import xml.etree.ElementTree as ET
import csv
import sys
from bs4 import BeautifulSoup

try:
    tree = ET.parse('/home/x79/sisvietnamvn_01/sisvietnamvn_Trang chính thức hiện tại/bnhvinakhoaquctsiscnth.WordPress.2026-07-28 (5).xml')
    root = tree.getroot()
except Exception as e:
    print(f"Error parsing XML: {e}")
    sys.exit(1)

ns = {
    'wp': 'http://wordpress.org/export/1.2/',
    'content': 'http://purl.org/rss/1.0/modules/content/'
}

attachments = {}
for channel in root.findall('channel'):
    for item in channel.findall('item'):
        pt = item.find('wp:post_type', ns)
        if pt is not None and pt.text == 'attachment':
            post_id = item.find('wp:post_id', ns)
            url = item.find('wp:attachment_url', ns)
            if post_id is not None and url is not None:
                attachments[post_id.text] = url.text

csv_mapping = {}
with open('/home/x79/sisvietnamvn_01/Media/HÌNH ẢNH BÁC SĨ NGỒI PHÒNG KHÁM/ALL_IMAGES/DANH SÁCH BÁC SĨ NGỒI PHÒNG KHÁM.csv', 'r', encoding='utf-8-sig') as f:
    reader = csv.DictReader(f)
    for row in reader:
        name = row.get('Họ và tên', '').strip()
        msnv = row.get('MSNV', '').strip()
        if name:
            csv_mapping[name.lower()] = msnv

def escape_sql(text):
    if not text:
        return "NULL"
    return "'" + text.replace("'", "''") + "'"

def parse_html_content(html):
    if not html:
        return None, None, None
    soup = BeautifulSoup(html, 'html.parser')
    
    education = ""
    experience = ""
    achievements = ""
    
    current_section = None
    
    # We will just split it simply for now, but a more robust way:
    # Just save the entire raw HTML into work_experience if we can't easily parse
    return escape_sql(html), "NULL", "NULL"

sql_statements = []
# TRUNCATE TABLE sis_doctor CASCADE? We should probably just insert.
# But there might be old data. Let's just generate INSERTs.

doctor_id = 1000

for channel in root.findall('channel'):
    for item in channel.findall('item'):
        pt = item.find('wp:post_type', ns)
        if pt is not None and pt.text == 'employee':
            title_el = item.find('title')
            raw_title = title_el.text.strip() if title_el is not None and title_el.text else ""
            
            # Extract name and title (e.g. "TS.BS TRẦN CHÍ CƯỜNG" -> title="TS.BS", name="TRẦN CHÍ CƯỜNG")
            parts = raw_title.split(' ', 1)
            doctor_title = parts[0] if '.' in parts[0] else ''
            doctor_name = parts[1] if doctor_title and len(parts) > 1 else raw_title
            if not doctor_title:
                doctor_name = raw_title
            
            doctor_name_lower = doctor_name.lower()
            
            msnv = ""
            for csv_name, code in csv_mapping.items():
                if csv_name in doctor_name_lower or doctor_name_lower in csv_name:
                    msnv = code
                    break
            
            meta_dict = {}
            for meta in item.findall('wp:postmeta', ns):
                k = meta.find('wp:meta_key', ns)
                v = meta.find('wp:meta_value', ns)
                if k is not None and k.text and v is not None:
                    meta_dict[k.text] = v.text or ""
            
            position = meta_dict.get('position', '')
            gender = meta_dict.get('gender', '')
            birthday = meta_dict.get('birthday', '')
            phone = meta_dict.get('phone_number', '')
            email = meta_dict.get('email', '')
            address = meta_dict.get('address', '')
            order = meta_dict.get('order', '0')
            image_id = meta_dict.get('image', '')
            avatar_url = attachments.get(image_id, '')
            
            content_el = item.find('content:encoded', ns)
            html_content = content_el.text if content_el is not None else ""
            exp, edu, ach = parse_html_content(html_content)
            
            if not order.isdigit():
                order = '0'
                
            # INSERT statement
            # Use sequenceGenerator.nextval or just let JPA handle it? Since we insert via SQL, we use sequence_generator.nextval
            sql = f"""
INSERT INTO sis_doctor (
    id, name, title, avatar_url, doctor_code, active, 
    position, gender, birthday, phone_number, email, address, display_order, 
    work_experience, created_by, created_date, last_modified_by, last_modified_date
) VALUES (
    {doctor_id}, {escape_sql(doctor_name)}, {escape_sql(doctor_title)}, {escape_sql(avatar_url)}, {escape_sql(msnv)}, 1,
    {escape_sql(position)}, {escape_sql(gender)}, {escape_sql(birthday)}, {escape_sql(phone)}, {escape_sql(email)}, {escape_sql(address)}, {order},
    {exp}, 'system', sysdate, 'system', sysdate
);"""
            sql_statements.append(sql.strip())
            doctor_id += 1

with open('/home/x79/sisvietnamvn_01/sisvietnamvn_main/import_doctors.sql', 'w', encoding='utf-8') as f:
    f.write("DELETE FROM sis_doctor;\n")
    f.write("\n".join(sql_statements))
    f.write("\nCOMMIT;\n")

print("Generated import_doctors.sql")