V10.588 提升比價救援穩定性與決策覆蓋

services/competitor_intel_repository.py

@@ -24,7 +24,8 @@ from sqlalchemy import inspect, text
 
 
 PCHOME_MATCH_SCORE_FLOOR = 0.76
-CATALOG_COMPARABLE_SCORE_FLOOR = 0.88
+CATALOG_COMPARABLE_SCORE_FLOOR = 0.85
+CATALOG_IDENTITY_REVIEW_SCORE_FLOOR = 0.95
 UNIT_COMPARABLE_STATUSES = {"unit_comparable", "refresh_unit_comparable"}
 UNIT_PRICE_DECISION_STATUSES = UNIT_COMPARABLE_STATUSES | {"manual_unit_price_required"}
 CATALOG_COMPARABLE_SIGNAL_REASONS = {
@@ -338,13 +339,21 @@ def _jsonb_any_array_predicate(jsonb_expr: str, values: set[str]) -> str:
 
 def _catalog_comparable_sql(alias: str = "la") -> str:
     diagnostic_codes = f"{alias}.diagnostic_codes"
+    signal_sql = _jsonb_any_array_predicate(diagnostic_codes, CATALOG_COMPARABLE_SIGNAL_REASONS)
+    identity_sql = _jsonb_any_array_predicate(diagnostic_codes, CATALOG_COMPARABLE_IDENTITY_REASONS)
+    block_sql = _jsonb_any_array_predicate(diagnostic_codes, CATALOG_COMPARABLE_BLOCK_REASONS)
     return f"""(
         {alias}.attempt_status = 'true_low_confidence'
         AND COALESCE({alias}.hard_veto, false) = false
-        AND COALESCE({alias}.best_match_score, 0) >= {CATALOG_COMPARABLE_SCORE_FLOOR}
-        AND {_jsonb_any_array_predicate(diagnostic_codes, CATALOG_COMPARABLE_SIGNAL_REASONS)}
-        AND {_jsonb_any_array_predicate(diagnostic_codes, CATALOG_COMPARABLE_IDENTITY_REASONS)}
-        AND NOT {_jsonb_any_array_predicate(diagnostic_codes, CATALOG_COMPARABLE_BLOCK_REASONS)}
+        AND {identity_sql}
+        AND NOT {block_sql}
+        AND (
+            (
+                COALESCE({alias}.best_match_score, 0) >= {CATALOG_COMPARABLE_SCORE_FLOOR}
+                AND {signal_sql}
+            )
+            OR COALESCE({alias}.best_match_score, 0) >= {CATALOG_IDENTITY_REVIEW_SCORE_FLOOR}
+        )
     )"""
 
 
@@ -1178,7 +1187,7 @@ def _cached_payload(cache_key: str, producer, ttl_seconds: int = COMPETITOR_INTE
 
 def fetch_competitor_coverage(engine) -> dict:
     return _cached_payload(
-        f"coverage:v13:floor={PCHOME_MATCH_SCORE_FLOOR}:catalog_floor={CATALOG_COMPARABLE_SCORE_FLOOR}:manual_reviews=1:rescore=1:review_no_fresh=1:decision_ready=1:open_queue=1:unknown_freshness=1:decision_support=1:catalog_plan=1",
+        f"coverage:v14:floor={PCHOME_MATCH_SCORE_FLOOR}:catalog_floor={CATALOG_COMPARABLE_SCORE_FLOOR}:identity_floor={CATALOG_IDENTITY_REVIEW_SCORE_FLOOR}:manual_reviews=1:rescore=1:review_no_fresh=1:decision_ready=1:open_queue=1:unknown_freshness=1:decision_support=1:catalog_plan=1",
         lambda: _fetch_competitor_coverage_uncached(engine),
     )
 
@@ -1435,6 +1444,7 @@ def _fetch_competitor_coverage_uncached(engine) -> dict:
         "manual_accept_rate": manual_review_summary["accept_rate"],
         "match_score_floor": PCHOME_MATCH_SCORE_FLOOR,
         "catalog_comparable_score_floor": CATALOG_COMPARABLE_SCORE_FLOOR,
+        "catalog_identity_review_score_floor": CATALOG_IDENTITY_REVIEW_SCORE_FLOOR,
     }
 
 

services/competitor_price_feeder.py

@@ -939,6 +939,9 @@ def _recover_low_score_with_fresh_search(
     momo_name: str,
     momo_price: float = None,
     existing_product_id: str = "",
+    max_terms: int | None = None,
+    max_pages: int | None = None,
+    max_seconds: float | None = None,
 ) -> tuple[Optional[tuple], list[str], int]:
     """
     當 legacy / known-id 候選重評仍低分時，再跑一次 fresh keyword search，
@@ -950,6 +953,9 @@ def _recover_low_score_with_fresh_search(
         momo_name,
         keywords=keywords,
         momo_price=momo_price,
+        max_terms=max_terms,
+        max_pages=max_pages,
+        max_seconds=max_seconds,
     )
     if existing_product_id:
         existing_key = _product_id_key(existing_product_id)
@@ -3053,6 +3059,7 @@ class CompetitorPriceFeeder:
         source: str = "pchome",
         label: str = "已確認身份價格刷新",
         allow_missing_recovery: bool = True,
+        bounded_recovery: bool = False,
     ) -> FeederResult:
         start = time.time()
 
@@ -3064,7 +3071,17 @@ class CompetitorPriceFeeder:
             return FeederResult(0, 0, 0, 0, 0, 0.0)
 
         from services.pchome_crawler import PChomeCrawler
-        crawler = PChomeCrawler(timeout=REQUEST_TIMEOUT, delay=RATE_DELAY)
+        crawler_timeout = BACKFILL_REQUEST_TIMEOUT if bounded_recovery else REQUEST_TIMEOUT
+        crawler = PChomeCrawler(timeout=crawler_timeout, delay=RATE_DELAY)
+        recovery_kwargs = (
+            {
+                "max_terms": BACKFILL_MAX_SEARCH_TERMS,
+                "max_pages": BACKFILL_SEARCH_MAX_PAGES,
+                "max_seconds": BACKFILL_MAX_SECONDS_PER_SKU,
+            }
+            if bounded_recovery
+            else {}
+        )
 
         requested_ids = [
             str(item.get("competitor_product_id") or "").strip()
@@ -3120,6 +3137,7 @@ class CompetitorPriceFeeder:
                         momo_name,
                         momo_price=momo_price,
                         existing_product_id=competitor_product_id,
+                        **recovery_kwargs,
                     )
                     if recovered:
                         best_product, score, diagnostics = recovered
@@ -3313,6 +3331,7 @@ class CompetitorPriceFeeder:
                             momo_name,
                             momo_price=momo_price,
                             existing_product_id=competitor_product_id,
+                            **recovery_kwargs,
                         )
                         if recovered:
                             recovered_product, recovered_score, recovered_diagnostics = recovered
@@ -3595,6 +3614,7 @@ class CompetitorPriceFeeder:
             source=source,
             label="identity_v2 過期價格搜尋救援",
             allow_missing_recovery=True,
+            bounded_recovery=True,
         )
 
     def run_retryable_candidate_revalidation(