wooo/ewoooc
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

修正 PChome 近門檻商品比對規則

Browse Source
This commit is contained in:
OoO
2026-05-24 10:22:59 +08:00
committed by AiderHeal Bot
parent 4ae3a39970
commit d2dea022ea
3 changed files with 315 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

194
services/marketplace_product_matcher.py
View File

@@ -313,6 +313,16 @@ SEARCH_NOISE_TOKENS = {
}
SEARCH_IDENTITY_ANCHORS = (
"控油清爽防曬棒",
"蔓越莓私密清潔慕斯",
"金縷梅私密清潔慕斯",
"光采奪目眼部飾底乳",
"男性私密沐浴露",
"私密沐浴露",
"hydsto 車載香薰",
"電動刮鬍刀 s101",
"磁吸控油定妝噴霧",
"修容打亮棒",
"私密潔淨凝露",
"柔霧裸唇膏",
"潤浸保濕清爽身體乳液",
@@ -436,6 +446,15 @@ SEARCH_BROAD_ANCHORS = {
}
VARIANT_SENSITIVE_KEYWORDS = {
"香氛蠟燭",
"芬香蠟燭",
"蠟燭",
"擴香",
"融蠟燈",
"車用香氛",
"香味",
"私密清潔慕斯",
"定妝噴霧",
"妝前防護乳",
"妝前乳",
"素顏霜",
@@ -453,9 +472,26 @@ VARIANT_SENSITIVE_KEYWORDS = {
"蜜粉餅",
"粉底棒",
"遮瑕棒",
"修容打亮棒",
}
VARIANT_OPTION_COLOR_WORDS = {
"茉莉花",
"梔子花",
"白麝香",
"黑麝香",
"青檸羅勒",
"炭木香",
"無花果",
"清甜柚香",
"杏仁牛奶",
"完熟白桃",
"原味",
"草莓",
"蔓越莓",
"金縷梅",
"柔焦霧面",
"水光亮面",
"黑色",
"棕色",
"咖啡色",
@@ -544,7 +580,13 @@ PRODUCT_TYPES = {
"私密噴霧": ("私密噴霧", "抑菌噴霧", "醒肌抑菌噴霧"),
"私密凝露": ("凝露", "激淨凝露", "緊實凝露", "亮白凝露"),
"護甲油": ("護甲油", "亮油", "top coat"),
"定妝噴霧": ("定妝噴霧", "setting spray"),
"修容打亮棒": ("修容打亮棒", "修容棒", "打亮棒"),
"刮鬍刀": ("刮鬍刀", "刮胡刀", "shaver", "razor"),
"體香膏": ("體香膏", "體香劑", "deodorant"),
"電動牙刷": ("電動牙刷", "聲波電動牙刷", "electric toothbrush"),
"洗手慕斯": ("洗手慕斯", "洗手泡泡", "hand wash foam"),
"私密慕斯": ("私密清潔慕斯", "私密防護慕絲", "私密慕斯"),
"足膜": ("足膜", "足部膜", "足部去角質"),
"妝前乳": ("妝前乳", "妝前防護乳", "妝前隔離", "primer"),
"素顏霜": ("素顏霜", "tone up cream"),
@@ -557,7 +599,7 @@ PRODUCT_TYPES = {
"粉底棒": ("粉底棒", "foundation stick"),
"精華": ("精華", "精華液", "essence", "serum", "安瓶"),
"化妝水": ("化妝水", "機能水", "toner", "lotion"),
"乳液": ("乳液", "emulsion", "milk"),
"乳液": ("乳液", "按摩乳", "emulsion", "milk"),
"面霜": ("面霜", "乳霜", "霜", "cream"),
"防曬": ("防曬", "spf", "uv", "sunscreen"),
"洗面乳": ("洗面乳", "洗顏", "潔面", "cleanser", "foam"),
@@ -728,6 +770,7 @@ def normalize_product_text(value: str) -> str:
)
text = text.replace("×", "x").replace("＊", "x").replace("*", "x")
text = text.replace("／", "/").replace("＆", "&")
text = text.replace("粧", "妝")
text = re.sub(r"[\u3000\r\n\t]+", " ", text)
text = text.lower()
text = re.sub(r"[?？]+", " ", text)
@@ -1850,6 +1893,34 @@ def score_marketplace_match(
):
score += 0.10
reasons.append("shared_identity_anchor_lactacyd_wash")
if (
"私密潔膚露" in shared_anchor
and {"femfresh", "芳芯"} & (left.brand_tokens | right.brand_tokens)
and brand_score >= 0.95
and not hard_veto
and price_penalty == 0
and type_score >= 0.95
and spec_score >= 0.85
and token_score >= 0.30
and sequence_score >= 0.45
and not variant_descriptor_conflict
):
score += 0.06
reasons.append("shared_identity_anchor_femfresh_wash")
if (
"私密沐浴露" in shared_anchor
and {"vigill", "婦潔"} & (left.brand_tokens | right.brand_tokens)
and brand_score >= 0.95
and not hard_veto
and price_penalty == 0
and type_score >= 0.95
and spec_score >= 0.70
and token_score >= 0.45
and sequence_score >= 0.55
and not variant_descriptor_conflict
):
score += 0.06
reasons.append("shared_identity_anchor_vigill_private_wash")
if (
"私密潔淨凝露" in shared_anchor
and {"relove"} <= (left.brand_tokens | right.brand_tokens)
@@ -1892,6 +1963,52 @@ def score_marketplace_match(
):
score += 0.12
reasons.append("shared_identity_anchor_karadium_eye_stick")
if (
_has_seche_vite_top_coat_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 >= 0.70
and not variant_descriptor_conflict
):
score += 0.04
reasons.append("shared_identity_anchor_seche_vite_top_coat")
if (
_has_xiaomi_s101_shaver_alignment(left, right)
and brand_score >= 0.95
and not hard_veto
and price_penalty == 0
and token_score >= 0.60
and not variant_descriptor_conflict
):
score += 0.04
reasons.append("shared_model_token_xiaomi_s101_shaver")
if (
_has_hinoki_roller_oil_alignment(left, right)
and brand_score >= 0.95
and not hard_veto
and price_penalty == 0
and type_score >= 0.95
and spec_score >= 0.85
and sequence_score >= 0.50
and not variant_descriptor_conflict
):
score += 0.04
reasons.append("shared_identity_anchor_hinoki_roller_oil")
if (
_has_brush_baby_wildones_toothbrush_alignment(left, right)
and brand_score >= 0.95
and not hard_veto
and price_penalty == 0
and type_score >= 0.95
and token_score >= 0.78
and sequence_score >= 0.90
and not variant_descriptor_conflict
):
score += 0.04
reasons.append("shared_model_token_brush_baby_wildones")
if (
shared_anchor
and len(shared_anchor.replace(" ", "")) >= 5
@@ -2108,14 +2225,15 @@ def _extract_anchor_phrases(token: str) -> list[str]:
if "瞬效" in cleaned and "b5" in cleaned and "玻尿酸" in cleaned and "精華" in cleaned:
phrases.append("瞬效b5玻尿酸精華")
for anchor in SEARCH_IDENTITY_ANCHORS:
if anchor not in cleaned:
anchor_phrase = _clean_search_phrase(anchor)
if not anchor_phrase or anchor_phrase not in cleaned:
continue
if re.search(r"[\u4e00-\u9fff]", anchor):
prefix_width = 0 if len(anchor) >= 5 else (4 if len(anchor) >= 3 else 6)
match = re.search(rf"([\u4e00-\u9fff]{{0,{prefix_width}}}{re.escape(anchor)})", cleaned)
phrase = match.group(1) if match else anchor
if re.search(r"[\u4e00-\u9fff]", anchor_phrase):
prefix_width = 0 if len(anchor_phrase) >= 5 else (4 if len(anchor_phrase) >= 3 else 6)
match = re.search(rf"([\u4e00-\u9fff]{{0,{prefix_width}}}{re.escape(anchor_phrase)})", cleaned)
phrase = match.group(1) if match else anchor_phrase
else:
phrase = anchor
phrase = anchor_phrase
phrase = _clean_search_phrase(phrase)
if phrase.startswith("款") and len(phrase) > 2:
phrase = phrase[1:]
@@ -2214,6 +2332,64 @@ def _has_serum_formulation_conflict(left: ProductIdentity, right: ProductIdentit
return bool(left_hit and right_hit and left_hit != right_hit)
def _has_taicend_baby_spray_equivalence(left: ProductIdentity, right: ProductIdentity) -> bool:
brand_tokens = {"taicend", "泰陞"}
return (
bool(left.brand_tokens & brand_tokens)
and bool(right.brand_tokens & brand_tokens)
and "屁屁噴" in left.searchable_name
and "屁屁噴" in right.searchable_name
and _has_overlapping_base_spec(left, right)
)
def _has_seche_vite_top_coat_alignment(left: ProductIdentity, right: ProductIdentity) -> bool:
brand_tokens = {"seche", "vite"}
return (
brand_tokens <= (left.brand_tokens | right.brand_tokens)
and bool(left.brand_tokens & brand_tokens)
and bool(right.brand_tokens & brand_tokens)
and "快乾亮油" in left.searchable_name
and "快乾亮油" in right.searchable_name
)
def _has_xiaomi_s101_shaver_alignment(left: ProductIdentity, right: ProductIdentity) -> bool:
brand_tokens = {"xiaomi", "小米", "小米有品"}
return (
bool(left.brand_tokens & brand_tokens)
and bool(right.brand_tokens & brand_tokens)
and "s101" in left.searchable_name
and "s101" in right.searchable_name
and "電動刮鬍刀" in left.searchable_name
and "電動刮鬍刀" in right.searchable_name
)
def _has_hinoki_roller_oil_alignment(left: ProductIdentity, right: ProductIdentity) -> bool:
return (
"檜山坊" in left.brand_tokens
and "檜山坊" in right.brand_tokens
and "檜木精油" in left.searchable_name
and "檜木精油" in right.searchable_name
and "滾珠瓶" in left.searchable_name
and "滾珠瓶" in right.searchable_name
and _has_overlapping_base_spec(left, right)
)
def _has_brush_baby_wildones_toothbrush_alignment(left: ProductIdentity, right: ProductIdentity) -> bool:
brand_tokens = {"brush", "baby", "wildones"}
return (
brand_tokens <= left.brand_tokens
and brand_tokens <= right.brand_tokens
and "電動牙刷" in left.searchable_name
and "電動牙刷" in right.searchable_name
and "0-10y" in left.searchable_name
and "0-10y" in right.searchable_name
)
def _is_variant_sensitive_identity(
left: ProductIdentity,
right: ProductIdentity,
@@ -2232,6 +2408,10 @@ def _is_variant_sensitive_identity(
def _has_variant_descriptor_conflict(left: ProductIdentity, right: ProductIdentity, shared_anchor: str) -> bool:
if _has_serum_formulation_conflict(left, right, shared_anchor):
return True
if _has_taicend_baby_spray_equivalence(left, right):
return False
if _has_brush_baby_wildones_toothbrush_alignment(left, right):
return False
if (
shared_anchor
and shared_anchor not in SEARCH_BROAD_ANCHORS