WeRecruit Jobs API.

import requests

locale, tenant = "fr", "walter-learning"
board_url = f"https://careers.werecruit.io/{locale}/{tenant}"

resp = requests.get(board_url, timeout=15)
if resp.status_code == 404:
    raise LookupError(f"tenant '{tenant}' not found")
resp.raise_for_status()
html = resp.text

import re

_START = re.compile(r"window\.allOffers\s*=\s*")

def extract_all_offers_json(html: str) -> str | None:
    m = _START.search(html)
    if not m or m.end() >= len(html) or html[m.end()] != "[":
        return None
    start = m.end()
    depth = 0
    in_string = escape = False
    for i in range(start, len(html)):
        c = html[i]
        if escape:
            escape = False
        elif in_string:
            if c == "\\":
                escape = True
            elif c == '"':
                in_string = False
        elif c == '"':
            in_string = True
        elif c == "[":
            depth += 1
        elif c == "]":
            depth -= 1
            if depth == 0:
                return html[start:i + 1]
    return None

import json

offers = json.loads(extract_all_offers_json(html))
jobs = []
for o in offers:
    # Require all three: a malformed offer must not silently drop half the alias pair.
    if not (o.get("Url") and o.get("Slug") and o.get("Id")):
        continue
    title = o.get("TitleTranslated") or next(iter((o.get("Title") or {}).values()), None)
    company = o.get("Subsidiary_Name") or o.get("Company_Name") or o.get("Holding_Name")
    jobs.append({
        "external_id": o["Slug"].strip(),        # stable key
        "native_id": o["Id"],                      # UUID (candidate only)
        "url": o["Url"].strip(),
        "title": title,
        "company": company,
        "employment_type": o.get("TypeTranslated"),
        "schedule": o.get("TimeTranslated"),
        "department": o.get("Address_Department"),
        "city": o.get("Address_City"),
        "region": o.get("Address_Region"),
        "postal_code": o.get("Address_PostalCode"),
        "country": o.get("Address_State"),         # ISO country code
        "published_at": o.get("PublicationStartDate"),
        "closes_at": o.get("EndDate"),
    })
print(f"parsed {len(jobs)} offers")

import html as ihtml
from bs4 import BeautifulSoup

def fetch_details(job_url: str) -> dict:
    resp = requests.get(job_url, timeout=15)
    resp.raise_for_status()
    soup = BeautifulSoup(resp.text, "html.parser")

    for tag in soup.find_all("script", type="application/ld+json"):
        try:
            data = json.loads(tag.string or "")
        except (ValueError, TypeError):
            continue
        for node in (data if isinstance(data, list) else [data]):
            if isinstance(node, dict) and node.get("@type") == "JobPosting":
                desc = BeautifulSoup(node.get("description", ""), "html.parser").get_text(" ", strip=True)
                return {
                    "title": ihtml.unescape(node.get("title", "")),
                    "description": desc,
                    "posted_at": node.get("datePosted"),
                    "employment_type": node.get("employmentType"),
                }

    # Fallback: some pages briefly render the offer before the JSON-LD is present.
    article = soup.select_one("article.single-offre")
    if article:
        h1 = article.select_one("h1")
        body = article.select_one(".description.rich-text")
        return {
            "title": h1.get_text(strip=True) if h1 else None,
            "description": body.get_text(" ", strip=True) if body else None,
        }
    return {}

import time

seen = {}   # slug -> native UUID
for job in jobs:
    prior = seen.get(job["external_id"])
    if prior is not None:
        if prior != job["native_id"]:
            print(f"identity collision on slug {job['external_id']} — review before merging")
        continue
    seen[job["external_id"]] = job["native_id"]

for job in [j for j in jobs if seen.get(j["external_id"]) == j["native_id"]]:
    job.update(fetch_details(job["url"]))
    time.sleep(0.4)   # mirror the scraper's ~400ms crawl delay