seo/scripts/config.py

"""
Configuration module for WordPress SEO automation.
Loads and validates environment variables and YAML configuration.
"""

import os
import yaml
from dotenv import load_dotenv
from pathlib import Path

# Load environment variables from .env file
load_dotenv()

class Config:
    """Configuration class for WordPress SEO automation."""

    # Load configuration from YAML file
    CONFIG_FILE = Path(__file__).parent.parent / 'config.yaml'
    
    if CONFIG_FILE.exists():
        with open(CONFIG_FILE, 'r', encoding='utf-8') as f:
            YAML_CONFIG = yaml.safe_load(f)
    else:
        YAML_CONFIG = {}

    # WordPress Settings (Primary site)
    WORDPRESS_URL = os.getenv('WORDPRESS_URL', YAML_CONFIG.get('primary_site', {}).get('url', '')).rstrip('/')
    WORDPRESS_USERNAME = os.getenv('WORDPRESS_USERNAME', YAML_CONFIG.get('primary_site', {}).get('username', ''))
    WORDPRESS_APP_PASSWORD = os.getenv('WORDPRESS_APP_PASSWORD', YAML_CONFIG.get('primary_site', {}).get('password', ''))

    # Multi-site WordPress Configuration
    WORDPRESS_SITES = {
        'mistergeek.net': {
            'url': os.getenv('WORDPRESS_MISTERGEEK_URL', YAML_CONFIG.get('wordpress_sites', {}).get('mistergeek.net', {}).get('url', 'https://www.mistergeek.net')),
            'username': os.getenv('WORDPRESS_MISTERGEEK_USERNAME', os.getenv('WORDPRESS_USERNAME', YAML_CONFIG.get('wordpress_sites', {}).get('mistergeek.net', {}).get('username', ''))),
            'password': os.getenv('WORDPRESS_MISTERGEEK_PASSWORD', os.getenv('WORDPRESS_APP_PASSWORD', YAML_CONFIG.get('wordpress_sites', {}).get('mistergeek.net', {}).get('password', ''))),
        },
        'webscroll.fr': {
            'url': os.getenv('WORDPRESS_WEBSCROLL_URL', YAML_CONFIG.get('wordpress_sites', {}).get('webscroll.fr', {}).get('url', 'https://www.webscroll.fr')),
            'username': os.getenv('WORDPRESS_WEBSCROLL_USERNAME', os.getenv('WORDPRESS_USERNAME', YAML_CONFIG.get('wordpress_sites', {}).get('webscroll.fr', {}).get('username', ''))),
            'password': os.getenv('WORDPRESS_WEBSCROLL_PASSWORD', os.getenv('WORDPRESS_APP_PASSWORD', YAML_CONFIG.get('wordpress_sites', {}).get('webscroll.fr', {}).get('password', ''))),
        },
        'hellogeek.net': {
            'url': os.getenv('WORDPRESS_HELLOGEEK_URL', YAML_CONFIG.get('wordpress_sites', {}).get('hellogeek.net', {}).get('url', 'https://www.hellogeek.net')),
            'username': os.getenv('WORDPRESS_HELLOGEEK_USERNAME', os.getenv('WORDPRESS_USERNAME', YAML_CONFIG.get('wordpress_sites', {}).get('hellogeek.net', {}).get('username', ''))),
            'password': os.getenv('WORDPRESS_HELLOGEEK_PASSWORD', os.getenv('WORDPRESS_APP_PASSWORD', YAML_CONFIG.get('wordpress_sites', {}).get('hellogeek.net', {}).get('password', ''))),
        }
    }

    # OpenRouter API Settings
    OPENROUTER_API_KEY = os.getenv('OPENROUTER_API_KEY', YAML_CONFIG.get('ai_model', {}).get('api_key', ''))
    AI_MODEL = os.getenv('AI_MODEL', YAML_CONFIG.get('ai_model', {}).get('name', 'anthropic/claude-3.5-sonnet'))

    # Script Settings
    BATCH_SIZE = int(os.getenv('BATCH_SIZE', str(YAML_CONFIG.get('script_settings', {}).get('batch_size', 100))))
    API_DELAY_SECONDS = float(os.getenv('API_DELAY_SECONDS', str(YAML_CONFIG.get('script_settings', {}).get('api_delay_seconds', 0.5))))

    # Analysis Settings
    ANALYSIS_MIN_POSITION = int(os.getenv('ANALYSIS_MIN_POSITION', str(YAML_CONFIG.get('analysis_settings', {}).get('min_position', 11))))
    ANALYSIS_MAX_POSITION = int(os.getenv('ANALYSIS_MAX_POSITION', str(YAML_CONFIG.get('analysis_settings', {}).get('max_position', 30))))
    ANALYSIS_MIN_IMPRESSIONS = int(os.getenv('ANALYSIS_MIN_IMPRESSIONS', str(YAML_CONFIG.get('analysis_settings', {}).get('min_impressions', 50))))
    ANALYSIS_TOP_N_POSTS = int(os.getenv('ANALYSIS_TOP_N_POSTS', str(YAML_CONFIG.get('analysis_settings', {}).get('top_n_posts', 20))))

    # Output directory
    OUTPUT_DIR = Path(os.getenv('OUTPUT_DIR', YAML_CONFIG.get('output_settings', {}).get('output_dir', './output')))

    @classmethod
    def validate(cls):
        """Validate that all required configuration is present."""
        errors = []

        if not cls.WORDPRESS_URL:
            errors.append("WORDPRESS_URL is required")

        if not cls.WORDPRESS_USERNAME:
            errors.append("WORDPRESS_USERNAME is required")

        if not cls.WORDPRESS_APP_PASSWORD:
            errors.append("WORDPRESS_APP_PASSWORD is required")

        if not cls.OPENROUTER_API_KEY:
            errors.append("OPENROUTER_API_KEY is required (get one from https://openrouter.ai/)")

        if errors:
            raise ValueError("Configuration errors:\n" + "\n".join(f"  - {e}" for e in errors))

        # Create output directory if it doesn't exist
        cls.OUTPUT_DIR.mkdir(exist_ok=True)

        return True

    @classmethod
    def get_wordpress_auth(cls):
        """Get WordPress authentication tuple."""
        return (cls.WORDPRESS_USERNAME, cls.WORDPRESS_APP_PASSWORD)

    @classmethod
    def get_api_base_url(cls):
        """Get WordPress REST API base URL."""
        return f"{cls.WORDPRESS_URL}/wp-json/wp/v2"

    @classmethod
    def get_site_config(cls, site_name):
        """Get configuration for a specific site."""
        return cls.WORDPRESS_SITES.get(site_name, {})

    @classmethod
    def get_all_sites(cls):
        """Get all configured WordPress sites."""
        return cls.WORDPRESS_SITES.keys()