlights
        self.highlighted_tags = []
        
        # Track if we need to update the mask
        self.needs_update = False
        total_newly_revealed = 0
        
        # Process all tags in the collection if this is first run or force_update
        # Otherwise just process new tags
        known_tags = set(self.processed_tags)
        all_tags_to_process = []
        
        # First pass: determine which tags to process
        for tag, info in collected_tags.items():
            # Only include tags with count > 0
            if info.get("count", 0) <= 0:
                continue
                
            # Mark new tags for processing
            if tag not in known_tags:
                # Get the tag's rarity
                rarity = info.get("rarity", "Canard")  # Default to lowest rarity
                all_tags_to_process.append((tag, rarity, True))  # True = new tag
        
        # If no newly discovered tags, but on first run, process all existing tags
        if not all_tags_to_process and (len(self.processed_tags) == 0 or force_update):
            for tag, info in collected_tags.items():
                if info.get("count", 0) <= 0:
                    continue
                    
                rarity = info.get("rarity", "Canard")
                all_tags_to_process.append((tag, rarity, False))  # False = not new
        
        # If nothing to process, return early
        if not all_tags_to_process:
            return 0
        
        # Sort tags by rarity (process rarest first)
        rarity_order = ["Impuritas Civitas", "Star of the City", "Urban Nightmare", 
                      "Urban Plague", "Urban Legend", "Urban Myth", "Canard"]
        
        def get_rarity_rank(tag_tuple):
            _, rarity, _ = tag_tuple
            if rarity in rarity_order:
                return rarity_order.index(rarity)
            return len(rarity_order)  # Unknown rarity goes at the end
            
        all_tags_to_process.sort(key=get_rarity_rank)
        
        # Process tags
        for tag, rarity, is_new in all_tags_to_process:
            # Reveal pixels for this tag
            newly_revealed = self.reveal_pixels_for_tag(tag, rarity)
            total_newly_revealed += newly_revealed
            
            # Mark this tag as processed
            self.processed_tags.add(tag)
        
        # Update and save the mask if we revealed new pixels
        if self.needs_update:
            self.update_mask()
            # Invalidate the image cache
            self._cache_valid = False
            
        # Update last update time
        self.last_update_time = time.time()
            
        # Report performance
        end_time = time.time()
        print(f"Mosaic update processed {len(all_tags_to_process)} tags, revealed {total_newly_revealed} pixels in {end_time - start_time:.3f}s")
        
        return total_newly_revealed
    
    def reveal_pixels_for_tag(self, tag, rarity):
        """
        Reveal pixels for a newly discovered tag.
        
        Args:
            tag: The tag name
            rarity: The tag's rarity
            
        Returns:
            Number of newly revealed pixels
        """
        # Calculate how many pixels to reveal
        pixels_to_reveal = self.calculate_pixels_to_reveal(rarity)
        
        # Get next set of pixels to reveal from priority map
        newly_revealed = []
        for pixel in self.priority_map:
            if pixel not in self.revealed_pixels:
                newly_revealed.append(pixel)
                self.revealed_pixels.add(pixel)
                if len(newly_revealed) >= pixels_to_reveal:
                    break
        
        # No need to update if no new pixels were revealed
        if not newly_revealed:
            return 0
        
        # Add tag to highlighted tags
        if newly_revealed:
            # Add a random revealed pixel to highlight
            highlight_pixel = random.choice(newly_revealed)
            self.highlighted_tags.append((tag, highlight_pixel[0], highlight_pixel[1], rarity))
            
        # Flag that we need to update the mask
        self.needs_update = True
        
        # Return count of newly revealed pixels
        return len(newly_revealed)
    
    def update_mask(self):
        """Update and save the mask based on current revealed pixels"""
        # Create a new mask image
        new_mask = Image.new('L', (self.width, self.height), 255)
        
        # Reveal pixels (set to 0/transparent where we want the image to show)
        pixels_updated = 0
        for x, y in self.revealed_pixels:
            if 0 <= x < self.width and 0 <= y < self.height:
                new_mask.putpixel((x, y), 0)
                pixels_updated += 1
        
        print(f"Updated mask with {pixels_updated} revealed pixels")
        
        # Save the updated mask
        self.mask = new_mask
        try:
            self.mask.save(self.save_path)
            print(f"Saved mask to {self.save_path}")
        except Exception as e:
            print(f"Error saving mask: {e}")
        
        # Invalidate the image cache since the mask has changed
        self._cache_valid = False
    
    def get_image(self, force_refresh=False):
        """
        Get the current mosaic image (template with mask applied).
        Uses caching to avoid regenerating the image unless needed.
        
        Args:
            force_refresh: Force regeneration of the image even if cached
            
        Returns:
            PIL Image of the current state
        """
        # Check if we need to regenerate the image
        if force_refresh or not self._cache_valid or self._cached_image is None:
            print(f"Regenerating mosaic image for {self.mosaic_name}")
            
            # If the mask file exists but differs from our current mask, reload it
            if os.path.exists(self.save_path):
                try:
                    file_mask = Image.open(self.save_path).convert('L')
                    if file_mask.size == self.mask.size:
                        # Check if the masks are different
                        diff = 0
                        file_data = file_mask.getdata()
                        mask_data = self.mask.getdata()
                        for i, (f_pixel, m_pixel) in enumerate(zip(file_data, mask_data)):
                            if f_pixel != m_pixel:
                                diff += 1
                                
                        if diff > 0:
                            print(f"Mask file differs from memory by {diff} pixels, reloading")
                            self.mask = file_mask
                except Exception as e:
                    print(f"Error comparing masks: {e}")
            
            # Create a copy of the template
            result = self.template_image.copy()
            
            # Create a solid color image for unrevealed areas
            mask_color_img = Image.new('RGB', result.size, self.mask_color)
            
            # Apply the mask (0=transparent, 255=opaque)
            result.paste(mask_color_img, (0, 0), self.mask)
            
            # Cache the result
            self._cached_image = result
            self._cache_valid = True
            
            return result
        else:
            # Return cached image
            return self._cached_image

    def load_or_create_mask(self):
        """Load an existing mask or create a new one"""
        if os.path.exists(self.save_path):
            try:
                mask = Image.open(self.save_path).convert('L')
                
                # Ensure mask size matches template
                if mask.size != self.template_image.size:
                    print(f"Resizing mask from {mask.size} to {self.template_image.size}")
                    mask = mask.resize(self.template_image.size)
                    # Save the resized mask
                    mask.save(self.save_path)
                    
                print(f"Loaded mask from {self.save_path}")
                
                # Count revealed pixels (where mask is 0/transparent)
                revealed_count = 0
                mask_data = mask.getdata()
                for i, pixel in enumerate(mask_data):
                    if pixel == 0:  # Fully transparent
                        x = i % self.width
                        y = i // self.width
                        self.revealed_pixels.add((x, y))
                        revealed_count += 1
                        
                print(f"Loaded mask has {revealed_count} revealed pixels out of {self.total_pixels}")
                return mask
            except Exception as e:
                print(f"Error loading mask: {e}")
        
        # Create a new fully opaque mask (nothing revealed)
        print(f"Creating new mask for {self.mosaic_name}")
        mask = Image.new('L', self.template_image.size, 255)
        # Save the empty mask
        mask.save(self.save_path)
        return mask
    
    def get_stats(self):
        """
        Get statistics about the mosaic completion.
        
        Returns:
            Dictionary with completion statistics
        """
        # Calculate completion percentage
        revealed_count = len(self.revealed_pixels)
        completion_percentage = min(100, (revealed_count / self.total_pixels) * 100)
        
        return {
            "revealed_pixels": revealed_count,
            "total_pixels": self.total_pixels,
            "completion_percentage": completion_percentage,
            "completion_pattern": get_completion_pattern(completion_percentage),
            "newly_highlighted": len(self.highlighted_tags),
            "has_new_tags": len(self.highlighted_tags) > 0
        }
    
    def calculate_pixels_to_reveal(self, rarity):
        """
        Calculate how many pixels to reveal based on tag rarity and total tags.
        
        Args:
            rarity: The rarity of the discovered tag
            
        Returns:
            Number of pixels to reveal
        """
        # Get total tags to distribute pixels among
        # This value should be set correctly during initialization
        total_tags_expected = max(1, self.total_tags)
        
        # Calculate base pixels per tag (equal distribution)
        base_pixels_per_tag = self.total_pixels / total_tags_expected
        
        # Apply rarity multiplier
        rarity_multiplier = 1.0
        if rarity == "Canard":
            rarity_multiplier = 0.5  # Half the average
        elif rarity == "Urban Myth":
            rarity_multiplier = 0.8  # Slightly below average
        elif rarity == "Urban Legend":
            rarity_multiplier = 1.0  # Average
        elif rarity == "Urban Plague":
            rarity_multiplier = 1.5  # Above average
        elif rarity == "Urban Nightmare":
            rarity_multiplier = 2.0  # Double
        elif rarity == "Star of the City":
            rarity_multiplier = 3.0  # Triple
        elif rarity == "Impuritas Civitas":
            rarity_multiplier = 5.0  # Five times
            
        # Calculate pixels to reveal, ensuring at least 100 pixels
        pixels_to_reveal = max(100, int(base_pixels_per_tag * rarity_multiplier))
        
        # Don't reveal too many at once for visual smoothness
        max_at_once = min(100000, int(self.total_pixels * 0.1))  # 10% of total or 100k, whichever is less
        pixels_to_reveal = min(pixels_to_reveal, max_at_once)
        
        # Make sure we don't try to reveal more than what's left
        unrevealed_pixels = self.total_pixels - len(self.revealed_pixels)
        pixels_to_reveal = min(pixels_to_reveal, unrevealed_pixels)
        
        return pixels_to_reveal
    
    def create_priority_map(self):
        """
        Create a priority map for pixel reveal order.
        
        Returns:
            List of pixel coordinates in priority order (highest to lowest)
        """
        # Convert the image to grayscale for brightness analysis
        gray_img = self.template_image.convert('L')
        
        # Get brightness values
        brightness_map = {}
        width, height = gray_img.size
        
        # Create center point and max distance for normalization
        center_x, center_y = width // 2, height // 2
        max_dist = math.sqrt(center_x**2 + center_y**2)
        
        # Calculate priority based on:
        # 1. Distance from center (higher = closer to center)
        # 2. Brightness (higher = brighter parts of image)
        for y in range(height):
            for x in range(width):
                # Distance from center (normalized 0-1)
                dx, dy = x - center_x, y - center_y
                distance = math.sqrt(dx**2 + dy**2)
                distance_factor = 1.0 - (distance / max_dist)
                
                # Brightness (normalized 0-1)
                brightness = gray_img.getpixel((x, y)) / 255.0
                
                # Edge detection factor - look for neighboring pixels with different brightness
                # This helps reveal edges of objects first
                edge_factor = 0.0
                if x > 0 and x < width-1 and y > 0 and y < height-1:
                    # Check surrounding pixels
                    neighbors = [
                        gray_img.getpixel((x-1, y)),
                        gray_img.getpixel((x+1, y)),
                        gray_img.getpixel((x, y-1)),
                        gray_img.getpixel((x, y+1))
                    ]
                    current = gray_img.getpixel((x, y))
                    # Calculate average difference with neighbors
                    diff_sum = sum(abs(current - n) for n in neighbors)
                    edge_factor = min(1.0, diff_sum / (4 * 255.0))
                
                # Calculate priority - weight factors according to importance
                # 40% distance from center, 40% brightness, 20% edge detection
                priority = (distance_factor * 0.4) + (brightness * 0.4) + (edge_factor * 0.2)
                
                # Store in map with a small random factor to prevent exact ties
                random_factor = random.random() * 0.01  # 1% randomness
                brightness_map[(x, y)] = priority + random_factor
        
        # Sort by priority (highest to lowest)
        sorted_pixels = sorted(brightness_map.items(), key=lambda x: x[1], reverse=True)
        
        # Return just the pixel coordinates in order
        return [pixel for pixel, _ in sorted_pixels]
    
    def create_default_template(self, width=1024, height=1024):
        """Create a default colorful template image"""
        # Create a new black canvas
        img = Image.new('RGB', (width, height), (0, 0, 0))
        draw = ImageDraw.Draw(img)
        
        # Draw a colorful circular pattern
        center_x, center_y = width // 2, height // 2
        max_radius = min(width, height) // 2 - 10
        
        # Draw colorful background gradients
        for y in range(height):
            for x in range(width):
                # Calculate distance from center
                dx, dy = x - center_x, y - center_y
                distance = math.sqrt(dx*dx + dy*dy)
                
                # Calculate angle
                angle = math.atan2(dy, dx)
                
                # Normalize distance
                norm_distance = min(1.0, distance / max_radius)
                
                # Create color based on angle and distance
                # This creates a colorful cosmic-like background
                hue = (math.degrees(angle) % 360) / 360.0
                saturation = 0.7 - 0.3 * norm_distance
                value = 0.2 + 0.3 * (1 - norm_distance)
                
                # Convert HSV to RGB
                h = hue * 6
                i = int(h)
                f = h - i
                p = value * (1 - saturation)
                q = value * (1 - saturation * f)
                t = value * (1 - saturation * (1 - f))
                
                if i == 0:
                    r, g, b = value, t, p
                elif i == 1:
                    r, g, b = q, value, p
                elif i == 2:
                    r, g, b = p, value, t
                elif i == 3:
                    r, g, b = p, q, value
                elif i == 4:
                    r, g, b = t, p, value
                else:
                    r, g, b = value, p, q
                
                r, g, b = int(r * 255), int(g * 255), int(b * 255)
                
                # Make the center brighter
                if distance < max_radius * 0.2:
                    # Central brightness
                    brightness = 1.0 - (distance / (max_radius * 0.2))
                    r = min(255, r + int(brightness * (255 - r)))
                    g = min(255, g + int(brightness * (255 - g)))
                    b = min(255, b + int(brightness * (255 - b)))
                
                # Add a bright spot at the center
                img.putpixel((x, y), (r, g, b))
        
        # Draw a central bright spot
        for r in range(50, 0, -1):
            brightness = 1.0 - (r / 50)
            color = (
                min(255, int(200 + brightness * 55)),
                min(255, int(150 + brightness * 105)),
                min(255, int(100 + brightness * 155))
            )
            draw.ellipse((center_x - r, center_y - r, center_x + r, center_y + r), fill=color)
        
        # Apply a slight blur for a smoother appearance
        img = img.filter(ImageFilter.GaussianBlur(radius=1.5))
        
        # Save the template
        img.save(self.template_path)
        print(f"Created default template at {self.template_path}")
        return img

def get_completion_pattern(completion_percentage):
    """
    Get a description of the completion pattern based on percentage.
    
    Args:
        completion_percentage: Percentage of completion (0-100)
        
    Returns:
        String description of what's visible
    """
    if completion_percentage < 1:
        return "the first glimpses of a hidden image"
    elif completion_percentage < 5:
        return "emerging fragments of a mysterious picture"
    elif completion_percentage < 15:
        return "a partial revelation of the concealed artwork"
    elif completion_percentage < 30:
        return "a quarter of the image taking shape"
    elif completion_percentage < 50:
        return "half of the picture becoming clear"
    elif completion_percentage < 75:
        return "most of the image revealed"
    elif completion_percentage < 95:
        return "nearly complete image with just a few hidden details"
    else:
        return "fully revealed artwork"

class AnimatedRevealMosaic(RevealMosaic):
    """Manages the progressive revealing of an animated GIF or video as tags are discovered"""
    
    def __init__(self, 
                total_tags=100, 
                template_path=None, 
                mosaic_name="main",
                save_path=None,
                mask_color=(0, 0, 0)):
        """Initialize with support for animated GIFs"""
        # Store animation-specific attributes
        self.is_animated = False
        self.frames = []
        self.frame_durations = []
        
        # Call parent initializer
        super().__init__(
            total_tags=total_tags,
            template_path=template_path,
            mosaic_name=mosaic_name,
            save_path=save_path,
            mask_color=mask_color
        )
    
    def load_template_image(self):
        """Load the template image with support for animated GIFs"""
        if os.path.exists(self.template_path):
            try:
                # Open the image
                img = Image.open(self.template_path)
                
                # Check if it's an animated GIF
                try:
                    # Get number of frames
                    self.is_animated = hasattr(img, 'n_frames') and img.n_frames > 1
                    
                    if self.is_animated:
                        print(f"Loading animated GIF with {img.n_frames} frames")
                        
                        # Store all frames
                        self.frames = []
                        self.frame_durations = []
                        
                        for frame_idx in range(img.n_frames):
                            img.seek(frame_idx)
                            # Store frame duration
                            self.frame_durations.append(img.info.get('duration', 100))  # Default 100ms
                            # Store frame as RGB
                            self.frames.append(img.convert('RGB').copy())
                        
                        # Return the first frame as the template_image
                        return self.frames[0]
                    else:
                        # Not animated, treat as regular image
                        print(f"Loaded static template image from {self.template_path}")
                        return img.convert('RGB')
                        
                except Exception as e:
                    print(f"Error processing animation: {e}")
                    # Fallback to static image
                    return img.convert('RGB')
                    
            except Exception as e:
                print(f"Error loading template image: {e}")
        
        # If no image exists or there was an error, create a default
        return self.create_default_template()
    
    def get_image(self, force_refresh=False):
        """
        Get the current mosaic image with support for animation.
        
        Args:
            force_refresh: Force regeneration of the image even if cached
            
        Returns:
            PIL Image or list of PIL Images for animated GIFs
        """
        # If not animated, use parent method
        if not self.is_animated or not self.frames:
            return super().get_image(force_refresh)
            
        # Check if we need to regenerate the image
        if force_refresh or not self._cache_valid or self._cached_image is None:
            print(f"Regenerating animated mosaic image for {self.mosaic_name}")
            
            # Create masked frames
            masked_frames = []
            
            # Create a solid color image for unrevealed areas
            mask_color_img = Image.new('RGB', self.frames[0].size, self.mask_color)
            
            # Apply mask to each frame
            for frame in self.frames:
                # Create a copy of the frame
                result_frame = frame.copy()
                
                # Apply the mask (0=transparent, 255=opaque)
                result_frame.paste(mask_color_img, (0, 0), self.mask)
                
                # Add to masked frames
                masked_frames.append(result_frame)
            
            # Store the frames and mark cache as valid
            self._cached_image = masked_frames
            self._cache_valid = True
            
            return masked_frames
        else:
            # Return cached frames
            return self._cached_image

def display_tag_mosaic():
    """Display the tag mosaic in the game UI with progressive image reveal"""
    import streamlit as st
    
    # Create a container for the mosaic display
    with st.container():
        st.subheader("🧩 Tag Collection Mosaic")
        
        # Add loading warning
        st.info("⏳ Note: Initial loading or switching templates may take some time for high-resolution images due to pixel processing.")
        
        # Add an expander for advanced settings
        with st.expander("Mosaic Settings", expanded=False):
            # Only render the uploader if the expander is open
            uploaded_file = st.file_uploader("Upload a custom template image", type=["png", "jpg", "jpeg", "gif"])
            if uploaded_file is not None:
                try:
                    # Process the uploaded image
                    image = Image.open(uploaded_file)
                    
                    # Check if it's an animated GIF
                    is_animated = hasattr(image, 'n_frames') and image.n_frames > 1
                    
                    if is_animated:
                        st.info(f"Animated GIF detected with {image.n_frames} frames. Processing may take longer.")
                    elif image.width * image.height > 2000000:  # More than 2 million pixels
                        st.warning(f"Large image detected ({image.width}x{image.height}). Processing may take longer.")
                    
                    # Create templates directory if it doesn't exist
                    ensure_directories()
                    # Save as template
                    template_path = os.path.join(DEFAULT_TEMPLATES_DIR, "main_template.gif" if is_animated else "main_template.png")
                    image.save(template_path)
                    st.success("Template updated! Tags will now progressively reveal this image.")
                    # Clear the current mask to start fresh
                    mask_path = os.path.join(DEFAULT_MOSAICS_DIR, "main_mosaic_mask.png")
                    if os.path.exists(mask_path):
                        os.remove(mask_path)
                    # Reinitialize the mosaic
                    if 'tag_mosaic' in st.session_state:
                        del st.session_state.tag_mosaic
                except Exception as e:
                    st.error(f"Error processing image: {e}")
        
        # Initialize the mosaic if not already in session state
        if 'tag_mosaic' not in st.session_state:
            # Try to load the total tags count from metadata if available
            total_tags = 70527  # Default
            try:
                if hasattr(st.session_state, 'model') and hasattr(st.session_state.model, 'dataset'):
                    if hasattr(st.session_state.model.dataset, 'tag_to_idx'):
                        total_tags = len(st.session_state.model.dataset.tag_to_idx)
            except Exception as e:
                print(f"Error getting tag count from metadata: {e}")
                
            # Check for animated template and use appropriate class
            template_path = os.path.join(DEFAULT_TEMPLATES_DIR, "main_template.gif")
            if os.path.exists(template_path):
                # Try to open as GIF and check if animated
                try:
                    img = Image.open(template_path)
                    is_animated = hasattr(img, 'n_frames') and img.n_frames > 1
                    if is_animated:
                        # Create the animated reveal mosaic
                        st.session_state.tag_mosaic = AnimatedRevealMosaic(
                            total_tags=total_tags,
                            template_path=template_path,
                            mosaic_name="main"
                        )
                        print("Using AnimatedRevealMosaic")
                    else:
                        # Fallback to standard reveal mosaic
                        st.session_state.tag_mosaic = RevealMosaic(
                            total_tags=total_tags,
                            mosaic_name="main"
                        )
                except Exception as e:
                    print(f"Error checking animation: {e}")
                    # Fallback to standard reveal mosaic
                    st.session_state.tag_mosaic = RevealMosaic(
                        total_tags=total_tags,
                        mosaic_name="main"
                    )
            else:
                # No animated GIF template, use standard reveal mosaic
                st.session_state.tag_mosaic = RevealMosaic(
                    total_tags=total_tags,
                    mosaic_name="main"
                )
        
        # Get the mosaic from session state
        mosaic = st.session_state.tag_mosaic
        
        # Make sure processed_tags is initialized
        if not hasattr(mosaic, 'processed_tags'):
            mosaic.processed_tags = set()
        
        # Display milestone tracker for main collection at the top
        if hasattr(st.session_state, 'collected_tags'):
            # Import the milestone tracker display function if needed
            from series_mosaics import display_milestone_tracker
            display_milestone_tracker("main", st.session_state.collected_tags, mosaic.total_tags)
        
        # Add update button
        update_requested = st.button("🔄 Update Mosaic")
        
        # Display the manually update message 
        if not update_requested:
            st.info("Click the 'Update Mosaic' button to process new tag discoveries and update the image.")
        
        # Update the mosaic with the latest collected tags only if requested
        newly_revealed = 0
        if update_requested and hasattr(st.session_state, 'collected_tags'):
            # Show processing spinner
            with st.spinner("Processing tag discoveries and updating mosaic..."):
                # Get optional metadata if available
                metadata = st.session_state.model.dataset if hasattr(st.session_state, 'model') else None
                newly_revealed = mosaic.update_with_tags(st.session_state.collected_tags, metadata, force_update=True)
                
                # Check for milestone rewards after updating
                from series_mosaics import check_and_award_milestone_rewards  # Import the reward function
                
                # For main collection, use total model tags as the total (if available)
                total_main_tags = mosaic.total_tags
                milestone, reward = check_and_award_milestone_rewards("main", st.session_state.collected_tags, total_main_tags)
                
                # Show appropriate messages based on update results
                if milestone is not None:
                    # Show milestone achievement message with celebration
                    st.balloons()
                    st.success(f"🏆 MILESTONE ACHIEVED! {milestone}% Completion of Main Collection!")
                    st.success(f"Rewarded with {reward} {ENKEPHALIN_ICON} {ENKEPHALIN_CURRENCY_NAME}!")
                    
                    # Force a rerun to update the UI with new enkephalin
                    st.rerun()
                elif newly_revealed > 0:
                    st.success(f"Successfully updated! Revealed {newly_revealed} new pixels.")
                else:
                    st.info("No new pixels to reveal.")
        
        # Get mosaic stats
        stats = mosaic.get_stats()
        
        # Show completion stats
        col1, col2 = st.columns(2)
        with col1:
            st.write(f"**Completion:** {stats['completion_percentage']:.2f}%")
            st.write(f"**Pixels Revealed:** {stats['revealed_pixels']} / {stats['total_pixels']}")
        with col2:
            st.write(f"**Status:** {stats['completion_pattern']}")
            if newly_revealed > 0:
                st.write(f"**Newly Revealed:** {newly_revealed} pixels")
        
        # Display the mosaic image
        mosaic_img = mosaic.get_image()
        
        # Check if we have an animated mosaic
        if hasattr(mosaic, 'is_animated') and mosaic.is_animated and isinstance(mosaic_img, list):
            # Convert animated frames to GIF for display
            img_bytes = io.BytesIO()
            
            # Save as animated GIF
            mosaic_img[0].save(
                img_bytes, 
                format='GIF', 
                save_all=True, 
                append_images=mosaic_img[1:], 
                duration=mosaic.frame_durations, 
                loop=0
            )
            
            img_bytes.seek(0)
            st.image(img_bytes, caption="Your Tag Collection Mosaic - Each discovery reveals more of the hidden image", 
                    use_container_width=True)
        else:
            # Handle static image as before
            img_bytes = io.BytesIO()
            mosaic_img.save(img_bytes, format='PNG')
            img_bytes.seek(0)
            
            st.image(img_bytes, caption="Your Tag Collection Mosaic - Each discovery reveals more of the hidden image", 
                    use_container_width=True)
        
        # Show legend for rarities
        st.write("**Rarity Legend:**")
        cols = st.columns(len(RARITY_LEVELS))
        for i, (rarity, info) in enumerate(RARITY_LEVELS.items()):
            with cols[i]:
                st.markdown(
                    f"<div style='background-color:{info['color']};height:20px;width:20px;display:inline-block;margin-right:5px;'></div> {rarity}",
                    unsafe_allow_html=True
                )
        
        # Show recently added tags
        if mosaic.highlighted_tags:
            with st.expander("Recently Added Tags", expanded=False):
                for tag, _, _, rarity in mosaic.highlighted_tags:
                    color = RARITY_LEVELS.get(rarity, {}).get("color", "#AAAAAA")
                    st.markdown(
                        f"<span style='color:{color};font-weight:bold;'>{tag}</span>",
                        unsafe_allow_html=True
                    )