sisvietnamvn_01/sisvietnamvn_main/fix_specialties.py

import oracledb
from bs4 import BeautifulSoup
import re
import xml.etree.ElementTree as ET

def transform_html(raw_html):
    if not raw_html:
        return "", None, None, None, None

    soup = BeautifulSoup(raw_html, "html.parser")
    
    # Extract all images
    images = []
    for img in soup.find_all('img'):
        src = img.get('src')
        if src:
            images.append(src)
        img.decompose() # Remove from flow
        
    hero_image = images[0] if len(images) > 0 else "https://sisvietnam.vn/wp-content/uploads/2021/07/DSC07517-1024x576.jpg"
    other_images = images[1:]

    # Extract Contact Info
    text = soup.get_text()
    email_match = re.search(r"(?i)Email\s*:\s*([\w\.-]+@[\w\.-]+)", text)
    phone_match = re.search(r"(?i)(?:SĐT|Điện thoại|ĐT|ĐIỆN THOẠI)\s*:\s*([\d\.\s-]+)", text)
    address_match = re.search(r"(?i)Địa chỉ\s*:\s*([^\n]+)", text)
    
    contact_email = email_match.group(1).strip() if email_match else None
    contact_phone = phone_match.group(1).strip() if phone_match else None
    contact_address = address_match.group(1).strip() if address_match else None

    # Parse sections based on wp:list or headings
    sections_data = []
    
    top_uls = soup.find_all('ul', recursive=False)
    if not top_uls and soup.find('div', class_='wp-content-container'):
        top_uls = soup.find('div', class_='wp-content-container').find_all('ul', recursive=False)

    for ul in top_uls:
        heading = "Nội dung"
        strong_tag = ul.find('strong')
        if strong_tag:
            heading = strong_tag.text.strip()
            strong_tag.decompose()
        
        content_html = ""
        for li in ul.find_all('li', recursive=False):
            inner_uls = li.find_all('ul')
            if inner_uls:
                for inner in inner_uls:
                    content_html += str(inner)
            else:
                content_html += str(li) # fallback to just raw text if no inner UL

        if not content_html:
            content_html = str(ul)
            
        sections_data.append({
            'heading': heading,
            'content': content_html
        })

    if not sections_data:
        sections_data.append({
            'heading': 'Thông tin chi tiết',
            'content': str(soup)
        })

    out = ""
    bg_colors = ["bg-white", "bg-gray-50", "bg-white", "bg-gray-100"]
    for i, sec in enumerate(sections_data):
        bg = bg_colors[i % len(bg_colors)]
        
        img_html = ""
        if i < len(other_images):
            img_src = other_images[i]
            img_html = f"""
            <div class="w-full lg:w-1/2">
              <div class="rounded-2xl overflow-hidden shadow-lg aspect-[4/3] relative">
                <img src="{img_src}" class="absolute inset-0 w-full h-full object-cover">
              </div>
            </div>
            """
            
        flex_dir = "lg:flex-row-reverse" if i % 2 == 1 else "lg:flex-row"
        text_width = "lg:w-1/2" if img_html else "lg:w-3/4 lg:mx-auto"

        out += f"""
<!-- SECTION {i+1} -->
<section class="xl:py-12 md:py-8 py-4 {bg}">
  <div class="container mx-auto px-4">
    <div class="flex flex-col gap-6 {flex_dir} lg:items-center">
      <div class="w-full {text_width}">
        <h2 class="display-6 text-primary-600 mb-4 md:mb-6 xl:mb-8">{sec['heading']}</h2>
        <div class="prose max-w-none text-gray-700 body-2" style="list-style-type: disc;">
          {sec['content']}
        </div>
      </div>
      {img_html}
    </div>
  </div>
</section>
"""
    return out, hero_image, contact_email, contact_address, contact_phone

def main():
    print("Loading original HTML from XML...")
    file_path = "/home/x79/sisvietnamvn_01/sisvietnamvn_Trang chính thức hiện tại/post_chuyenkhoa.xml"
    tree = ET.parse(file_path)
    root = tree.getroot()
    channel = root.find("channel")
    ns = {"wp": "http://wordpress.org/export/1.2/", "content": "http://purl.org/rss/1.0/modules/content/"}

    slug_to_raw = {}
    for item in channel.findall("item"):
        post_name = item.find("wp:post_name", ns).text
        content_node = item.find("content:encoded", ns)
        if content_node is not None and content_node.text:
            slug_to_raw[post_name] = content_node.text

    print(f"Loaded {len(slug_to_raw)} raw posts from XML")

    dsn = "localhost:1521/sisvietnam"
    conn = oracledb.connect(user="sisvietnam", password="sisvietnam", dsn=dsn)
    cursor = conn.cursor()
    
    cursor.execute("SELECT slug, title FROM sis_page WHERE page_type = 'CUSTOM'")
    rows = cursor.fetchall()
    
    count = 0
    warnings = []
    
    for row in rows:
        slug = row[0]
        title = row[1]
        
        # If slug doesn't match exactly, maybe some slight difference? We'll try exact match
        if slug not in slug_to_raw:
            print(f"Slug {slug} not found in XML! Skipping.")
            continue
            
        raw_html = slug_to_raw[slug]
        
        # Wrap it in wp-content-container for parsing if needed
        if 'wp-content-container' not in raw_html:
             raw_html = f'<div class="wp-content-container">{raw_html}</div>'

        transformed_content, hero_image, email, address, phone = transform_html(raw_html)
        
        if "Thông tin chi tiết" in transformed_content and "wp-content-container" not in transformed_content:
            warnings.append(slug)
            
        # Update DB
        update_sql = """
            UPDATE sis_page 
            SET content = :1, hero_image = :2, contact_email = :3, contact_address = :4, contact_phone = :5 
            WHERE slug = :6
        """
        cursor.execute(update_sql, [transformed_content, hero_image, email, address, phone, slug])
        count += 1
        print(f"Restored and Updated {slug}")
        
    conn.commit()
    print(f"Successfully fixed {count} pages.")

if __name__ == "__main__":
    main()