liferay_scripts/update_robots/update_robots.py

import psycopg2
import re
import argparse
import os
import logging

# Default values
DEFAULT_CONFIG_FILE = "/home/life/Portal-Bundle/portal-setup-wizard.properties"
DEFAULT_TEMPLATE_FILE = "template-robots.txt"
DEFAULT_PLACEHOLDER = "{{HOSTNAME}}"
DEFAULT_OUTPUT_DIR = "currents_robots"
LOG_FILE = "update_robots.log"

# Function to read database connection parameters from the configuration file
def read_db_config(config_file_path):
    config = {}
    with open(config_file_path, 'r') as file:
        for line in file:
            if line.strip() and '=' in line and not line.startswith("#"):
                key, value = line.strip().split('=', 1)
                config[key.strip()] = value.strip()
    return config

# Function to read the content of the template file
def read_template_file(file_path):
    with open(file_path, 'r') as file:
        return file.read().strip().replace('\n', '_SAFE_NEWLINE_CHARACTER_').replace('\r\n', '_SAFE_NEWLINE_CHARACTER_')

# Function to update the typesettings field for all groups containing false-robots.txt
def update_typesettings(cursor, template_content, placeholder_pattern, execute_updates, specific_hostname=None):
    try:
        # Query to extract the necessary data with a join
        query = """
            SELECT g.groupid, g.typesettings, v.hostname
            FROM public.group_ g
            JOIN public.layoutset l ON g.groupid = l.groupid
            JOIN public.virtualhost v ON l.layoutsetid = v.layoutsetid
            WHERE g.typesettings LIKE '%false-robots.txt%'
        """
        if specific_hostname:
            query += " AND v.hostname = %s"
            cursor.execute(query, (specific_hostname,))
        else:
            cursor.execute(query)
        
        rows = cursor.fetchall()

        for row in rows:
            groupid, current_typesettings, hostname = row

            # Replace the placeholder in the template with the hostname value
            new_false_robots = template_content.replace(placeholder_pattern, hostname)

            # Update the value of false-robots.txt
            updated_typesettings = re.sub(
                r'false-robots\.txt=[^¶]*',
                f'false-robots.txt={new_false_robots}',
                current_typesettings,
                flags=re.DOTALL
            )

            # Update query
            update_query = "UPDATE public.group_ SET typesettings = %s WHERE groupid = %s;"

            if execute_updates:
                cursor.execute(update_query, (updated_typesettings, groupid))
                executed_query = cursor.mogrify(update_query, (updated_typesettings, groupid)).decode('utf-8')
                restore_query = cursor.mogrify(update_query, (current_typesettings, groupid)).decode('utf-8')
                logging.info(f"Executed: \n{executed_query}\n")
                logging.info(f"### Restore command: \n{restore_query}\n")
            else:
                print(cursor.mogrify(update_query, (updated_typesettings, groupid)).decode('utf-8'))

    except Exception as e:
        print(f"Error during update: {e}")
        logging.error(f"Error during update: {e}")

# Function to print the list of current vhosts
def print_current_vhosts(cursor):
    try:
        cursor.execute("""
            SELECT v.hostname
            FROM public.virtualhost v
        """)
        rows = cursor.fetchall()
        print("List of current vhosts:")
        for row in rows:
            print(row[0])
    except Exception as e:
        print(f"Error retrieving vhosts: {e}")

# Function to print the list of current false-robots.txt values
def print_current_robots(cursor):
    try:
        cursor.execute("""
            SELECT g.typesettings, v.hostname
            FROM public.group_ g
            JOIN public.layoutset l ON g.groupid = l.groupid
            JOIN public.virtualhost v ON l.layoutsetid = v.layoutsetid
            WHERE g.typesettings LIKE '%false-robots.txt%'
        """)
        rows = cursor.fetchall()
        print("List of current false-robots.txt values and related hostnames:")
        for row in rows:
            match = re.search(r'false-robots\.txt=([^¶]*)', row[0])
            if match:
                print(f"Hostname: {row[1]} - Robots: {match.group(1)}")
    except Exception as e:
        print(f"Error retrieving false-robots.txt values: {e}")

# Function to save the current vhosts to files
def save_current_vhosts(cursor, output_dir):
    try:
        os.makedirs(output_dir, exist_ok=True)
        cursor.execute("""
            SELECT v.hostname, g.typesettings
            FROM public.virtualhost v
            JOIN public.layoutset l ON v.layoutsetid = l.layoutsetid
            JOIN public.group_ g ON l.groupid = g.groupid
            WHERE g.typesettings LIKE '%false-robots.txt%'
        """)
        rows = cursor.fetchall()
        for row in rows:
            hostname, typesettings = row
            file_path = os.path.join(output_dir, f"robots_{hostname}.txt")
            with open(file_path, 'w') as file:
                match = re.search(r'false-robots\.txt=([^¶]*)', typesettings)
                if match:
                    file.write(match.group(1))
                    print(f"Saved {hostname} robots.txt to {file_path}")

            file_path = os.path.join(output_dir, f"typesettings_{hostname}.txt")
            with open(file_path, 'w') as file:
                file.write(typesettings)
            print(f"Saved {hostname} typesettings to {file_path}")
    except Exception as e:
        print(f"Error saving vhosts: {e}")

# Function to print the differences between current and new false-robots.txt values
def print_differences(cursor, template_content, placeholder_pattern, specific_hostname=None):
    try:
        # Query to extract the necessary data with a join
        query = """
            SELECT g.typesettings, v.hostname
            FROM public.group_ g
            JOIN public.layoutset l ON g.groupid = l.groupid
            JOIN public.virtualhost v ON l.layoutsetid = v.layoutsetid
            WHERE g.typesettings LIKE '%false-robots.txt%'
        """
        if specific_hostname:
            query += " AND v.hostname = %s"
            cursor.execute(query, (specific_hostname,))
        else:
            cursor.execute(query)
        
        rows = cursor.fetchall()

        print("Differences between current and new false-robots.txt values:")
        for row in rows:
            current_typesettings, hostname = row
            match = re.search(r'false-robots\.txt=([^¶]*)', current_typesettings)
            if match:
                current_false_robots = match.group(1)
                new_false_robots = template_content.replace(placeholder_pattern, hostname)
                if current_false_robots != new_false_robots:
                    print(f"Hostname: {hostname}")
                    print(f"Current: {current_false_robots}")
                    print(f"New: {new_false_robots}")
                    print("-" * 40)

    except Exception as e:
        print(f"Error during difference check: {e}")

# Main function
def main():
    parser = argparse.ArgumentParser(description="Update the false-robots.txt field in the group_ table.",
                                     formatter_class=argparse.RawTextHelpFormatter,
                                     epilog=f"""
Examples of usage:

1. Use a configuration file for database parameters and print the update queries:
   python3 update_robots.py --config-file {DEFAULT_CONFIG_FILE} --template-file {DEFAULT_TEMPLATE_FILE}

2. Specify database parameters directly from the command line and print the update queries:
   python3 update_robots.py --db-host postgres --db-port 5432 --db-name liferay_db --db-user infra_bundle_dev --db-password pass_db --template-file {DEFAULT_TEMPLATE_FILE}

3. Execute the update queries:
   python3 update_robots.py --config-file {DEFAULT_CONFIG_FILE} --template-file {DEFAULT_TEMPLATE_FILE} --execute

4. Print the list of current vhosts:
   python3 update_robots.py --config-file {DEFAULT_CONFIG_FILE} --list-vhosts

5. Print the list of current false-robots.txt values and related hostnames:
   python3 update_robots.py --config-file {DEFAULT_CONFIG_FILE} --list-robots

6. Save the current vhosts to files:
   python3 update_robots.py --config-file {DEFAULT_CONFIG_FILE} --save-vhosts --output-dir {DEFAULT_OUTPUT_DIR}

7. Update the false-robots.txt for a specific hostname:
   python3 update_robots.py --config-file {DEFAULT_CONFIG_FILE} --template-file {DEFAULT_TEMPLATE_FILE} --hostname specific.hostname.com

8. Print the differences between current and new false-robots.txt values:
   python3 update_robots.py --config-file {DEFAULT_CONFIG_FILE} --template-file {DEFAULT_TEMPLATE_FILE} --print-differences
    """)
    
    parser.add_argument("--config-file", default=DEFAULT_CONFIG_FILE, help=f"Path to the configuration file with database parameters (default: {DEFAULT_CONFIG_FILE})")
    parser.add_argument("--db-host", help="Database host")
    parser.add_argument("--db-port", type=int, help="Database port")
    parser.add_argument("--db-name", help="Database name")
    parser.add_argument("--db-user", help="Database user")
    parser.add_argument("--db-password", help="Database password")
    parser.add_argument("--template-file", default=DEFAULT_TEMPLATE_FILE, help=f"Path to the template file for false-robots.txt (default: {DEFAULT_TEMPLATE_FILE})")
    parser.add_argument("--placeholder", default=DEFAULT_PLACEHOLDER, help=f"Placeholder pattern to replace in the template file (default: {DEFAULT_PLACEHOLDER})")
    parser.add_argument("--execute", action="store_true", help="Execute the update queries (default is to print the queries)")
    parser.add_argument("--print-updates", action="store_true", help="Print only (not execute) the update queries (default is to print the queries)")
    parser.add_argument("--list-vhosts", action="store_true", help="Print the list of current vhosts")
    parser.add_argument("--list-robots", action="store_true", help="Print the list of current false-robots.txt values and related hostnames")
    parser.add_argument("--save-vhosts", action="store_true", help=f"Save the current vhosts to files (default directory: {DEFAULT_OUTPUT_DIR})")
    parser.add_argument("--output-dir", default=DEFAULT_OUTPUT_DIR, help=f"Output directory to save the vhosts (default: {DEFAULT_OUTPUT_DIR})")
    parser.add_argument("--hostname", help="Specific hostname to update")
    parser.add_argument("--print-differences", action="store_true", help="Print the differences between current and new false-robots.txt values")
    
    args = parser.parse_args()

    if not any(vars(args).values()):
        parser.print_help()
        return

    # Read the template file content
    template_content = read_template_file(args.template_file)

    # Initialize database connection parameters with None
    db_host = db_port = db_name = db_user = db_password = None

    # Determine the database connection parameters
    if args.config_file:
        db_config = read_db_config(args.config_file)
        db_host = db_config['jdbc.default.url'].split('/')[2].split(':')[0]
        db_port = int(db_config['jdbc.default.url'].split('/')[2].split(':')[1])
        db_name = db_config['jdbc.default.url'].split('/')[-1]
        db_user = db_config['jdbc.default.username']
        db_password = db_config['jdbc.default.password']
        
    if args.db_host:
        db_host = args.db_host

    if args.db_port:
        db_port = args.db_port

    if args.db_name:
        db_name = args.db_name

    if args.db_user:
        db_user = args.db_user

    if args.db_password:
        db_password = args.db_password

    # Configure logging
    logging.basicConfig(filename=LOG_FILE, level=logging.INFO, 
                        format='%(asctime)s %(message)s')

    # Connect to the database
    conn = psycopg2.connect(
        host=db_host,
        port=db_port,
        dbname=db_name,
        user=db_user,
        password=db_password
    )
    cursor = conn.cursor()

    if args.list_vhosts:
        print_current_vhosts(cursor)
    elif args.list_robots:
        print_current_robots(cursor)
    elif args.save_vhosts:
        save_current_vhosts(cursor, args.output_dir)
    elif args.print_differences:
        print_differences(cursor, template_content, args.placeholder, args.hostname)
    elif args.execute:
        update_typesettings(cursor, template_content, args.placeholder, True, args.hostname)
        print("robots.txt updated. Clean the Liferay database cache on all the instances (Configuration => Server Administration => Clear the database cache)")
    elif args.print_updates:
        update_typesettings(cursor, template_content, args.placeholder, False, args.hostname)
    else:
        parser.print_help()
        return
    
    conn.commit()
    cursor.close()
    conn.close()

if __name__ == "__main__":
    main()