osm-labo/wiki_compare/suggest_translation.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

"""
suggest_translation.py

This script reads the outdated_pages.json file generated by wiki_compare.py,
identifies English wiki pages that don't have a French translation,
and posts a message on Mastodon suggesting that the page needs translation.

Usage:
    python suggest_translation.py [--dry-run]

Options:
    --dry-run    Run the script without actually posting to Mastodon

Output:
    - A post on Mastodon suggesting a wiki page for translation
    - Log messages about the selected page and posting status
"""

import json
import random
import argparse
import logging
import os
from datetime import datetime
import requests

# Configure logging
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s',
    datefmt='%Y-%m-%d %H:%M:%S'
)
logger = logging.getLogger(__name__)

# Constants
OUTDATED_PAGES_FILE = "outdated_pages.json"
MASTODON_API_URL = "https://mastodon.instance/api/v1/statuses"  # Replace with actual instance
MASTODON_ACCESS_TOKEN = os.environ.get("MASTODON_ACCESS_TOKEN")

def load_outdated_pages():
    """
    Load the outdated pages from the JSON file
    
    Returns:
        list: List of dictionaries containing outdated page information
    """
    try:
        with open(OUTDATED_PAGES_FILE, 'r', encoding='utf-8') as f:
            pages = json.load(f)
        logger.info(f"Successfully loaded {len(pages)} pages from {OUTDATED_PAGES_FILE}")
        return pages
    except (IOError, json.JSONDecodeError) as e:
        logger.error(f"Error loading pages from {OUTDATED_PAGES_FILE}: {e}")
        return []

def find_missing_translations(pages):
    """
    Find English pages that don't have a French translation
    
    Args:
        pages (list): List of dictionaries containing page information
        
    Returns:
        list: List of pages that need translation
    """
    # Filter pages to include only those with a missing French page
    missing_translations = [page for page in pages if 
                           page.get('reason') == 'French page missing' and 
                           page.get('en_page') is not None and 
                           page.get('fr_page') is None]
    
    logger.info(f"Found {len(missing_translations)} pages without French translation")
    return missing_translations

def select_random_page_for_translation(pages):
    """
    Randomly select a page for translation from the list
    
    Args:
        pages (list): List of dictionaries containing page information
        
    Returns:
        dict: Randomly selected page or None if no suitable pages found
    """
    if not pages:
        logger.warning("No pages found that need translation")
        return None
    
    # Randomly select a page
    selected_page = random.choice(pages)
    logger.info(f"Randomly selected page for key '{selected_page['key']}' for translation")
    
    return selected_page

def create_mastodon_post(page):
    """
    Create a Mastodon post suggesting a page for translation
    
    Args:
        page (dict): Dictionary containing page information
        
    Returns:
        str: Formatted Mastodon post text
    """
    key = page['key']
    en_url = page['en_page']['url']
    fr_url = en_url.replace('/wiki/Key:', '/wiki/FR:Key:')
    
    # Get word count and sections from English page
    word_count = page['en_page']['word_count']
    sections = page['en_page']['sections']
    
    # Format the post
    post = f"""🔍 Clé OSM sans traduction française : #{key}

Cette page wiki importante n'a pas encore de traduction française !

📊 Statistiques de la page anglaise :
• {word_count} mots
• {sections} sections

Vous pouvez aider en créant la traduction française ici :
{fr_url}

Page anglaise à traduire :
{en_url}

#OpenStreetMap #OSM #Wiki #Traduction #Contribution"""
    
    return post

def post_to_mastodon(post_text, dry_run=False):
    """
    Post the message to Mastodon
    
    Args:
        post_text (str): Text to post
        dry_run (bool): If True, don't actually post to Mastodon
        
    Returns:
        bool: True if posting was successful or dry run, False otherwise
    """
    if dry_run:
        logger.info("DRY RUN: Would have posted to Mastodon:")
        logger.info(post_text)
        return True
    
    if not MASTODON_ACCESS_TOKEN:
        logger.error("MASTODON_ACCESS_TOKEN environment variable not set")
        return False
    
    headers = {
        "Authorization": f"Bearer {MASTODON_ACCESS_TOKEN}",
        "Content-Type": "application/json"
    }
    
    data = {
        "status": post_text,
        "visibility": "public"
    }
    
    try:
        response = requests.post(MASTODON_API_URL, headers=headers, json=data)
        response.raise_for_status()
        logger.info("Successfully posted to Mastodon")
        return True
    except requests.exceptions.RequestException as e:
        logger.error(f"Error posting to Mastodon: {e}")
        return False

def main():
    """Main function to execute the script"""
    parser = argparse.ArgumentParser(description="Suggest an OSM wiki page for translation on Mastodon")
    parser.add_argument("--dry-run", action="store_true", help="Run without actually posting to Mastodon")
    args = parser.parse_args()
    
    logger.info("Starting suggest_translation.py")
    
    # Load pages
    pages = load_outdated_pages()
    if not pages:
        logger.error("No pages found. Run wiki_compare.py first.")
        return
    
    # Find pages that need translation
    pages_for_translation = find_missing_translations(pages)
    if not pages_for_translation:
        logger.error("No pages found that need translation.")
        return
    
    # Select a random page for translation
    selected_page = select_random_page_for_translation(pages_for_translation)
    if not selected_page:
        logger.error("Could not select a page for translation.")
        return
    
    # Create the post text
    post_text = create_mastodon_post(selected_page)
    
    # Post to Mastodon
    success = post_to_mastodon(post_text, args.dry_run)
    
    if success:
        logger.info("Script completed successfully")
    else:
        logger.error("Script completed with errors")

if __name__ == "__main__":
    main()