backend/app/api/v1/parcels.py

@@ -1330,7 +1330,13 @@ def analyze_parcel(
             }
         )
 
-    # 5) Конкуренты в радиусе 3 км из DOM.РФ.
+    # 5) Конкуренты в радиусе 3 км из DOM.РФ с ценами из objective_lots.
+    # OBJ-3: обогащаем каждый ЖК данными objective_lots через маппинг
+    # domrf_kn_objects.obj_id → objective_complex_mapping.domrf_obj_id
+    #   → objective_lots.project_name.
+    # Агрегат: avg_price_per_m2_rub (81% coverage), avg_area_pd, units_sold,
+    # units_available — для UI-блока «Конкуренты».
+    # LEFT JOIN — ЖК без маппинга остаются в выдаче (поля = NULL).
     # NB: domrf_kn_objects имеет ~3 snapshot per obj_id → DISTINCT ON по
     # latest snapshot, иначе дубликаты ЖК в выдаче.
     competitor_rows = (
@@ -1341,26 +1347,46 @@ def analyze_parcel(
                   FROM domrf_kn_objects
                  WHERE latitude IS NOT NULL
                  ORDER BY obj_id, snapshot_date DESC NULLS LAST
+            ),
+            obj_pricing AS (
+                SELECT
+                    cm.domrf_obj_id,
+                    ROUND(AVG(ol.price_per_m2_rub)::numeric, 0)  AS avg_price_per_m2_rub,
+                    ROUND(AVG(ol.area_pd)::numeric, 1)            AS avg_area_pd,
+                    COUNT(*) FILTER (WHERE ol.is_sold)            AS units_sold,
+                    COUNT(*) FILTER (WHERE NOT ol.is_sold)        AS units_available,
+                    COUNT(*) FILTER (
+                        WHERE ol.price_per_m2_rub IS NOT NULL
+                    )                                             AS lots_with_price
+                  FROM objective_complex_mapping cm
+                  JOIN objective_lots ol ON ol.project_name = cm.objective_complex_name
+                 GROUP BY cm.domrf_obj_id
             )
-            SELECT obj_id,
+            SELECT o.obj_id,
-                   comm_name,
+                   o.comm_name,
-                   dev_name,
+                   o.dev_name,
-                   obj_class,
+                   o.obj_class,
-                   flat_count,
+                   o.flat_count,
-                   district_name,
+                   o.district_name,
-                   site_status,
+                   o.site_status,
-                   ready_dt,
+                   o.ready_dt,
+                   p.avg_price_per_m2_rub,
+                   p.avg_area_pd,
+                   p.units_sold,
+                   p.units_available,
+                   p.lots_with_price,
                    ST_Distance(
                        ST_SetSRID(ST_MakePoint(o.longitude, o.latitude), 4326)::geography,
                        ST_Centroid(ST_GeomFromText(:wkt, 4326))::geography
                    ) AS distance_m
               FROM latest_obj o
+              LEFT JOIN obj_pricing p ON p.domrf_obj_id = o.obj_id
              WHERE ST_DWithin(
                        ST_SetSRID(ST_MakePoint(o.longitude, o.latitude), 4326)::geography,
                        ST_Centroid(ST_GeomFromText(:wkt, 4326))::geography,
                        3000
                    )
-             ORDER BY CASE site_status WHEN 'Строящиеся' THEN 0 ELSE 1 END,
+             ORDER BY CASE o.site_status WHEN 'Строящиеся' THEN 0 ELSE 1 END,
                       distance_m ASC
              LIMIT 20
         """),
@@ -2146,7 +2172,7 @@ def analyze_parcel(
         "competitors": [dict(c) for c in competitor_rows],
         # D4 (#36): 24-month pipeline competition
         "pipeline_24mo": pipeline_24mo,
-        # D2 (#34): velocity-score из domrf_kn_sale_graph
+        # D2 (#34): velocity-score из objective_corpus_room_month (OBJ-3 migrated)
         "velocity": velocity_data,
         "noise": {
             "score": round(noise_score, 2),

backend/tests/test_velocity.py

@@ -2,8 +2,9 @@
 
 Mock-based — не требуют живой БД.
 Источник данных — objective_corpus_room_month (мигрировано с domrf_kn_sale_graph).
-Mock shape совместим: sales query возвращает те же aliases (obj_id, total_sqm,
+Mock shape совместим: sales query возвращает aliases (obj_id, total_sqm,
-months_with_data, period_start, period_end) через GROUP BY domrf_obj_id.
+months_with_data, period_start, period_end, has_mapping) через LEFT JOIN
+all_competitors + mapped (OBJ-2: ALL competitors included, unmapped has_mapping=False).
 Третий вызов execute — bucket_rows (obj_id, room_bucket, units_sold, sqm_sold).
 """
 
@@ -46,6 +47,7 @@ def _sales_row(
     months: int,
     start: str = "2024-11-01",
     end: str = "2025-04-01",
+    has_mapping: bool = True,
 ) -> MagicMock:
     r = MagicMock()
     start_d = datetime.date.fromisoformat(start)
@@ -56,6 +58,9 @@ def _sales_row(
         "months_with_data": months,
         "period_start": start_d,
         "period_end": end_d,
+        # OBJ-2: LEFT JOIN all_competitors — все конкуренты включены,
+        # has_mapping=True если есть маппинг в objective_complex_mapping.
+        "has_mapping": has_mapping,
     }[k]
     return r
 
@@ -199,9 +204,14 @@ def test_score_capped_at_1():
 
 
 def test_score_zero_when_no_sales_sqm():
-    """total_sqm=0 → None (нет данных, не score=0)."""
+    """total_sqm=0 → VelocityResult с velocity_data_available=False, score=0.
+
+    OBJ-2: функция больше не возвращает None при нулевых продажах —
+    возвращает «пустое» состояние (velocity_data_available=False, source='none').
+    Rosreestr-fallback пропускается т.к. cad_quarter не передан.
+    """
     comp_rows = [_comp_row(1)]
-    # total_sqm=0 — нет продаж → должен вернуть None
+    # total_sqm=0 — нет продаж → velocity_data_available=False
     sales_rows = [_sales_row(1, total_sqm=0.0, months=5)]
 
     db = _make_db(comp_rows=comp_rows, sales_rows=sales_rows)
@@ -212,7 +222,10 @@ def test_score_zero_when_no_sales_sqm():
     ):
         result = compute_velocity(db, parcel_geom_wkt=_PARCEL_WKT)
 
-    assert result is None
+    assert result is not None
+    assert result.velocity_data_available is False
+    assert result.velocity_score == pytest.approx(0.0)
+    assert result.velocity_source == "none"
 
 
 def test_as_dict_structure():
@@ -228,6 +241,8 @@ def test_as_dict_structure():
         period_end="2025-02",
         sample_competitors=[],
         by_room_bucket={"1": {"units": 10, "sqm": 450.0, "complexes_count": 2}},
+        velocity_data_available=True,
+        velocity_source="objective",
     )
     d = vr.as_dict()
     assert "competitors_count" in d
@@ -239,6 +254,9 @@ def test_as_dict_structure():
     assert d["velocity_score"] == pytest.approx(0.333, abs=1e-3)
     assert "by_room_bucket" in d
     assert d["by_room_bucket"]["1"]["units"] == 10
+    # OBJ-2: новые поля velocity_data_available и velocity_source
+    assert d["velocity_data_available"] is True
+    assert d["velocity_source"] == "objective"
 
 
 def test_sample_competitors_top5():