eClass MCP Server

""" HTML parsing utilities for eClass MCP Server. Uses BeautifulSoup to extract data from eClass and CAS SSO HTML responses. """ import logging from typing import Dict, List, Optional, Tuple from bs4 import BeautifulSoup logger = logging.getLogger('eclass_mcp_server.html_parsing') def extract_sso_link(html_content: str, base_url: str) -> Optional[str]: """ Extract the SSO login link from the eClass login page. Returns: Absolute SSO login URL, or None if not found. """ soup = BeautifulSoup(html_content, 'html.parser') sso_link = None # Look for UoA login button for link in soup.find_all('a'): if 'Είσοδος με λογαριασμό ΕΚΠΑ' in link.text or 'ΕΚΠΑ' in link.text: sso_link = link.get('href') break # Fallback: look for CAS form action if not sso_link: for form in soup.find_all('form'): if form.get('action') and 'cas.php' in form.get('action'): sso_link = form.get('action') break if not sso_link: logger.warning("Could not find SSO login link on the login page") return None # Ensure URL is absolute if not sso_link.startswith(('http://', 'https://')): sso_link = f"{base_url}/{sso_link.lstrip('/')}" logger.debug(f"Extracted SSO link: {sso_link}") return sso_link def extract_cas_form_data( html_content: str, current_url: str, sso_base_url: str = "https://sso.uoa.gr" ) -> Tuple[Optional[str], Optional[str], Optional[str]]: """ Extract execution parameter and form action from CAS login page. Returns: Tuple of (execution_value, form_action, error_message). On parsing failure, appropriate values will be None. """ soup = BeautifulSoup(html_content, 'html.parser') # Check for error message error_msg = soup.find('div', {'id': 'msg'}) error_text = error_msg.text.strip() if error_msg else None # Find execution token execution_input = soup.find('input', {'name': 'execution'}) if not execution_input: logger.warning("Could not find execution parameter on SSO page") return None, None, error_text or "Could not find execution parameter on SSO page" execution = execution_input.get('value') # Find login form form = soup.find('form', {'id': 'fm1'}) or soup.find('form') if not form: logger.warning("Could not find login form on SSO page") return execution, None, error_text or "Could not find login form on SSO page" action = form.get('action') if not action: action = current_url elif not action.startswith(('http://', 'https://')): action = f"{sso_base_url}/{action.lstrip('/')}" return execution, action, error_text def verify_login_success(html_content: str) -> bool: """ Verify if login was successful by checking for portfolio page content. """ return ('Μαθήματα' in html_content or 'portfolio' in html_content.lower() or 'course' in html_content.lower()) def extract_courses(html_content: str, base_url: str) -> List[Dict[str, str]]: """ Extract course information from the portfolio page. Returns: List of dicts with 'name' and 'url' keys. """ courses = [] soup = BeautifulSoup(html_content, 'html.parser') # Try specific course selectors course_elements = ( soup.select('.course-title') or soup.select('.lesson-title') or soup.select('.course-box .title') or soup.select('.course-info h4') ) if course_elements: for course_elem in course_elements: course_link = course_elem.find('a') or course_elem if course_link and course_link.get('href'): course_name = course_link.text.strip() course_url = course_link.get('href') if course_name: courses.append({ 'name': course_name, 'url': _make_absolute_url(course_url, base_url) }) else: # Fallback: find course links by URL pattern for link in soup.find_all('a'): href = link.get('href', '') if 'courses' in href or 'course.php' in href: course_name = link.text.strip() if course_name: courses.append({ 'name': course_name, 'url': _make_absolute_url(href, base_url) }) logger.debug(f"Extracted {len(courses)} courses") return courses def _make_absolute_url(url: str, base_url: str) -> str: """Convert relative URL to absolute URL.""" if url.startswith(('http://', 'https://')): return url return f"{base_url}/{url.lstrip('/')}" def format_course_list(courses: List[Dict[str, str]]) -> str: """ Format course list for display. Returns: Formatted string with numbered courses and URLs. """ lines = [] for i, course in enumerate(courses, 1): lines.append(f"{i}. {course['name']}") lines.append(f" URL: {course['url']}") return "\n".join(lines)