PageUp Jobs API.

import requests

HOST = "careers.pageuppeople.com"
PAGE_ITEMS = 100

# PageUp fronts every payload as server-rendered HTML, but returns the same
# content as JSON when the request carries these XHR headers. The JSON wrapper
# also gives a deterministic total 'count' that drives pagination.
XHR_HEADERS = {
    "X-Requested-With": "XMLHttpRequest",
    "Accept": "application/json, text/javascript, */*; q=0.01",
}

# Each tenant picks exactly ONE (cwci, locale) path shape. Probe page 1 of
# each in turn until one responds with count > 0.
PORTAL_PROBES = [
    ("cw", "en-us"), ("ci", "en"), ("cw", "en-au"), ("ci", "en-au"),
    ("cw", "en-gb"), ("ci", "en-us"), ("cw", "en"), ("ci", "en-gb"),
]

def listing_url(tenant, cwci, locale, page):
    return (f"https://{HOST}/{tenant}/{cwci}/{locale}/listing/"
            f"?page={page}&page-items={PAGE_ITEMS}")

def discover_portal(session, tenant):
    for cwci, locale in PORTAL_PROBES:
        resp = session.get(listing_url(tenant, cwci, locale, 1),
                           headers=XHR_HEADERS, timeout=30)
        if resp.status_code in (403, 404):
            continue
        resp.raise_for_status()
        data = resp.json()
        if data.get("count", 0) > 0:
            return cwci, locale, data
    raise RuntimeError(f"No PageUp portal responded for tenant {tenant}")

session = requests.Session()
tenant = "889"  # numeric tenant id — the leading path segment of the careers URL
cwci, locale, first_page = discover_portal(session, tenant)
print(f"Portal: /{tenant}/{cwci}/{locale}  total jobs: {first_page['count']}")

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

def job_id_from_url(url):
    # Detail path shape: /{tenant}/{cwci}/{locale}/job/{jobId}/{slug}
    path = urlparse(url).path
    if "/job/" not in path:
        return None
    tail = path.split("/job/", 1)[1].strip("/").split("/")
    return tail[0] if tail and tail[0].isdigit() else None

def clean(el):
    return " ".join(el.get_text().split()) if el else None

def parse_rows(results_html, seen):
    # 'results' is a sequence of BARE <tr> rows. Wrap them in <table><tbody>
    # or the parser drops the <td> siblings (location/department/close-date).
    soup = BeautifulSoup(f"<table><tbody>{results_html}</tbody></table>", "html.parser")
    rows = []
    for anchor in soup.select("a.job-link"):
        href = anchor.get("href")
        if not href:
            continue
        url = urljoin(f"https://{HOST}/", href)
        job_id = job_id_from_url(url)
        if not job_id or job_id in seen:
            continue
        seen.add(job_id)
        tr = anchor.find_parent("tr")
        rows.append({
            "job_id": job_id,
            "title": clean(anchor),
            "listing_url": url,
            "department": clean(tr.select_one(".job-department, .department")) if tr else None,
            "location": clean(tr.select_one(".location")) if tr else None,
        })
    return rows

def fetch_all_listings(session, tenant):
    cwci, locale, data = discover_portal(session, tenant)
    total = data.get("count", 0)
    seen = set()
    listings = parse_rows(data.get("results", ""), seen)

    page = 1
    while len(listings) < total:
        page += 1
        resp = session.get(listing_url(tenant, cwci, locale, page),
                           headers=XHR_HEADERS, timeout=30)
        resp.raise_for_status()
        results = resp.json().get("results")
        if not results:
            break
        before = len(listings)
        listings.extend(parse_rows(results, seen))
        if len(listings) == before:
            break  # no new rows — stop to avoid an infinite loop
    return listings

listings = fetch_all_listings(session, "889")
print(f"Collected {len(listings)} postings")

import time

def fetch_detail(session, listing):
    resp = session.get(listing["listing_url"], headers=XHR_HEADERS, timeout=30)
    if resp.status_code in (404, 410):
        return None  # posting removed — treat as a delisting signal
    resp.raise_for_status()
    fragment = resp.json().get("results")
    if not fragment:
        return None

    soup = BeautifulSoup(fragment, "html.parser")
    body = soup.select_one("#job-details")
    apply_anchor = soup.select_one("a.apply-link")
    apply_url = listing["listing_url"]
    if apply_anchor and apply_anchor.get("href"):
        apply_url = urljoin(listing["listing_url"], apply_anchor["href"])

    return {
        **listing,
        "title": clean(soup.select_one("h2")) or listing["title"],
        "external_job_no": clean(soup.select_one(".job-externalJobNo")),
        "employment_type": clean(soup.select_one(".work-type")),
        "category": clean(soup.select_one(".categories")),
        "locations": [clean(el) for el in soup.select(".location")],
        "apply_url": apply_url,
        "description_html": body.decode_contents() if body else None,
    }

for listing in listings[:3]:
    job = fetch_detail(session, listing)
    if job:
        print(job["title"], "-", job["employment_type"])
    time.sleep(0.5)  # respect ~500ms between requests