fix(tradein-estimator): atomic IMV link, batched yandex history, drop redundant CAST

3 fixes inside services/estimator.py from 2026-05-24 audit:

* IMV link UPDATE moved inside main tx (finding #4) — closes race where two
  concurrent estimates sharing a cache_key could overwrite each other's
  estimate_id after the main commit.
* _save_yandex_history_items batched under single try/except (finding #5) —
  prior per-item db.rollback() destroyed the parent transaction; next
  iteration could run on a rolled-back session.
* Drop redundant CAST inside IS NOT NULL predicates in _fetch_analogs
  (finding #9) — CAST(NULL AS T) IS NOT NULL is equivalent to NULL IS NOT NULL
  but evaluates per-row. THEN-branch CASTs preserved (typed arithmetic).

test: update test_save_history_items_db_error_continues → batch rollback semantics

tradein-mvp/backend/app/services/estimator.py

@@ -343,8 +343,15 @@ def _save_yandex_history_items(
     Idempotent via UNIQUE (source, ext_item_id); we synthesize ext_item_id from
     (address|date|area|floor) hash since Yandex history items don't carry an
     explicit ID.
+
+    Batch semantics (closes finding #5 from 2026-05-24 audit): all items
+    inserted under a single try/except; on any failure the whole batch is
+    rolled back as a unit (prior per-item rollback() destroyed the parent tx).
     """
-    saved = 0
+    if not result.history_items:
+        return 0
+
+    rows = []
     for item in result.history_items:
         # Synthesize stable ext_item_id (no native ID in valuation page)
         ext_seed = (
@@ -352,48 +359,51 @@ def _save_yandex_history_items(
             f"{item.start_price}|{item.last_price}"
         )
         ext_item_id = hashlib.sha256(ext_seed.encode("utf-8")).hexdigest()[:32]
-        try:
+        rows.append({
-            db.execute(
+            "ext_id": ext_item_id,
-                text(
+            "rooms": item.rooms,
-                    """
+            "area": item.area_m2,
-                    INSERT INTO house_placement_history (
+            "floor": item.floor,
-                        source, ext_item_id,
+            "start_price": item.start_price,
-                        rooms, area_m2, floor,
+            "last_price": item.last_price,
-                        start_price, start_price_date,
+            "publish_date": item.publish_date,
-                        last_price, last_price_date,
+            "exposure": item.exposure_days,
-                        exposure_days,
+            "raw": json.dumps(item.model_dump(mode="json"), ensure_ascii=False),
-                        raw_payload
+        })
-                    ) VALUES (
+
-                        'yandex_valuation', :ext_id,
+    sql = text(
-                        :rooms, :area, :floor,
+        """
-                        :start_price, :publish_date,
+        INSERT INTO house_placement_history (
-                        :last_price, :publish_date,
+            source, ext_item_id,
-                        :exposure,
+            rooms, area_m2, floor,
-                        CAST(:raw AS jsonb)
+            start_price, start_price_date,
-                    )
+            last_price, last_price_date,
-                    ON CONFLICT (source, ext_item_id) DO NOTHING
+            exposure_days,
-                    """
+            raw_payload
-                ),
+        ) VALUES (
-                {
+            'yandex_valuation', :ext_id,
-                    "ext_id": ext_item_id,
+            :rooms, :area, :floor,
-                    "rooms": item.rooms,
+            :start_price, :publish_date,
-                    "area": item.area_m2,
+            :last_price, :publish_date,
-                    "floor": item.floor,
+            :exposure,
-                    "start_price": item.start_price,
+            CAST(:raw AS jsonb)
-                    "last_price": item.last_price,
+        )
-                    "publish_date": item.publish_date,
+        ON CONFLICT (source, ext_item_id) DO NOTHING
-                    "exposure": item.exposure_days,
+        """
-                    "raw": json.dumps(item.model_dump(mode="json"), ensure_ascii=False),
+    )
-                },
+
-            )
+    try:
-            saved += 1
+        for row in rows:
-        except Exception as e:
+            db.execute(sql, row)
-            logger.warning("yandex_valuation: failed to save history item: %s", e)
-            db.rollback()
-            continue
-    if saved:
         db.commit()
-    return saved
+        return len(rows)
+    except Exception as e:
+        logger.warning(
+            "yandex_valuation: failed to save history batch (%d items): %s",
+            len(rows), e,
+        )
+        db.rollback()
+        return 0
 
 
 # ── Public ───────────────────────────────────────────────────────────────────
@@ -667,25 +677,25 @@ async def estimate_quality(
             "expires_at": expires_at,
         },
     )
-    db.commit()
 
-    # Link saved IMV evaluation к этому estimate_id (для analytics joining)
+    # Link saved IMV evaluation к этому estimate_id атомарно с основным INSERT
+    # (closes finding #4 from 2026-05-24 audit — prior code committed estimate first,
+    # then UPDATEd IMV in a separate tx, racing against concurrent estimators
+    # sharing the same cache_key).
     if imv_eval is not None:
-        try:
+        db.execute(
-            db.execute(
+            text(
-                text(
+                """
-                    """
+                UPDATE avito_imv_evaluations
-                    UPDATE avito_imv_evaluations
+                SET estimate_id = CAST(:estimate_id AS uuid)
-                    SET estimate_id = CAST(:estimate_id AS uuid)
+                WHERE cache_key = :cache_key
-                    WHERE cache_key = :cache_key
+                  AND (estimate_id IS NULL OR estimate_id = CAST(:estimate_id AS uuid))
-                      AND (estimate_id IS NULL OR estimate_id = CAST(:estimate_id AS uuid))
+                """
-                    """
+            ),
-                ),
+            {"estimate_id": str(estimate_id), "cache_key": imv_eval.cache_key},
-                {"estimate_id": str(estimate_id), "cache_key": imv_eval.cache_key},
+        )
-            )
+
-            db.commit()
+    db.commit()
-        except Exception as e:
-            logger.warning("imv: failed to link estimate_id to evaluation: %s", e)
 
     logger.info(
         "estimate: id=%s addr=%s rooms=%d area=%.1f → median=%d (n=%d, conf=%s)%s%s",
@@ -819,13 +829,13 @@ def _fetch_analogs(
                         -- без типа → PostgreSQL "could not determine data type of parameter"
                         -- (AmbiguousParameter). Явный тип снимает неоднозначность.
                         + CASE
-                            WHEN CAST(:target_year AS integer) IS NOT NULL
+                            WHEN :target_year IS NOT NULL
                                  AND year_built IS NOT NULL
                             THEN abs(year_built - CAST(:target_year AS integer)) / 12.0
                             ELSE 0
                           END
                         + CASE
-                            WHEN CAST(:target_house_type AS text) IS NOT NULL
+                            WHEN :target_house_type IS NOT NULL
                                  AND house_type IS NOT NULL
                                  AND house_type <> CAST(:target_house_type AS text)
                             THEN 1.5
@@ -838,13 +848,13 @@ def _fetch_analogs(
                             ST_Distance(geom::geography, ST_MakePoint(:lon, :lat)::geography)
                                 / 1000.0
                             + CASE
-                                WHEN CAST(:target_year AS integer) IS NOT NULL
+                                WHEN :target_year IS NOT NULL
                                      AND year_built IS NOT NULL
                                 THEN abs(year_built - CAST(:target_year AS integer)) / 12.0
                                 ELSE 0
                               END
                             + CASE
-                                WHEN CAST(:target_house_type AS text) IS NOT NULL
+                                WHEN :target_house_type IS NOT NULL
                                      AND house_type IS NOT NULL
                                      AND house_type <> CAST(:target_house_type AS text)
                                 THEN 1.5

tradein-mvp/backend/tests/test_estimator_yandex_integration.py

@@ -188,10 +188,12 @@ def test_save_history_items_ext_id_stable_across_calls():
     assert ext_ids_1 == ext_ids_2
 
 
-def test_save_history_items_db_error_continues():
+def test_save_history_items_db_error_rolls_back_batch():
-    """One item failing doesn't abort the others."""
+    """Any item failing rolls back the whole batch — batch semantics (finding #5)."""
     db = MagicMock()
     db.execute.side_effect = [RuntimeError("first row fails"), None]
     result = _sample_result()
     saved = _save_yandex_history_items(db, result)
-    assert saved == 1  # second item still saved
+    assert saved == 0  # whole batch rolled back
+    db.rollback.assert_called_once()
+    db.commit.assert_not_called()