SilkRoad OpenHire Jobs API.

from urllib.parse import urlparse, parse_qs

DOMAIN = "silkroad.com"
# The modern SilkRoad product lives on these hosts and is a different contract.
EXCLUDED = {"jobs.silkroad.com", "jobs-ca.silkroad.com"}

def parse_openhire(url: str) -> dict:
    parsed = urlparse(url)
    host = parsed.netloc.lower()
    if host in EXCLUDED or not host.endswith("." + DOMAIN):
        raise ValueError("not a legacy OpenHire host")
    if "/epostings" not in parsed.path.lower():
        raise ValueError("not an /epostings route")

    # ColdFusion query keys are case-insensitive and the wild data mixes cases.
    query = {k.lower(): v[0] for k, v in parse_qs(parsed.query).items()}
    version = (query.get("version") or "1").lstrip("0") or "0"
    if not version.isdigit():
        raise ValueError("version must be numeric")

    return {
        "tenant": host.split(".")[0],
        "version": version,
        "job_id": query.get("jobid"),
        "fuseaction": (query.get("fuseaction") or "").lower(),
    }

print(parse_openhire("https://kcsouthern.silkroad.com/epostings/index.cfm"
                     "?fuseaction=app.jobinfo&jobid=1234&version=2"))

import requests
from bs4 import BeautifulSoup

def board_url(tenant: str, version: str) -> str:
    return (f"https://{tenant}.silkroad.com/epostings/index.cfm"
            f"?fuseaction=app.jobsearch&version={version}")

def job_url(tenant: str, version: str, job_id: str) -> str:
    return (f"https://{tenant}.silkroad.com/epostings/index.cfm"
            f"?fuseaction=app.jobinfo&jobid={job_id}&version={version}")

session = requests.Session()
session.headers["Accept"] = "text/html,application/xhtml+xml"

resp = session.get(board_url("southland", "1"), timeout=60)
resp.raise_for_status()
soup = BeautifulSoup(resp.text, "html.parser")
print(f"{len(resp.text)} bytes in one response")

def column_index(soup) -> dict:
    """Header anchor ids stay English (header_location) even when the
    visible header text is translated, e.g. on the Spanish CPKC board."""
    index = {}
    for position, header in enumerate(soup.select("a[id^='header_']")):
        index[header.get("id").lower()] = position
    return index

def parse_rows(soup) -> list[dict]:
    columns = column_index(soup)
    anchors = soup.select("a[id^='jobTitle_']")
    if not anchors and soup.select_one("a#header_jobTitle") is None:
        raise RuntimeError("this is not an OpenHire search page")

    rows = []
    for anchor in anchors:
        job_id = anchor.get("id")[len("jobTitle_"):]
        if not job_id.isdigit():
            continue
        cells = anchor.find_parent("tr").select("td") if anchor.find_parent("tr") else []

        def cell(name):
            position = columns.get(name)
            if position is None or position >= len(cells):
                return None
            return " ".join(cells[position].get_text().split())

        rows.append({
            "job_id": job_id,
            "title": " ".join(anchor.get_text().split()),
            "tracking_code": cell("header_trackingcode"),
            "location": cell("header_location"),
        })
    return rows

listings = parse_rows(soup)
print(f"{len(listings)} postings")

import time

DETAIL_IDS = {
    "title": "#jobTitleDiv",
    "description": "#jobDesciptionDiv",   # vendor's own typo — do not "fix" it
    "required_skills": "#jobRequiredSkillsDiv",
    "required_experience": "#jobExperienceRqdDiv",
    "location": "#jobPositionLocationDiv",
    "tracking_code": "#jobCodeDiv",
    "position_type": "#translatedJobPostingTypeDiv",
}

def fetch_detail(tenant: str, version: str, job_id: str) -> dict | None:
    url = job_url(tenant, version, job_id)
    resp = session.get(url, timeout=45)
    if resp.status_code in (404, 410):
        return None
    resp.raise_for_status()
    page = BeautifulSoup(resp.text, "html.parser")

    # A pulled posting answers HTTP 200, so status code alone proves nothing.
    if page.select_one("#jobStatusChangeDiv") and page.select_one("#jobTitleDiv") is None:
        return None

    if page.select_one("#jobTitleDiv") is None:
        raise RuntimeError("neither a posting nor the status-changed marker — parse error")

    detail = {"job_id": job_id, "listing_url": url}
    for field, selector in DETAIL_IDS.items():
        node = page.select_one(selector)
        detail[field] = node.decode_contents() if field == "description" and node else (
            " ".join(node.get_text().split()) if node else None
        )
    return detail

for row in listings[:3]:
    print(fetch_detail("southland", "1", row["job_id"]))
    time.sleep(0.6)