o2o-site-AEO/solution/backend/services/collector/tour_api_adapter.py

"""한국관광공사 TourAPI(KorService2) 어댑터 — 숙박·음식점."""
import re
from typing import Optional
from urllib.parse import unquote, urlencode

import httpx

from common.enums import LinkChannel, PlaceCategory
from common.logger import LOG
from config.server_configs import external_api_config
from services.collector.base import CollectedFact, CollectedMedia, RawSource

BASE_URL = "https://apis.data.go.kr/B551011/KorService2"

# fact 의 출처로 남기는 주소.
SOURCE_URL = (
    "https://apis.data.go.kr/B551011/KorService2/detailCommon2"
    "?contentId={content_id}&MobileOS=ETC&MobileApp=o2o-web4ai&_type=json"
)
# 화면에 적을 출처 이름.
SOURCE_LABEL = "한국관광공사 TourAPI"

REQUEST_TIMEOUT = 25
MAX_MEDIA = 30

CONTENT_TYPE_LODGING = "32"
CONTENT_TYPE_RESTAURANT = "39"
_SUPPORTED = (CONTENT_TYPE_LODGING, CONTENT_TYPE_RESTAURANT)

# 상업적 이용이 허용된 공공누리 유형만 담는다.
COMMERCIAL_OK_LICENSES = frozenset({"type1", "type3"})
# 변형(크롭·리사이즈)까지 금지된 유형.
NO_DERIVATIVE_LICENSES = frozenset({"type3", "type4"})

# 이 어댑터가 받는 주소.
_COTID = re.compile(r"[?&]cotid=([0-9a-zA-Z\-]+)", re.I)
_TOUR_SCHEME = re.compile(r"^tour://(\d+)/(\d+)$", re.I)
_HOSTS = ("visitkorea.or.kr",)

_BR = re.compile(r"<br\s*/?>", re.I)
_TAGS = re.compile(r"<[^>]+>")
_HREF = re.compile(r'href=["\']([^"\']+)["\']', re.I)
# "13실", "16명", "대지 면적 11,570㎡" 처럼 숫자에 단위가 붙어 온다.
_LEADING_NUM = re.compile(r"^\s*([\d,]+(?:\.\d+)?)")

# "가능 / 불가" 로 시작하는 필드들.
_NEG_HEAD = ("불가", "없음", "미제공", "불가능", "없슴")
_POS_HEAD = ("가능", "있음", "제공", "무료", "부분")


class TourApiAdapter:
    """TourAPI 콘텐츠 → fact·사진 후보."""

    id = "tour_api"

    # 계약
    def can_handle(self, url: str) -> bool:
        u = (url or "").strip().lower()
        if _TOUR_SCHEME.match(u):
            return True
        return any(h in u for h in _HOSTS) and bool(_COTID.search(u))

    async def fetch(self, url: str, category: Optional[PlaceCategory] = None) -> RawSource:
        channel = LinkChannel.ETC
        key = (external_api_config.tour_api_key or "").strip()
        if not key:
            return RawSource.failure(url, self.id, "TOUR_API_KEY 가 설정되지 않았다", channel)

        parsed = self._parse_url(url)
        if not parsed:
            return RawSource.failure(url, self.id, f"콘텐츠 id 를 찾지 못했다: {url}", channel)
        content_id, content_type = parsed

        try:
            async with httpx.AsyncClient(timeout=REQUEST_TIMEOUT) as client:
                common = await self._call(client, key, "detailCommon2", contentId=content_id)
                if not common:
                    return RawSource.failure(url, self.id, f"detailCommon2 결과 없음(contentId={content_id})", channel)
                common = common[0]

                # URL 에 타입이 없으면 공통정보가 알려준다.
                content_type = content_type or str(common.get("contenttypeid") or "")
                if content_type not in _SUPPORTED:
                    return RawSource.failure(
                        url, self.id,
                        f"지원하지 않는 콘텐츠 타입: {content_type} (숙박 32 · 음식점 39 만 받는다)",
                        channel,
                    )

                intro = await self._call(client, key, "detailIntro2", contentId=content_id, contentTypeId=content_type)
                info = await self._call(client, key, "detailInfo2", contentId=content_id, contentTypeId=content_type)
                images = await self._call(client, key, "detailImage2", contentId=content_id, imageYN="Y")
        except httpx.TimeoutException:
            return RawSource.failure(url, self.id, f"{REQUEST_TIMEOUT}s 안에 응답이 없다", channel)
        except httpx.HTTPError as ex:
            return RawSource.failure(url, self.id, f"네트워크 오류: {ex}", channel)
        except RuntimeError as ex:
            return RawSource.failure(url, self.id, str(ex), channel)

        source_url = SOURCE_URL.format(content_id=content_id)
        intro_row = intro[0] if intro else {}

        facts = self._common_facts(common)
        if content_type == CONTENT_TYPE_LODGING:
            facts += self._lodging_facts(intro_row)
            facts += self._room_facts(info)
        else:
            facts += self._restaurant_facts(intro_row)

        media = self._media(common, images, info if content_type == CONTENT_TYPE_LODGING else [])

        LOG.i(
            f"[tour_api] {common.get('title')} — fact {len(facts)}건 · 사진 {len(media)}장 "
            f"(contentId={content_id} type={content_type})"
        )
        return RawSource(
            url=source_url,
            adapter_id=self.id,
            channel=channel,
            text=self._plain(common.get("overview")),
            facts=facts,
            media=media,
        )

    # HTTP
    async def _call(self, client: httpx.AsyncClient, key: str, op: str, **params) -> list[dict]:
        """오퍼레이션 1회."""
        query = urlencode(
            {"serviceKey": unquote(key), "MobileOS": "ETC", "MobileApp": "o2o-web4ai",
             "_type": "json", "numOfRows": "50", "pageNo": "1", **params},
            safe="",
        )
        res = await client.get(f"{BASE_URL}/{op}?{query}")
        if res.status_code != 200:
            raise RuntimeError(f"{op} HTTP {res.status_code}")

        try:
            payload = res.json()
        except ValueError:
            # 인증 실패·쿼터 초과는 XML 로 온다.
            raise RuntimeError(f"{op} 응답이 JSON 이 아니다: {res.text[:160]}")

        header = payload.get("response", {}).get("header", {})
        code = str(header.get("resultCode") or "")
        if code not in ("0000", "00"):
            raise RuntimeError(f"{op} 실패 [{code}] {header.get('resultMsg')}")

        items = (payload.get("response", {}).get("body", {}).get("items") or {}).get("item")
        if isinstance(items, dict):
            return [items]
        return items or []

    @staticmethod
    def _parse_url(url: str) -> Optional[tuple[str, str]]:
        """(contentId, contentTypeId) — 타입을 모르면 빈 문자열."""
        m = _TOUR_SCHEME.match((url or "").strip())
        if m:
            return m.group(2), m.group(1)
        m = _COTID.search(url or "")
        return (m.group(1), "") if m else None

    # 값 다듬기
    @classmethod
    def _plain(cls, value) -> str:
        """<br> 은 줄바꿈으로, 나머지 태그는 제거."""
        if value is None:
            return ""
        text = _TAGS.sub("", _BR.sub("\n", str(value)))
        return re.sub(r"[ \t]+", " ", text).strip()

    @classmethod
    def _yn(cls, value) -> Optional[bool]:
        """'Y'/'N' 또는 '1'/'0'."""
        v = str(value or "").strip().upper()
        if v in ("Y", "1"):
            return True
        if v in ("N", "0"):
            return False
        return None

    @classmethod
    def _head_bool(cls, value) -> Optional[bool]:
        """'가능 (…)' / '불가 (…)' 처럼 **맨 앞 토큰**이 부호인 필드."""
        text = cls._plain(value)
        if not text:
            return None
        head = text[:6]
        if any(neg in head for neg in _NEG_HEAD):
            return False
        if any(pos in head for pos in _POS_HEAD):
            return True
        return None

    @classmethod
    def _number(cls, value) -> Optional[str]:
        """'13실' · '11,570㎡' → '13' · '11570'."""
        m = _LEADING_NUM.match(cls._plain(value))
        if not m:
            return None
        num = m.group(1).replace(",", "")
        return num if num not in ("", "0") else None

    @staticmethod
    def _fact(key: str, value: Optional[str], *, unit: Optional[str] = None) -> Optional[CollectedFact]:
        if value is None or not str(value).strip():
            return None
        return CollectedFact(
            key=key, value=str(value).strip(),
            scope="unit" if unit else "place", unit_name=unit,
        )

    @classmethod
    def _collect(cls, pairs) -> list[CollectedFact]:
        return [f for f in (cls._fact(*p[:2], unit=p[2] if len(p) > 2 else None) for p in pairs) if f]

    # 업종별 매핑
    def _common_facts(self, common: dict) -> list[CollectedFact]:
        """모든 업종 공통."""
        return []

    def _lodging_facts(self, intro: dict) -> list[CollectedFact]:
        """숙박 detailIntro2 → 사업장 단위 fact."""
        return self._collect([
            ("check_in_time", self._plain(intro.get("checkintime"))[:40]),
            ("check_out_time", self._plain(intro.get("checkouttime"))[:40]),
            ("cancel_policy", self._plain(intro.get("refundregulation"))[:1000]),
            ("cooking_allowed", self._bool_str(self._head_bool(intro.get("chkcooking")))),
            ("parking", self._bool_str(self._head_bool(intro.get("parkinglodging")))),
            ("pickup_service", self._bool_str(self._head_bool(intro.get("pickup")))),
            ("bbq_available", self._bool_str(self._yn(intro.get("barbecue")))),
            ("facilities", self._plain(intro.get("subfacility"))[:500]),
            ("total_rooms", self._number(intro.get("roomcount"))),
            ("accommodation_capacity", self._number(intro.get("accomcountlodging"))),
            # "약 23평" · "대지 면적 11,570㎡" 처럼 단위가 제각각이라 숫자로 뽑지 않고 원문을 싣는다.
            ("building_scale", self._plain(intro.get("scalelodging"))[:200]),
        ])

    def _restaurant_facts(self, intro: dict) -> list[CollectedFact]:
        """음식점 detailIntro2 → 사업장 단위 fact."""
        facts = self._collect([
            ("business_hours", self._plain(intro.get("opentimefood"))[:500]),
            ("closed_days", self._plain(intro.get("restdatefood"))[:200]),
            ("signature_menu", self._plain(intro.get("firstmenu"))[:200]),
            ("parking", self._bool_str(self._head_bool(intro.get("parkingfood")))),
            ("takeout", self._bool_str(self._head_bool(intro.get("packing")))),
            ("reservation_required", self._bool_str(self._head_bool(intro.get("reservationfood")))),
        ])
        # treatmenu 는 "돔베고기 / 멸치국수 / 물만두 등" 처럼 취급 메뉴를 이어 붙여 준다.
        for name in self._split_menu(intro.get("treatmenu")):
            facts.append(CollectedFact(key="menu_name", value=name, scope="unit", unit_name=name))
        return facts

    @classmethod
    def _split_menu(cls, value) -> list[str]:
        """'게살해물요리 / 난자완스 / 특색냉채 등' → 3건."""
        text = cls._plain(value)
        if not text:
            return []
        out: list[str] = []
        for token in re.split(r"[/,·]", text):
            name = re.sub(r"\s*등\s*$", "", token).strip()
            if len(name) >= 2 and name not in out:
                out.append(name)
        return out[:20]

    def _room_facts(self, rooms: list[dict]) -> list[CollectedFact]:
        """숙박 detailInfo2 → **객실 단위** fact."""
        facts: list[CollectedFact] = []
        seen: list[str] = []
        for row in rooms:
            name = self._plain(row.get("roomtitle"))
            if not name or name in seen:
                continue
            seen.append(name)
            facts += self._collect([
                ("room_type", name, name),
                ("standard_capacity", self._number(row.get("roombasecount")), name),
                ("max_capacity", self._number(row.get("roommaxcount")), name),
                ("room_size", self._number(row.get("roomsize2")), name),
                ("weekday_price", self._number(row.get("roomoffseasonminfee1")), name),
                ("weekend_price", self._number(row.get("roomoffseasonminfee2")), name),
                ("peak_price", self._number(row.get("roompeakseasonminfee1")), name),
                ("has_kitchen", self._bool_str(self._yn(row.get("roomcook"))), name),
                ("has_aircon", self._bool_str(self._yn(row.get("roomaircondition"))), name),
                # 객실 편의시설 Y/N.
                ("has_bathroom", self._bool_str(self._yn(row.get("roombathfacility"))), name),
                ("has_tv", self._bool_str(self._yn(row.get("roomtv"))), name),
                ("has_internet", self._bool_str(self._yn(row.get("roominternet"))), name),
                ("has_refrigerator", self._bool_str(self._yn(row.get("roomrefrigerator"))), name),
                ("has_hairdryer", self._bool_str(self._yn(row.get("roomhairdryer"))), name),
                ("has_toiletries", self._bool_str(self._yn(row.get("roomtoiletries"))), name),
                ("has_table", self._bool_str(self._yn(row.get("roomtable"))), name),
                ("has_sofa", self._bool_str(self._yn(row.get("roomsofa"))), name),
                ("has_home_theater", self._bool_str(self._yn(row.get("roomhometheater"))), name),
            ])
        return facts

    @staticmethod
    def _bool_str(value: Optional[bool]) -> Optional[str]:
        """None(모름)과 False(아니오)를 구분한다."""
        return None if value is None else ("true" if value else "false")

    # 사진
    def _media(self, common: dict, images: list[dict], rooms: list[dict]) -> list[CollectedMedia]:
        """대표사진 + 추가사진 + 객실사진."""
        out: list[CollectedMedia] = []
        seen: set[str] = set()
        dropped = 0

        def add(url, license_code, label=None, unit=None) -> None:
            nonlocal dropped
            url = (url or "").strip()
            if not url or url in seen:
                return
            code = str(license_code or "").strip().lower()
            # 라이선스를 모르면 담지 않는다.
            if code not in COMMERCIAL_OK_LICENSES:
                dropped += 1
                return
            seen.add(url)
            out.append(CollectedMedia(
                origin_url=url, label=label or None, unit_name=unit,
                license=code,
            ))

        add(common.get("firstimage"), common.get("cpyrhtDivCd"), label=self._plain(common.get("title")))
        for img in images:
            add(img.get("originimgurl"), img.get("cpyrhtDivCd"), label=self._plain(img.get("imgname")))
        for row in rooms:
            unit = self._plain(row.get("roomtitle")) or None
            for i in range(1, 6):
                add(row.get(f"roomimg{i}"), row.get(f"cpyrhtDivCd{i}"),
                    label=self._plain(row.get(f"roomimg{i}alt")), unit=unit)
            if len(out) >= MAX_MEDIA:
                break

        if dropped:
            LOG.i(f"[tour_api] 상업적 이용이 허용되지 않은 사진 {dropped}장 제외(공공누리 유형)")
        return out[:MAX_MEDIA]