up wiki compare

wiki_compare/wiki_compare.py

@@ -26,6 +26,8 @@ import os
 from datetime import datetime
 from bs4 import BeautifulSoup
 import logging
+import matplotlib.pyplot as plt
+import numpy as np
 
 # Configure logging
 logging.basicConfig(
@@ -42,6 +44,7 @@ WIKI_BASE_URL_FR = "https://wiki.openstreetmap.org/wiki/FR:Key:"
 TOP_KEYS_FILE = "top_keys.json"
 WIKI_PAGES_CSV = "wiki_pages.csv"
 OUTDATED_PAGES_FILE = "outdated_pages.json"
+STALENESS_HISTOGRAM_FILE = "staleness_histogram.png"
 # Number of wiki pages to examine
 NUM_WIKI_PAGES = 100
 
@@ -255,6 +258,67 @@ def fetch_wiki_page(key, language='en'):
         logger.error(f"Error fetching wiki page for key '{key}' in {language}: {e}")
         return None
 
+def generate_staleness_histogram(wiki_pages):
+    """
+    Generate a histogram of staleness scores by 10% ranges
+    
+    Args:
+        wiki_pages (list): List of dictionaries containing page information with staleness scores
+        
+    Returns:
+        None: Saves the histogram to a file
+    """
+    logger.info("Generating histogram of staleness scores by 10% ranges...")
+    
+    # Extract staleness scores
+    staleness_scores = []
+    for page in wiki_pages:
+        if page and 'staleness_score' in page:
+            staleness_scores.append(page['staleness_score'])
+    
+    if not staleness_scores:
+        logger.warning("No staleness scores found. Cannot generate histogram.")
+        return
+    
+    # Determine the maximum score for binning
+    max_score = max(staleness_scores)
+    # Round up to the nearest 10 to ensure all scores are included
+    max_bin_edge = np.ceil(max_score / 10) * 10
+    
+    # Create bins for 10% ranges
+    bins = np.arange(0, max_bin_edge + 10, 10)
+    
+    # Count scores in each bin
+    hist, bin_edges = np.histogram(staleness_scores, bins=bins)
+    
+    # Create histogram
+    plt.figure(figsize=(12, 6))
+    
+    # Create bar chart
+    plt.bar(range(len(hist)), hist, align='center')
+    
+    # Set x-axis labels for each bin
+    bin_labels = [f"{int(bin_edges[i])}-{int(bin_edges[i+1])}%" for i in range(len(bin_edges)-1)]
+    plt.xticks(range(len(hist)), bin_labels, rotation=45)
+    
+    # Set labels and title
+    plt.xlabel('Tranches de score de décrépitude (en %)')
+    plt.ylabel('Nombre de pages')
+    plt.title('Répartition du score de décrépitude par tranches de 10%')
+    
+    # Add grid for better readability
+    plt.grid(axis='y', linestyle='--', alpha=0.7)
+    
+    # Adjust layout
+    plt.tight_layout()
+    
+    # Save figure
+    plt.savefig(STALENESS_HISTOGRAM_FILE)
+    logger.info(f"Histogram saved to {STALENESS_HISTOGRAM_FILE}")
+    
+    # Close the figure to free memory
+    plt.close()
+
 def analyze_wiki_pages(pages):
     """
     Analyze wiki pages to determine which ones need updating
@@ -621,6 +685,9 @@ def main():
             fr_page['staleness_score'] = 0
             processed_wiki_pages.append(fr_page)
     
+    # Generate histogram of staleness scores
+    generate_staleness_histogram(processed_wiki_pages)
+    
     # Save processed wiki pages to CSV
     try:
         with open(WIKI_PAGES_CSV, 'w', newline='', encoding='utf-8') as f: