feat: parse project metadata from PF links

app/scrapers/base.py

@@ -35,6 +35,9 @@ class ScrapedListing:
     permit_number: str | None
     agent_name: str | None
     agency_name: str | None
+    building: str | None = None
+    bedrooms: int | None = None
+    size_sqft: float | None = None
     is_active: bool = True
 
 

app/scrapers/propertyfinder.py

@@ -100,6 +100,110 @@ def _extract_permit(item: dict) -> str | None:
     return None
 
 
+def _parse_int(value) -> int | None:
+    if value is None:
+        return None
+    if isinstance(value, bool):
+        return None
+    if isinstance(value, (int, float)):
+        return int(value)
+    text = str(value).strip().lower()
+    if text in {"studio", "студия"}:
+        return 0
+    m = re.search(r"\d+", text)
+    return int(m.group(0)) if m else None
+
+
+def _extract_bedrooms(item: dict) -> int | None:
+    for key in ("bedrooms", "bedroom", "beds", "rooms", "bedroom_count", "bedrooms_count"):
+        value = item.get(key)
+        if isinstance(value, dict):
+            value = value.get("value") or value.get("count") or value.get("name")
+        parsed = _parse_int(value)
+        if parsed is not None:
+            return parsed
+
+    for node in _walk(item):
+        if not isinstance(node, dict):
+            continue
+        name = str(node.get("name") or node.get("label") or node.get("key") or "").lower()
+        if "bed" not in name and "спал" not in name:
+            continue
+        parsed = _parse_int(node.get("value") or node.get("count") or node.get("text"))
+        if parsed is not None:
+            return parsed
+    return None
+
+
+def _area_to_sqft(value, unit: str | None = None) -> float | None:
+    parsed = parse_price(value)
+    if parsed is None:
+        return None
+    unit_text = (unit or "").lower()
+    if any(token in unit_text for token in ("sqm", "sq m", "m2", "m²", "метр")):
+        return round(parsed * 10.7639, 2)
+    return parsed
+
+
+def _extract_size_sqft(item: dict) -> float | None:
+    for key in ("size", "area", "property_size", "built_up_area", "builtup_area", "plot_area"):
+        value = item.get(key)
+        unit = None
+        if isinstance(value, dict):
+            unit = value.get("unit") or value.get("unit_label") or value.get("unitLabel")
+            value = value.get("value") or value.get("amount") or value.get("text")
+        parsed = _area_to_sqft(value, unit)
+        if parsed is not None:
+            return parsed
+
+    for node in _walk(item):
+        if not isinstance(node, dict):
+            continue
+        name = str(node.get("name") or node.get("label") or node.get("key") or "").lower()
+        if not any(token in name for token in ("size", "area", "sqft", "sq ft", "площад")):
+            continue
+        parsed = _area_to_sqft(
+            node.get("value") or node.get("amount") or node.get("text"),
+            str(node.get("unit") or node.get("unit_label") or ""),
+        )
+        if parsed is not None:
+            return parsed
+    return None
+
+
+def _location_candidate(node: dict) -> tuple[int, str] | None:
+    rank = _LOC_TYPE_PRIORITY.get(str(node.get("type", "")).upper(), -1)
+    name = str(node.get("name") or "").strip()
+    if rank < 0 or not name:
+        return None
+    return rank, name
+
+
+def _extract_building_from(node) -> str | None:
+    best_name: str | None = None
+    best_rank = -1
+    for item in _walk(node):
+        if not isinstance(item, dict):
+            continue
+        candidate = _location_candidate(item)
+        if not candidate:
+            continue
+        rank, name = candidate
+        if rank > best_rank:
+            best_rank, best_name = rank, name
+    return best_name
+
+
+def _extract_building(data: dict, item: dict) -> str | None:
+    for key in ("location", "location_tree", "locations", "locationTree", "community"):
+        value = item.get(key)
+        if value:
+            building = _extract_building_from(value)
+            if building:
+                return building
+    return _extract_building_from(data)
+
+
 def _find_permit_on_page(data: dict) -> str | None:
     """The DLD permit number lives in a regulatory block rendered as an image,
     but its plain value is still in __NEXT_DATA__: the dict that carries a
@@ -212,6 +316,9 @@ class PropertyFinderScraper:
             permit_number=_find_permit_on_page(data) or _extract_permit(best),
             agent_name=agent_name,
             agency_name=agency_name,
+            building=_extract_building(data, best),
+            bedrooms=_extract_bedrooms(best),
+            size_sqft=_extract_size_sqft(best),
             is_active=True,
         )
 
@@ -336,6 +443,9 @@ class PropertyFinderScraper:
                         permit_number=_extract_permit(node),
                         agent_name=agent_name,
                         agency_name=agency_name,
+                        building=_extract_building(data, node),
+                        bedrooms=_extract_bedrooms(node),
+                        size_sqft=_extract_size_sqft(node),
                         is_active=True,
                     )
                 )