sterge ghilimelele 2

sterge ghilimelele 2.py

import os
import regex

def remove_quotes_from_meta_tags(html_content):
    # Pattern pentru <meta name="description" content="...">
    pattern1 = regex.compile(r'(<meta name="description" content=")(.*?)(">)', regex.DOTALL)

    # Pattern pentru <meta property="og:description" content="...">
    pattern2 = regex.compile(r'(<meta property="og:description" content=")(.*?)(">)', regex.DOTALL)

    # Pattern pentru JSON "description": "..."
    pattern3 = regex.compile(r'("description": ")(.*?)(",)', regex.DOTALL)

    def clean_content(match):
        # Eliminarea ghilimelelor, apostrofurilor și caracterului '|' din conținut
        cleaned_content = (match.group(2)
                         .replace('„', '')
                         .replace('"', '')
                         .replace('"', '')
                         .replace("'", '')
                         .replace("'", '')
                         .replace("'", '')
                         .replace('|', ''))
        return match.group(1) + cleaned_content + match.group(3)

    # Aplicarea tuturor pattern-urilor
    content = pattern1.sub(clean_content, html_content)
    content = pattern2.sub(clean_content, content)
    content = pattern3.sub(clean_content, content)

    return content

def process_html_files(folder_path):
    for filename in os.listdir(folder_path):
        if filename.lower().endswith('.html'):
            file_path = os.path.join(folder_path, filename)

            print(f'Procesare: {file_path}')

            with open(file_path, 'r', encoding='utf-8') as file:
                content = file.read()

            new_content = remove_quotes_from_meta_tags(content)

            with open(file_path, 'w', encoding='utf-8') as file:
                file.write(new_content)

    print('Toate fișierele au fost modificate.')

# Calea către folderul cu fișierele HTML
folder_path = r'e:\Carte\BB\17 - Site Leadership\alte\Ionel Balauta\Aryeht\Task 1 - Traduce tot site-ul\Doar Google Web\Andreea\Meditatii\2023\OANA\Folder-Oana\test'

# Procesarea fișierelor
process_html_files(folder_path)