V10.393 refine bundle plus matching

config.py

@@ -325,7 +325,7 @@ YOUTUBE_API_KEY = os.getenv('YOUTUBE_API_KEY', '')
 # ==========================================
 # 系統版本與路徑
 # ==========================================
-SYSTEM_VERSION = "V10.392"
+SYSTEM_VERSION = "V10.393"
 LOG_FILE_PATH = os.path.join(BASE_DIR, 'logs/system.log')
 public_url = PUBLIC_URL  # 用於模板顯示
 

docs/memory/history_logs.md

@@ -13,6 +13,7 @@
 ## 📅 詳細更新日誌 (考古存檔)
 
 ### 2026-05-21：瀏覽器測試守門與 PChome 熱路徑優化
+- **V10.393 組合包 `+` 判定修正與 catalog 補強**: marketplace matcher 的組合包件數判定會先排除 `SPF50+`、`PA++++` 等防曬係數加號，以及 `NTT80+AL414` 這類純型號碼串，避免把防曬品與 OPI 套組的規格/型號加號誤判成多一個商品組件；CeraVe 三件組 vs 兩件組仍維持 `multi_component_count_conflict` hard veto。同版收緊品牌 alias 判定，避免只有品牌名就觸發商品線加成，並補 Baan 貝恩嬰兒修護唇膏「原味/草莓」catalog listing 放行。
 - **V10.392 組合包件數防錯配**: marketplace matcher 新增 `multi_component_count_conflict`，當 MOMO 與 PChome 都是 `+`/`＋` 組合包但組件數不同時直接進 `not_comparable`，避免三件組被拿去跟兩件組做總價告警；同步把該原因加入 evidence flags，讓告警與審核畫面可以清楚顯示「組合包件數不同」。
 - **V10.391 多款任選 catalog listing 防錯配**: marketplace matcher 新增 `catalog_variant_listing_alignment`，當 MOMO/PChome 雙方都是多款/多色/多香味任選 listing，且商品線、規格與類型一致時，可放行香氛擴香罐、香氛蠟燭等 catalog 型同款；同時把 Relove 菸鹼醯胺 vs 胺基酸私密清潔凝露列為變體衝突，並讓 competitor feeder 不再只因 `strong_exact_spec_match` 就把低分候選視為 recoverable，避免只同規格但品線不同的商品回寫正式比價。
 - **V10.390 PChome 近門檻商品比對規則**: marketplace matcher 補 17 組近門檻真同款召回與錯配防線，包含 OBgE 防曬棒、ARTMIS 私密清潔慕斯、Seche Vite 快乾亮油、TAICEND 屁屁噴、femfresh / VIGILL 私密清潔、Solone 眼部飾底乳、HYDSTO 車載香薰、小米 S101 刮鬍刀、PRAMY 定妝噴霧、I'M MEME 修容打亮棒、檜山坊滾珠精油、ARM&HAMMER 體香膏、Brush Baby WildOnes 電動牙刷與 Palmer's 按摩乳；同時把香氛/私密慕斯/定妝噴霧 finish 差異列為 variant-sensitive，避免不同香味、蔓越莓 vs 金縷梅、柔焦霧面 vs 水光亮面被誤推成直接價格告警。

services/marketplace_product_matcher.py

@@ -826,6 +826,19 @@ def _extract_model_tokens(text: str) -> set[str]:
     return tokens
 
 
+def _brand_alias_present(text: str, alias_norm: str, text_tokens: set[str]) -> bool:
+    if not alias_norm:
+        return False
+    if re.search(r"[\u4e00-\u9fff]", alias_norm):
+        return alias_norm in text
+    if " " not in alias_norm and alias_norm in text_tokens:
+        return True
+    if re.fullmatch(r"[a-z0-9][a-z0-9\s'&.-]*", alias_norm):
+        pattern = r"(?<![a-z0-9])" + re.escape(alias_norm).replace(r"\ ", r"\s+") + r"(?![a-z0-9])"
+        return bool(re.search(pattern, text))
+    return alias_norm in text
+
+
 def _known_brand_tokens(text: str) -> set[str]:
     tokens: set[str] = set()
     try:
@@ -843,9 +856,10 @@ def _known_brand_tokens(text: str) -> set[str]:
         for alias in aliases:
             alias_map[alias.lower()] = canonical
 
+    text_tokens = _tokenize(text)
     for alias, canonical in alias_map.items():
         alias_norm = normalize_product_text(alias)
-        if alias_norm and alias_norm in text:
+        if _brand_alias_present(text, alias_norm, text_tokens):
             tokens.add(canonical)
             tokens.update(
                 token for token in _tokenize(alias_norm)
@@ -1233,7 +1247,7 @@ def _has_bundle_offer(identity: ProductIdentity) -> bool:
 
 
 def _has_multi_component(identity: ProductIdentity) -> bool:
-    text = identity.normalized_name
+    text = _component_separator_text(identity)
     return bool(
         "+" in text
         or "＋" in text
@@ -1241,8 +1255,21 @@ def _has_multi_component(identity: ProductIdentity) -> bool:
     )
 
 
-def _multi_component_count(identity: ProductIdentity) -> int:
+def _component_separator_text(identity: ProductIdentity) -> str:
     text = identity.normalized_name
+    text = re.sub(r"spf\s*(\d+)\s*[+＋]+", r"spf\1", text, flags=re.I)
+    text = re.sub(r"pa\s*[+＋]{1,5}", "pa", text, flags=re.I)
+    text = re.sub(
+        r"\b[a-z]{1,6}\d{1,6}\s*[+＋]\s*[a-z]{1,6}\d{1,6}\b",
+        lambda match: re.sub(r"[+＋]", " ", match.group(0)),
+        text,
+        flags=re.I,
+    )
+    return text
+
+
+def _multi_component_count(identity: ProductIdentity) -> int:
+    text = _component_separator_text(identity)
     if not ("+" in text or "＋" in text):
         return 1
     parts = [
@@ -1478,7 +1505,7 @@ def _has_strong_product_line_signal(
     token_score: float,
     chinese_name_score: float,
 ) -> bool:
-    shared_core = left.core_tokens & right.core_tokens
+    shared_core = (left.core_tokens & right.core_tokens) - left.brand_tokens - right.brand_tokens
     shared_latin_or_model = {
         token for token in shared_core
         if re.fullmatch(r"[a-z][a-z0-9-]{3,}", token)
@@ -1486,7 +1513,7 @@ def _has_strong_product_line_signal(
     }
     if shared_latin_or_model and token_score >= 0.50:
         return True
-    return token_score >= 0.56 and chinese_name_score >= 0.45
+    return bool(shared_core) and token_score >= 0.56 and chinese_name_score >= 0.45
 
 
 def _has_safe_exact_spec_signal(
@@ -2062,6 +2089,18 @@ def score_marketplace_match(
     ):
         score += 0.06
         reasons.append("catalog_variant_listing_alignment")
+    if (
+        _has_baan_baby_lip_catalog_alignment(left, right)
+        and brand_score >= 0.95
+        and not hard_veto
+        and price_penalty == 0
+        and type_score >= 0.95
+        and token_score >= 0.70
+        and sequence_score >= 0.45
+        and not variant_descriptor_conflict
+    ):
+        score += 0.05
+        reasons.append("catalog_variant_listing_alignment_baan_lip")
     if (
         shared_anchor
         and len(shared_anchor.replace(" ", "")) >= 5
@@ -2443,6 +2482,20 @@ def _has_brush_baby_wildones_toothbrush_alignment(left: ProductIdentity, right:
     )
 
 
+def _has_baan_baby_lip_catalog_alignment(left: ProductIdentity, right: ProductIdentity) -> bool:
+    brand_tokens = {"baan", "貝恩"}
+    left_options = _explicit_variant_option_tokens(left)
+    right_options = _explicit_variant_option_tokens(right)
+    return (
+        bool(left.brand_tokens & brand_tokens)
+        and bool(right.brand_tokens & brand_tokens)
+        and "嬰兒修護唇膏" in left.searchable_name
+        and "嬰兒修護唇膏" in right.searchable_name
+        and {"原味", "草莓"} <= left_options
+        and {"原味", "草莓"} <= right_options
+    )
+
+
 def _is_relove_private_cleanser_line(left: ProductIdentity, right: ProductIdentity) -> bool:
     return (
         "relove" in (left.brand_tokens | right.brand_tokens)
@@ -2496,6 +2549,8 @@ def _has_variant_descriptor_conflict(left: ProductIdentity, right: ProductIdenti
         return False
     if _has_brush_baby_wildones_toothbrush_alignment(left, right):
         return False
+    if _has_baan_baby_lip_catalog_alignment(left, right):
+        return False
     if _is_relove_private_cleanser_line(left, right):
         return False
     if (

tests/test_marketplace_product_matcher.py

@@ -797,6 +797,35 @@ def test_marketplace_matcher_rejects_fragrance_formula_and_finish_variant_mismat
         assert expected_reason in diagnostics.reasons
 
 
+def test_marketplace_matcher_ignores_sunscreen_and_model_plus_markers_for_bundle_count():
+    from services.marketplace_product_matcher import score_marketplace_match
+
+    cases = [
+        (
+            "【O.P.I】受損型硬甲油記得卸組-亮麗增強基礎護甲油15mL+專業去光水110mL-NTT80+AL414(禮物/官方直營)",
+            "OPI 護甲油記得卸組（亮麗增強基礎護甲油15mL+去光水110mL）",
+        ),
+        (
+            "【Neutrogena 露得清】新品上市★水感透亮防曬乳SPF50+ PA++++(40mlx2入組)",
+            "露得清水感透亮防曬乳SPF50+ 40ml *2入",
+        ),
+        (
+            "【AGE20】輕透光潤色素顏霜(SPF50+/PA++++/台灣總代理)",
+            "AGE20’S 輕透光潤色素顏霜SPF50+/PA++++",
+        ),
+        (
+            "【Estee Lauder 雅詩蘭黛】粉持久完美鎖妝氣墊粉餅(SPF45/PA+++/一盒一蕊/IU同款)",
+            "【ESTEE LAUDER雅詩蘭黛】粉持久完美鎖妝氣墊粉餅SPF45/PA+++",
+        ),
+    ]
+
+    for momo_name, competitor_name in cases:
+        diagnostics = score_marketplace_match(momo_name, competitor_name, momo_price=1000, competitor_price=900)
+        assert diagnostics.hard_veto is False
+        assert diagnostics.score >= 0.76
+        assert "multi_component_count_conflict" not in diagnostics.reasons
+
+
 def test_marketplace_matcher_promotes_multi_variant_catalog_listings():
     from services.marketplace_product_matcher import score_marketplace_match
 
@@ -822,6 +851,31 @@ def test_marketplace_matcher_promotes_multi_variant_catalog_listings():
         assert "catalog_variant_listing_alignment" in diagnostics.reasons
 
 
+def test_marketplace_matcher_requires_non_brand_product_line_evidence():
+    from services.marketplace_product_matcher import score_marketplace_match
+
+    diagnostics = score_marketplace_match(
+        "【CEZANNE】柔潤腮紅",
+        "Cezanne",
+    )
+
+    assert diagnostics.score < 0.76
+    assert "strong_product_line_match" not in diagnostics.reasons
+
+
+def test_marketplace_matcher_promotes_baan_lip_catalog_with_same_options():
+    from services.marketplace_product_matcher import score_marketplace_match
+
+    diagnostics = score_marketplace_match(
+        "【Baan 貝恩】嬰兒修護唇膏(草莓/原味) 4.5g/個 (新東海藥局)",
+        "【貝恩】嬰兒修護唇膏(原味/草莓)",
+    )
+
+    assert diagnostics.score >= 0.76
+    assert diagnostics.hard_veto is False
+    assert "catalog_variant_listing_alignment_baan_lip" in diagnostics.reasons
+
+
 def test_marketplace_matcher_rejects_refill_core_vs_case_only_pack():
     from services.marketplace_product_matcher import score_marketplace_match