PeopleBank Jobs API.

import requests

def fetch_board(board_url: str, session: requests.Session) -> str:
    # do_search renders the full public advert set on one page (no pagination).
    resp = session.get(board_url, params={"do_search": "Show Vacancies"}, timeout=90)
    if resp.status_code == 403:
        # 403 is the WAF challenge, not an empty board -> retry with backoff.
        raise RuntimeError("PeopleBank WAF challenge on board — back off and retry")
    resp.raise_for_status()
    return resp.text

session = requests.Session()
session.headers.update({
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64)",
    "Accept": "text/html,application/xhtml+xml",
})
html = fetch_board(
    "https://careers.krispykreme.co.uk/pb3/corporate/krispykreme/advertsearch.php",
    session,
)

import re
from urllib.parse import urljoin, urlparse, parse_qs
from bs4 import BeautifulSoup

PUBLIC_KEY = re.compile(r"^[A-Za-z0-9@_-]+$")

def parse_listings(html: str, board_url: str) -> list[dict]:
    soup = BeautifulSoup(html, "html.parser")
    anchors = soup.select("a[href*='pbk24w12.main'][href*='p=']")
    if not anchors:
        # Distinguish an explicitly empty board from a blocked/changed page.
        text = soup.get_text(" ", strip=True).lower()
        if re.search(r"no vacancies|no positions|0 (?:jobs|positions)", text):
            return []
        raise RuntimeError("No advert links found — page was blocked or changed")

    listings, seen = [], set()
    for a in anchors:
        detail_url = urljoin(board_url, a.get("href", ""))
        key = (parse_qs(urlparse(detail_url).query).get("p") or [""])[0]
        if not PUBLIC_KEY.match(key) or key in seen:
            continue
        seen.add(key)
        row = a.find_parent(class_="result") or a.find_parent(["tr", "li"]) or a.parent
        title_node = row.select_one(".result-headline, .job-title, h2, h3") if row else None
        loc_node = row.select_one(".location, .job-location, .res-location") if row else None
        close_text = row.get_text(" ", strip=True) if row else ""
        close = re.search(r"(\d{1,2}/\d{1,2}/\d{4})", close_text)
        listings.append({
            "external_id": key,  # opaque public posting key, stable across routes
            "title": (title_node or a).get_text(strip=True),
            "detail_url": detail_url,
            "location": (row.get("data-location") if row else None)
                or (loc_node.get_text(strip=True) if loc_node else None),
            "postcode": loc_node.get("data-postcode") if loc_node else None,
            "closes_at": close.group(1) if close else None,  # dd/mm/yyyy (en-GB)
        })
    return listings

INTERNAL_ID = re.compile(r"(?:[?&]|&)P_nAdId=(\d+)", re.IGNORECASE)

def fetch_detail(session: requests.Session, listing: dict) -> dict:
    resp = session.get(listing["detail_url"], timeout=90)
    if resp.status_code in (404, 410):
        return {**listing, "removed": True}      # advert closed, not a block
    if resp.status_code == 403:
        raise RuntimeError("PeopleBank WAF challenge on detail — back off and retry")
    resp.raise_for_status()
    soup = BeautifulSoup(resp.text, "html.parser")

    # Newer .job-* template vs legacy De Vere .ad-display wrapper.
    branded = soup.select_one(".ad-display > .ad-content")
    title_node = soup.select_one("h1.job-title, h1")
    if title_node:
        title = title_node.get_text(strip=True)
    elif branded and branded.select_one("h2"):
        title = branded.select_one("h2").get_text(strip=True)
    else:
        title = listing.get("title", "")

    body = soup.select_one(".job-content")
    if body:
        description = body.get_text("\n", strip=True)
    elif branded:
        parts = [d.get_text(" ", strip=True) for d in branded.find_all("div", recursive=False)]
        description = "\n\n".join(p for p in parts if p)
    else:
        description = ""
    if not title or not description:
        raise RuntimeError("PeopleBank detail omitted title or job-content")

    apply_a = soup.select_one("a[href*='candidate.apply'][href*='p=']")
    apply_url = (urljoin(listing["detail_url"], apply_a["href"])
                 if apply_a and apply_a.has_attr("href") else None)
    internal = INTERNAL_ID.search(resp.text)
    salary_node = soup.select_one(".job-sup")

    return {
        **listing,
        "title": title,
        "description": description,
        "apply_url": apply_url,
        "native_advert_id": internal.group(1) if internal else None,
        "contract_salary_text": salary_node.get_text(" ", strip=True) if salary_node else None,
    }

import time

def scrape(board_url: str) -> list[dict]:
    session = requests.Session()
    session.headers.update({"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64)"})

    listings = parse_listings(fetch_board(board_url, session), board_url)
    jobs = []
    for listing in listings:
        detail = fetch_detail(session, listing)   # single-threaded on purpose
        if detail.get("removed"):
            continue                              # skip closed adverts
        jobs.append(detail)
        time.sleep(2)                             # ~2s between requests
    return jobs

for job in scrape("https://www.peoplebank.com/pb3/corporate/careconcern/advertsearch.php"):
    print(job["external_id"], job["title"])