aaas-2025-presenters-export.py

import time
import logging
import pandas as pd
from selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.webdriver.common.keys import Keys
from selenium.webdriver.chrome.options import Options
from bs4 import BeautifulSoup

logging.basicConfig(
    level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s"
)
logger = logging.getLogger()

urls = [
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35359",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34815",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34816",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35145",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35143",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35139",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35152",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35358",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35280",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35136",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35002",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35001",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35141",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35387",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35511",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35370",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34211",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34402",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34445",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33902",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34464",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34160",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34411",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34340",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34392",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33981",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34278",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33947",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34323",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34320",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35153",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35133",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34187",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34389",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34024",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33992",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34141",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34368",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34013",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34451",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34277",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34076",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34468",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34516",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34495",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35315",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35441",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35091",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35093",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35560",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35501",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35643",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34814",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34419",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33822",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34379",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34287",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34016",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34467",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33970",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34388",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34456",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34378",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34303",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34817",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34466",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34565",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33892",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35078",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34819",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34327",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34194",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34317",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34422",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34390",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34315",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34202",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34407",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34294",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34497",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34450",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34434",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34432",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34263",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35270",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34138",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34100",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34469",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33934",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34236",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34354",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34252",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/33928",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34427",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34491",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34439",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35313",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34463",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34000",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34510",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34169",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35148",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35379",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34873",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/34870",
    "https://aaas.confex.com/aaas/2025/meetingapp.cgi/Session/35121",
]

results = []

# Set up the Selenium WebDriver w/ Chrome
options = Options()
options.headless = True
driver = webdriver.Chrome(options=options)

for url in urls:
    try:
        logger.info(f"Processing URL: {url}")

        driver.get(url)

        # Wait 10 sec for the page to load fully
        time.sleep(10) 

        html_content = driver.page_source

        logger.debug(f"HTML content for {url}: {html_content[:500]}...")

        soup = BeautifulSoup(html_content, "html.parser")

        session_title = (
            soup.select_one("p.favoriteItem").text.strip()
            if soup.select_one("p.favoriteItem")
            else "N/A"
        )
        session_date = (
            soup.select_one("p.SlotDate .defaultTZ").text.strip()
            if soup.select_one("p.SlotDate .defaultTZ")
            else "N/A"
        )
        session_time = (
            soup.select_one("p.SlotTime .defaultTZ").text.strip()
            if soup.select_one("p.SlotTime .defaultTZ")
            else "N/A"
        )

        people = soup.select(".PersonList a")

        if not people:
            logger.warning(f"No people found on session page: {url}")
            continue

        for person in people:
            person_name = person.text.strip()
            person_href = person.get("href", "")
            role_affiliation = person.find_next("span", class_="roleAffiliation")
            person_title = role_affiliation.find("ul") if role_affiliation else None
            person_title = person_title.text.strip() if person_title else "N/A"

            results.append(
                {
                    "person_name": person_name,
                    "person_title": person_title,
                    "person_href": person_href,
                    "session_title": session_title,
                    "session_date": session_date,
                    "session_time": session_time,
                }
            )

        logger.info(f"Successfully processed session at {url}")

    except Exception as e:
        logger.error(f"Error occurred while processing {url}: {e}")

df = pd.DataFrame(results)
df.to_excel("session_people_with_titles.xlsx", index=False)
logger.info("Data has been written to session_people_with_titles.xlsx")
driver.quit()