참고소스 수정본

crawler_platform/app/adapters/__init__.py

@@ -0,0 +1,2 @@
+"""Domain adapters keep domain-specific semantics out of ontology core."""
+

crawler_platform/app/adapters/base.py

@@ -0,0 +1,35 @@
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from typing import Any
+
+
+@dataclass(frozen=True, slots=True)
+class DomainAdapter:
+    """Domain-specific semantics plugged into the generic ontology engine."""
+
+    domain: str
+    relation_rules: dict[str, Any] = field(default_factory=dict)
+    entity_type_aliases: dict[str, str] = field(default_factory=dict)
+    entity_name_aliases: dict[str, dict[str, str]] = field(default_factory=dict)
+
+    def relation_rule(self, predicate: str) -> Any | None:
+        return self.relation_rules.get(predicate)
+
+    def normalize_entity_type(self, raw_type: str) -> str | None:
+        key = compact_key(raw_type)
+        return self.entity_type_aliases.get(key)
+
+    def normalize_entity_name(self, raw_name: str, entity_type: str | None = None) -> str | None:
+        if not entity_type:
+            return None
+        aliases = self.entity_name_aliases.get(entity_type, {})
+        return aliases.get(canonical_key(raw_name))
+
+
+def compact_key(value: Any) -> str:
+    return "".join(ch for ch in str(value or "").strip().lower() if ch.isalnum())
+
+
+def canonical_key(value: Any) -> str:
+    return " ".join(str(value or "").strip().lower().split())

crawler_platform/app/adapters/ecommerce/__init__.py

@@ -0,0 +1,2 @@
+"""E-commerce domain adapters."""
+

crawler_platform/app/adapters/ecommerce/perfume.py

@@ -0,0 +1,158 @@
+from __future__ import annotations
+
+from crawler_platform.app.adapters.base import DomainAdapter
+from crawler_platform.app.core.ontology.relation_schema import RelationRule
+
+
+PRODUCT_TYPES = {"Perfume", "Product"}
+
+PERFUME_RELATION_RULES: dict[str, RelationRule] = {
+    "hasBrand": RelationRule(
+        "hasBrand",
+        PRODUCT_TYPES,
+        {"Brand"},
+        page_types={"ProductPage"},
+        source_zones={"product_title", "product_summary", "product_description", "product_detail"},
+        min_confidence=0.72,
+    ),
+    "hasTopNote": RelationRule(
+        "hasTopNote",
+        PRODUCT_TYPES,
+        {"Note"},
+        page_types={"ProductPage"},
+        source_zones={"product_description", "product_detail"},
+        min_confidence=0.78,
+    ),
+    "hasMiddleNote": RelationRule(
+        "hasMiddleNote",
+        PRODUCT_TYPES,
+        {"Note"},
+        page_types={"ProductPage"},
+        source_zones={"product_description", "product_detail"},
+        min_confidence=0.78,
+    ),
+    "hasBaseNote": RelationRule(
+        "hasBaseNote",
+        PRODUCT_TYPES,
+        {"Note"},
+        page_types={"ProductPage"},
+        source_zones={"product_description", "product_detail"},
+        min_confidence=0.78,
+    ),
+    "hasAccord": RelationRule(
+        "hasAccord",
+        PRODUCT_TYPES,
+        {"Accord"},
+        page_types={"ProductPage"},
+        source_zones={"product_description", "product_detail"},
+        min_confidence=0.74,
+    ),
+    "evokesMood": RelationRule(
+        "evokesMood",
+        PRODUCT_TYPES,
+        {"Mood"},
+        page_types={"ProductPage"},
+        source_zones={"product_description", "product_detail", "review_body"},
+        min_confidence=0.75,
+    ),
+    "suitableForSeason": RelationRule(
+        "suitableForSeason",
+        PRODUCT_TYPES,
+        {"Season"},
+        page_types={"ProductPage"},
+        source_zones={"product_description", "product_detail", "review_body"},
+        min_confidence=0.75,
+    ),
+    "suitableForOccasion": RelationRule(
+        "suitableForOccasion",
+        PRODUCT_TYPES,
+        {"Occasion"},
+        page_types={"ProductPage"},
+        source_zones={"product_description", "product_detail", "review_body"},
+        min_confidence=0.75,
+    ),
+    "hasReviewKeyword": RelationRule(
+        "hasReviewKeyword",
+        PRODUCT_TYPES,
+        {"Review"},
+        page_types={"ProductPage", "ReviewPage"},
+        source_zones={"review_body", "product_description", "product_detail"},
+        min_confidence=0.78,
+    ),
+    "soldBy": RelationRule(
+        "soldBy",
+        {"Perfume", "Product", "Brand"},
+        {"Brand"},
+        page_types={"ProductPage", "BrandStoryPage"},
+        source_zones={"product_detail", "product_description", "brand_story_body"},
+        min_confidence=0.82,
+    ),
+    "hasPrice": RelationRule(
+        "hasPrice",
+        PRODUCT_TYPES,
+        literal_value=True,
+        page_types={"ProductPage"},
+        source_zones={"product_summary", "product_detail"},
+        min_confidence=0.82,
+    ),
+    "similarTo": RelationRule(
+        "similarTo",
+        PRODUCT_TYPES,
+        PRODUCT_TYPES,
+        page_types={"ProductPage"},
+        source_zones={"product_description", "product_detail"},
+        min_confidence=0.86,
+    ),
+}
+
+PERFUME_TYPE_ALIASES = {
+    "perfumeproduct": "Perfume",
+    "fragrance": "Perfume",
+    "product": "Perfume",
+    "fragrancenote": "Note",
+    "ingredient": "Note",
+    "scentnote": "Note",
+    "note": "Note",
+    "brand": "Brand",
+    "accord": "Accord",
+    "mood": "Mood",
+    "season": "Season",
+    "occasion": "Occasion",
+    "reviewkeyword": "Review",
+    "review": "Review",
+    "price": "Price",
+}
+
+PERFUME_NAME_ALIASES = {
+    "Brand": {
+        "forment": "FORMENT",
+        "theforment": "FORMENT",
+        "the forment": "FORMENT",
+        "forment korea": "FORMENT",
+    },
+    "Note": {
+        "musk": "Musk",
+        "powdery": "Powdery",
+        "floral": "Floral",
+        "citrus": "Citrus",
+    },
+    "Accord": {
+        "musk": "Musk",
+        "powdery": "Powdery",
+        "floral": "Floral",
+        "citrus": "Citrus",
+    },
+    "Mood": {
+        "fresh": "Fresh",
+        "cozy": "Cozy",
+        "romantic": "Romantic",
+        "elegant": "Elegant",
+    },
+}
+
+PERFUME_ADAPTER = DomainAdapter(
+    domain="perfume",
+    relation_rules=PERFUME_RELATION_RULES,
+    entity_type_aliases=PERFUME_TYPE_ALIASES,
+    entity_name_aliases=PERFUME_NAME_ALIASES,
+)

crawler_platform/app/adapters/registry.py

@@ -0,0 +1,15 @@
+from __future__ import annotations
+
+from crawler_platform.app.adapters.base import DomainAdapter
+from crawler_platform.app.adapters.ecommerce.perfume import PERFUME_ADAPTER
+
+
+ADAPTERS: dict[str, DomainAdapter] = {
+    PERFUME_ADAPTER.domain: PERFUME_ADAPTER,
+}
+
+
+def adapter_for_domain(domain: str | None) -> DomainAdapter | None:
+    if not domain:
+        return None
+    return ADAPTERS.get(str(domain).strip().lower())