Networx / IRIS Recruitment Jobs API.

import re
import requests

def bootstrap(careers_url: str) -> dict:
    # The cid is the trailing -{digits} of the /Careers/{page}-{cid} path.
    client_id = int(re.search(r"-(\d+)$", careers_url.split("?")[0]).group(1))
    origin = "https://" + careers_url.split("/", 3)[2]

    session = requests.Session()
    session.headers["User-Agent"] = "Mozilla/5.0"

    # 1. Root GET seeds ASP.NET_SessionId + ApplicationGatewayAffinity cookies.
    session.get(origin + "/", timeout=15).raise_for_status()

    # 2. The scoped careers page carries the anti-forgery form token + onboarding id.
    html = session.get(careers_url, timeout=15).text
    form_token = re.search(
        r'name="__RequestVerificationToken"[^>]*value="([^"]+)"', html).group(1)
    onboarding = re.search(r"InitialiseVacancySearch\([^)]*?(-?\d+)\s*\)", html)

    return {
        "session": session,
        "origin": origin,
        "careers_url": careers_url,
        "client_id": client_id,
        "form_token": form_token,
        "onboarding_page_id": int(onboarding.group(1)) if onboarding else None,
    }

import json

def build_data(ctx: dict, page: int) -> str:
    data = {
        "ClientID": ctx["client_id"],
        "DynamicFields": [],
        "SearchResultFields": [],
        "CurrentPage": page,
        "PageSearchResults": True,
        "SearchResultPageSize": 25,
        "keywords": "",
        "Locations": ["0"],
        "Salary_All+Salaries": ["0"],
    }
    if ctx["onboarding_page_id"] is not None:
        data["OnboardingPageID"] = ctx["onboarding_page_id"]
    return json.dumps(data)

def post_search(ctx: dict, path: str, page: int) -> dict:
    body = {
        "__RequestVerificationToken": ctx["form_token"],
        "data": build_data(ctx, page),
        "hdnNewWorld": "True",
    }
    headers = {
        "Origin": ctx["origin"],
        "Referer": ctx["careers_url"],
        "X-Requested-With": "XMLHttpRequest",
        "__RequestVerificationToken": ctx["form_token"],
    }
    resp = ctx["session"].post(ctx["origin"] + path, data=body, headers=headers, timeout=15)
    if "DisplayError" in resp.url:
        raise RuntimeError("Anti-forgery token rejected — re-bootstrap the session")
    return resp.json()

ctx = bootstrap("https://hymans.current-vacancies.com/Careers/Hymans%20external%20VSP-2054")
total = post_search(ctx, "/Careers/SearchVacanciesCount", 1)["Count"]
print(f"{total} live vacancies")

import math

def scrape_listings(ctx: dict) -> list[dict]:
    total = post_search(ctx, "/Careers/SearchVacanciesCount", 1)["Count"]
    pages = math.ceil(total / 25)
    listings = []
    for page in range(1, pages + 1):
        payload = post_search(ctx, "/Careers/SearchVacancies", page)
        if not payload.get("OK"):
            raise RuntimeError(f"SearchVacancies returned OK=false on page {page}")
        rows = payload.get("Data") or []
        for item in rows:
            job_id = item.get("VacancyPublishInfoID")
            if not job_id:
                continue  # no publish id -> can't build a detail URL, skip
            postcode = item.get("Postcode")
            listings.append({
                "id": str(job_id),
                "title": (item.get("VacancyTitle") or "").strip(),
                "reference": item.get("Reference"),
                "location": item.get("Location"),
                "postcode": None if postcode in (None, "NO CODE") else postcode,
                "salary": item.get("Salary"),
                # Contract type lands under a per-tenant dynamic key; match by suffix.
                "contract_type": next(
                    (v for k, v in item.items() if k.endswith("_ContractType")), None),
                # ApplyLink may point at a sibling brand host sharing the same cid.
                "detail_url": item.get("ApplyLink")
                    or f"{ctx['origin']}/Jobs/Advert/{job_id}?cid={ctx['client_id']}",
                "posted": item.get("CreatedDate"),
                "closes": item.get("ExpiryDate"),
            })
        expected = 25 if page < pages else total - (page - 1) * 25
        if len(rows) != expected:
            raise RuntimeError(f"page {page}: got {len(rows)} rows, expected {expected}")
    return listings

import json
from bs4 import BeautifulSoup

def fetch_detail(session: requests.Session, detail_url: str) -> dict:
    html = session.get(detail_url, timeout=15).text
    soup = BeautifulSoup(html, "html.parser")

    posting = None
    for tag in soup.find_all("script", type="application/ld+json"):
        try:
            blob = json.loads(tag.string or "{}")
        except json.JSONDecodeError:
            continue
        for node in (blob if isinstance(blob, list) else [blob]):
            if isinstance(node, dict) and node.get("@type") == "JobPosting":
                posting = node
                break

    posting = posting or {}
    org = posting.get("hiringOrganization") or {}
    # The advert body backs the description when JSON-LD leaves it blank.
    body = soup.select_one("#top .col-lg-7") or soup.select_one("#top")

    return {
        "title": posting.get("title"),
        "employment_type": posting.get("employmentType"),
        "posted_at": posting.get("datePosted"),
        "closes_at": posting.get("validThrough"),
        "company": org.get("name"),
        "description_html": posting.get("description")
            or (body.decode_contents() if body else None),
    }