src/utils/examples.py

"""
Module for handling example data management and processing.
Provides functionality to load and organize landmark examples.
"""

import json
import os
from collections import OrderedDict

# Define the desired category order
CATEGORY_ORDER = [
    "Historical Sites & Monuments",
    "Modern Architecture & Urban Landmarks",
    "National Parks & Natural Wonders",
    "Bridges",
    "Other"
]

def load_categories():
    """Load landmark categories from JSON file in specified order."""
    with open('landmark_categories.json') as f:
        categories = json.load(f)
    
    # Create an OrderedDict with categories in the specified order
    ordered_categories = OrderedDict()
    for category in CATEGORY_ORDER:
        if category in categories:
            ordered_categories[category] = categories[category]
    
    return ordered_categories

def create_example_data(categories):
    """
    Create organized example data for the interface.
    
    Args:
        categories (dict): Dictionary of landmark categories and their landmarks
    
    Returns:
        dict: Examples organized by category with format {category: [(image_path, label_info)]}
    
    Raises:
        ValueError: If categories is empty or not a dictionary
    """
    if not isinstance(categories, dict) or not categories:
        raise ValueError("Categories must be a non-empty dictionary")

    category_examples = {}
    
    def generate_filename_variations(landmark_name):
        """
        Generate possible filename variations for landmark images.
        Handles different naming conventions (spaces, underscores, etc.)
        """
        base = landmark_name.lower().strip()
        return [
            base.replace(' ', '_'),   # underscore_format
            base.replace(' ', '-'),   # dash-format
            base.replace(' ', ''),    # nocapsnospacesformat
            '_'.join(base.split())    # underscore_all_words
        ]
    
    # Process each category and its landmarks
    for category, landmarks in categories.items():
        category_examples[category] = []
        
        # Find matching image files for each landmark
        for landmark in landmarks:
            if not isinstance(landmark, str):
                continue
                
            # Try different filename variations until a match is found
            for filename in generate_filename_variations(landmark):
                image_path = os.path.join("examples", f"{filename}.jpg")
                if os.path.exists(image_path):
                    # Create label info dictionary for the example
                    label_info = {
                        "label": landmark,
                        "confidences": [(landmark, 1.0)]
                    }
                    category_examples[category].append([image_path, label_info])
                    break
    
    return category_examples