.*?

]*>(.*?)

', re.S) # sections « toggler » de la fiche :

Titre

# suivi d'items

libellé[valeurs]

_SECTION_ITEM_RE = re.compile( r"

\s*(.*?)(?:\s*(.*?))?", re.S) # sections « à proximité » -> details ; le reste (Inclusions, Caractéristiques # intérieures/extérieures, Détail, Animaux, Sécurité, Buanderie, Autres # services, Critères importants…) -> amenities _PROXIMITY_TITLES = {"À proximité", "Transport"} _PETS_NO = {"Aucun animal autorisé", "Aucun animal accepté", "Animaux refusés"} _PETS_YES = {"Animaux acceptés", "Animaux autorisés", "Tous animaux autorisés"} def _clean_text(fragment: str) -> str: """HTML -> texte propre (br/p/li = sauts de ligne, entités décodées).""" import html as _h txt = re.sub(r"", "\n", fragment) txt = re.sub(r"", "\n", txt) txt = re.sub(r"]*>", "- ", txt) txt = re.sub(r"<[^>]+>", " ", txt) txt = _h.unescape(txt) txt = re.sub(r"[ \t]+", " ", txt) txt = re.sub(r" ?\n ?", "\n", txt) return re.sub(r"\n{3,}", "\n\n", txt).strip() def _toggler_sections(html: str) -> list[tuple[str, list[tuple[str, str]]]]: """[(titre de section, [(libellé, valeur)])] depuis les blocs toggler.""" import html as _h out = [] for seg in re.split(r"gr-toggler-trigger>", html)[1:]: title = _h.unescape(seg.split("<", 1)[0]).strip() items = [] for lab, val in _SECTION_ITEM_RE.findall(seg): lab = _h.unescape(re.sub(r"<[^>]+>", " ", lab)).strip() val = _h.unescape(re.sub(r"<[^>]+>", " ", val or "")).strip() if lab and len(lab) <= 60: items.append((lab, val)) if title and items: out.append((title, items)) return out def _parse_fiche(html: str) -> dict: """Champs riches d'une fiche Kangalou (SSR).""" out: dict = {} m = _GPS_RE.search(html) if m: out["lat"], out["lng"] = float(m.group(1)), float(m.group(2)) m = _ADDR_RE.search(html) if m: import html as _h # certaines fiches double-encodent (« Carrières ») : deux passes out["address"] = _h.unescape(_h.unescape(m.group(1))).strip() flat = du.flatten(html) vals = {k: rx.search(flat) for k, rx in _FLAT_FIELD.items()} if vals["loyer"]: label = re.sub(r"\s+", " ", vals["loyer"].group(1)).strip() out["price_label"] = label try: out["price"] = float(re.sub(r"[^\d,.]", "", label.split("$")[0]) .replace(",", "")) except ValueError: pass if not (100 <= (out.get("price") or 0) <= 20000): out.pop("price", None) if vals["chambres"]: out["bedrooms"] = float(vals["chambres"].group(1)) if vals["sdb"]: out["bathrooms"] = float(vals["sdb"].group(1).replace(",", ".")) if vals["dispo"]: out["availability"] = vals["dispo"].group(1).strip(" .") if vals["etage"]: out.setdefault("details", {})["floor"] = int(vals["etage"].group(1)) if vals["superficie"]: try: m2 = float(vals["superficie"].group(1) .replace(" ", "").replace(" ", "").replace(",", ".")) if 8 <= m2 <= 2000: out["area_sqft"] = round(m2 * 10.7639) except ValueError: pass m = _DESC_FULL_RE.search(html) or _DESC_RE.search(html) if m: desc = _clean_text(m.group(1)) if desc: out["description"] = desc[:6000] # sections « toggler » : commodités structurées + proximité + animaux. # Libellés jamais interprétés (texte source) ; « libellé : valeurs » # quand la section fournit des valeurs (ex. Type de plancher : Céramique). amen: list[str] = [] proximity: dict[str, str] = {} pets_labels: list[str] = [] for title, items in _toggler_sections(html): if title in _PROXIMITY_TITLES: for lab, val in items: proximity[lab] = val continue for lab, val in items: entry = f"{lab} : {val}" if val else lab if entry not in amen: amen.append(entry) if title == "Animaux": pets_labels += [lab for lab, _ in items] if amen: out["amenities"] = amen if proximity: out.setdefault("details", {})["proximity"] = proximity # animaux / meublé : depuis les libellés structurés seulement (le texte # libre est laissé à la normalisation centrale) if pets_labels: if any(l in _PETS_NO for l in pets_labels): out["pets"] = "non" elif any(l in _PETS_YES for l in pets_labels): out["pets"] = "oui" else: # chat seulement, petit chien, en cage… out["pets"] = "conditions" if "Meublé" in amen: out["furnished"] = True # galerie (slider--gallery ; les vignettes asnav dupliquées sont exclues) ig = html.find("slider--gallery") if ig >= 0: seg = html[ig:] cut = seg.find("slider__asnav") if cut > 0: seg = seg[:cut] photos = [] for u in _GALLERY_RE.findall(seg): if u not in photos: photos.append(u) if photos: out["images"] = photos vt = du.virtual_tour(html) if vt: out.setdefault("details", {})["virtual_tour"] = vt return out class KangalouConnector(BaseConnector): source_id = "kangalou" request_delay = 0.5 timeout = 90 # la réponse pins fait ~6 Mo def _pins(self) -> list[dict]: import time import requests # le POST n'a pas l'escalade résiliente de get() : retenter les # lenteurs transitoires du site (ReadTimeout constaté 2026-08-24) for attempt in range(3): try: resp = self.post(PINS_API, data={"pinsOnly": "true"}, headers={"Accept": "application/json", "X-Requested-With": "XMLHttpRequest"}) break except (requests.ConnectionError, requests.Timeout): if attempt == 2: raise time.sleep(15 * (attempt + 1)) return (resp.json() or {}).get("pins") or [] @staticmethod def _unit_type(pin: dict) -> str: an = (pin.get("an") or "").strip() # « 4 ½ », « 9 ½ + » if an: ut = normalize_unit_type(an) if re.match(r"^\d½$|^6½\+$", ut or ""): return ut return _TYPES.get(pin.get("ti"), "") def fetch(self) -> list[Listing]: pins = self._pins() if MAX_PINS: pins = pins[:MAX_PINS] # key=v2 : parseur enrichi 2026-08-18 (galerie complète déjà en v1 ; # ajout loyer/pets/meublé/proximité + description multi-blocs) — force # une re-visite du parc complet une seule fois cache = du.TtlDetailCache(self, budget=DETAIL_LIMIT, ttl_days=TTL_DAYS, key="v2") out: list[Listing] = [] try: for pin in pins: if pin.get("ir"): continue # déjà loué : pas une offre active if pin.get("ti") in _EXCLUDED_TI or not pin.get("id"): continue # commercial / garage / malformé url = pin.get("l") or "" if not url: continue eid = str(pin["id"]) price = None try: price = float(str(pin.get("mc")).replace(",", "")) except (TypeError, ValueError): pass if price is not None and not (100 <= price <= 20000): price = None # affichage « sur demande » etc. lat = lng = None try: lat, lng = float(pin["lt"]), float(pin["lg"]) except (KeyError, TypeError, ValueError): pass photo = pin.get("p") or "" lst = Listing( source=self.source_id, external_id=eid, url=url, title=(pin.get("n") or "").strip(), city=(pin.get("c") or "").strip(), unit_type=self._unit_type(pin), price=price, price_label=(f"{price:,.0f} $/mois".replace(",", " ") if price else ""), images=[photo] if photo else [], lat=lat, lng=lng, ) payload = cache.get(eid, url, _parse_fiche) if payload: du.apply_detail(lst, payload) if payload.get("address"): lst.title = payload["address"] out.append(lst) finally: cache.close() return out