"""Immonot.com — Notaires de France property listings. Picked over pap.fr (Cloudflare-blocked 2026-05-25) because immonot serves HTML cleanly to standard requests and represents a distinct market slice: notary-listed properties, often estate sales / family transfers / divorce splits — direct seller-to-buyer with the notary as intermediary. No agent commission. Frequently the cheapest comparable properties in a département. URL pattern (verified 2026-05-25): https://www.immonot.com/recherche-annonces-par-departement/VENT/MAIS/{dept_code}/Achat-Maison-{dept_slug}.html VENT = vente (sale, vs LOC = location/rent) MAIS = maison (vs APPT = appartement, TERR = terrain) Card structure: div.il-card-content per listing, with: - a.reset-link[href^="/annonce-..."] (detail URL) - h2.il-card-title > strong.il-card-locale (city + postcode) - div.il-card-price > strong (price text) - Card text contains: "Intérieur X m 2", "Extérieur Y m", "Pièces N", "Chb. M" Pagination not yet handled — first page returns ~12 results per département, which is enough for the weekly cadence and small enough to not annoy the site. """ from __future__ import annotations import re import sys import time from pathlib import Path from typing import Iterator import requests from bs4 import BeautifulSoup SCRIPT_DIR = Path(__file__).resolve().parent.parent if str(SCRIPT_DIR) not in sys.path: sys.path.insert(0, str(SCRIPT_DIR)) from sources._base import PropertyHit, SearchCriteria, Source # noqa: E402 # Region slug → (département INSEE code, département slug for the URL) REGION_MAP = { 'drome': ('26', 'drome'), 'ardeche': ('07', 'ardeche'), 'herault': ('34', 'herault'), 'gard': ('30', 'gard'), 'cotes-d-armor': ('22', 'cotes-d-armor'), 'morbihan': ('56', 'morbihan'), 'finistere': ('29', 'finistere'), 'dordogne': ('24', 'dordogne'), 'lot': ('46', 'lot'), 'creuse': ('23', 'creuse'), 'correze': ('19', 'correze'), 'charente': ('16', 'charente'), 'charente-maritime': ('17', 'charente-maritime'), # Added 2026-05-27 — Massif Central + Pyrenean + south-Languedoc widening 'cantal': ('15', 'cantal'), 'aveyron': ('12', 'aveyron'), 'lozere': ('48', 'lozere'), 'tarn': ('81', 'tarn'), 'tarn-et-garonne': ('82', 'tarn-et-garonne'), 'pyrenees-atlantiques': ('64', 'pyrenees-atlantiques'), 'hautes-pyrenees': ('65', 'hautes-pyrenees'), 'aude': ('11', 'aude'), 'pyrenees-orientales': ('66', 'pyrenees-orientales'), } BASE = 'https://www.immonot.com' SEARCH_URL_TEMPLATE = BASE + '/recherche-annonces-par-departement/VENT/MAIS/{code}/Achat-Maison-{slug}.html' UA = ('Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) ' 'AppleWebKit/605.1.15 (KHTML, like Gecko) ' 'Version/17.4 Safari/605.1.15') # Regexes for the card text body (city + price come from CSS selectors) RE_SURFACE = re.compile(r'Int[ée]rieur\s+([\d ,.]+)\s*m\s*[²2]') RE_LAND = re.compile(r'Ext[ée]rieur\s+([\d ,.]+)\s*m') RE_ROOMS = re.compile(r'Pi[èe]ces\s+(\d+)') RE_BEDS = re.compile(r'Chb\.\s*(\d+)') RE_PRICE_DIGITS = re.compile(r'(\d[\d\s]*)\s*€') RE_CITY_POSTCODE = re.compile(r'^(.+?)\s*-\s*(\d{5})\s*$') def _to_int(text: str) -> int | None: """Strip spaces/commas/dots, return int or None.""" if not text: return None digits = re.sub(r'[^\d]', '', text) return int(digits) if digits else None class ImmonotSource(Source): name = 'immonot' countries = ['FR'] requires_auth = False def __init__(self): self._session = requests.Session() self._session.headers.update({ 'User-Agent': UA, 'Accept': 'text/html,application/xhtml+xml', 'Accept-Language': 'fr-FR,fr;q=0.9', }) def health(self) -> tuple[bool, str]: """GET probe against the search base URL (HEAD returns 405 on immonot).""" try: r = self._session.get(BASE + '/', timeout=8, allow_redirects=True) if r.status_code >= 400: return False, f'HTTP {r.status_code}' return True, 'reachable' except requests.RequestException as e: return False, f'{type(e).__name__}' def search(self, criteria: SearchCriteria, known_urls: set[str] | None = None) -> Iterator[PropertyHit]: if criteria.country != 'FR': return # immonot is France-only if not criteria.region or criteria.region not in REGION_MAP: return # unknown département slug known = known_urls or set() code, slug = REGION_MAP[criteria.region] url = SEARCH_URL_TEMPLATE.format(code=code, slug=slug) try: r = self._session.get(url, timeout=20) except requests.RequestException as e: print(f" immonot: {type(e).__name__} on {criteria.region}") return if r.status_code != 200: print(f" immonot: HTTP {r.status_code} on {criteria.region}") return soup = BeautifulSoup(r.text, 'html.parser') cards = soup.select('div.il-card-content') yielded = 0 for card in cards: if criteria.limit and yielded >= criteria.limit: break link = card.select_one('a.reset-link[href^="/annonce-"]') if not link: continue href = link.get('href', '') prop_url = BASE + href if href.startswith('/') else href if prop_url in known: continue # skip already-seen — saves parsing + keeps limit semantics honest # City + postcode from .il-card-locale city = None country_code = 'FR' locale_el = card.select_one('strong.il-card-locale') if locale_el: locale_text = locale_el.get_text(strip=True) m = RE_CITY_POSTCODE.match(locale_text) if m: city = m.group(1).strip() # Price from .il-card-price price = None price_el = card.select_one('div.il-card-price') if price_el: price_text = price_el.get_text(' ', strip=True) m = RE_PRICE_DIGITS.search(price_text) if m: price = _to_int(m.group(1)) # Upstream filters if price is not None: if criteria.min_price and price < criteria.min_price: continue if criteria.max_price and price > criteria.max_price: continue # Surface, land, rooms, bedrooms from card text card_text = card.get_text(' ', strip=True) building_size = None land_size = None rooms = None bedrooms = None m = RE_SURFACE.search(card_text) if m: # "200,35" → 200 (truncate decimal); accept "200 35" too val = m.group(1).replace(',', '.').replace(' ', '') try: building_size = int(float(val)) except ValueError: pass m = RE_LAND.search(card_text) if m: val = m.group(1).replace(',', '.').replace(' ', '') try: land_size = int(float(val)) except ValueError: pass m = RE_ROOMS.search(card_text) if m: rooms = int(m.group(1)) m = RE_BEDS.search(card_text) if m: bedrooms = int(m.group(1)) # Apply size + bedroom upstream filters if criteria.min_building_m2 and (building_size or 0) < criteria.min_building_m2: continue if criteria.min_land_m2 and (land_size or 0) < criteria.min_land_m2: continue if criteria.min_bedrooms and (bedrooms or 0) < criteria.min_bedrooms: continue yield PropertyHit( url=prop_url, source='immonot', title=f"{city or criteria.region.title()} - immonot", price=price, city=city, country=country_code, building_size=building_size, land_size=land_size, rooms=rooms, bedrooms=bedrooms, search_region=criteria.region, ) yielded += 1 time.sleep(1) # be polite — single département per second