app/services/platform_api_registry.py

"""Platform API credential registry — providers, env fallbacks, resolution order."""

from __future__ import annotations

import json
import re
from dataclasses import dataclass
from typing import Literal

import httpx
import structlog
from sqlalchemy.ext.asyncio import AsyncSession

from app.config import get_settings
from app.core.crypto.credential_crypto import decrypt_secret, mask_secret
from app.repositories.platform_api_credential_repo import PlatformApiCredentialRepository
from app.services.platform_pricing import (
    PriceContext,
    PlatformPrice,
    extract_price_from_html,
    format_price,
)

logger = structlog.get_logger(__name__)

_CACHE_PREFIX = "platform_api_cred:v1:"
_CACHE_TTL = 300

ProviderId = Literal["google_books", "rainforest", "scrapingbee"]

PROVIDER_SCOPE_PREFIX = "@provider:"

PROVIDER_DEFINITIONS: dict[str, dict] = {
    "scrapingbee": {
        "display_name": "ScrapingBee",
        "description": "Shared proxy for B&N, Apple, Kobo, ThriftBooks, AbeBooks, Bookshop",
    },
    "rainforest": {
        "display_name": "Rainforest API",
        "description": "Amazon product data API",
    },
    "google_books": {
        "display_name": "Google Books API",
        "description": "Google Books volumes and pricing",
    },
}

PLATFORM_API_META: dict[str, dict] = {
    "amazon": {
        "display_name": "Amazon",
        "providers": ["rainforest", "scrapingbee"],
        "default_provider": "rainforest",
        "env_key": "RAINFOREST_API_KEY",
    },
    "google_books": {
        "display_name": "Google Books",
        "providers": ["google_books"],
        "default_provider": "google_books",
        "env_key": "GOOGLE_BOOKS_API_KEY",
    },
    "barnes_noble": {
        "display_name": "Barnes & Noble",
        "providers": ["scrapingbee"],
        "default_provider": "scrapingbee",
        "env_key": None,
    },
    "apple_books": {
        "display_name": "Apple Books",
        "providers": ["scrapingbee"],
        "default_provider": "scrapingbee",
        "env_key": None,
    },
    "kobo": {
        "display_name": "Kobo",
        "providers": ["scrapingbee"],
        "default_provider": "scrapingbee",
        "env_key": None,
    },
    "thriftbooks": {
        "display_name": "ThriftBooks",
        "providers": ["scrapingbee"],
        "default_provider": "scrapingbee",
        "env_key": None,
    },
    "abebooks": {
        "display_name": "AbeBooks",
        "providers": ["scrapingbee"],
        "default_provider": "scrapingbee",
        "env_key": None,
    },
    "bookshop": {
        "display_name": "Bookshop.org",
        "providers": ["scrapingbee"],
        "default_provider": "scrapingbee",
        "env_key": None,
    },
    "goodreads": {
        "display_name": "Goodreads",
        "providers": [],
        "default_provider": None,
        "env_key": None,
    },
    "open_library": {
        "display_name": "Open Library",
        "providers": [],
        "default_provider": None,
        "env_key": None,
    },
}


@dataclass
class EffectiveCredential:
    """Resolved credential ready for outbound API calls."""

    platform_id: str
    provider: str
    api_key: str
    api_secret: str | None
    source: Literal["superadmin", "environment"]


@dataclass
class PlatformApiCredentialView:
    """Masked credential for SuperAdmin API responses."""

    platform_id: str
    display_name: str
    provider: str | None
    api_key_hint: str | None
    has_secret: bool
    is_enabled: bool
    source: str
    available_providers: list[str]


def provider_scope_id(provider: str) -> str:
    """Canonical DB key for a shared provider credential."""
    return f"{PROVIDER_SCOPE_PREFIX}{provider}"


def is_provider_scope(platform_id: str) -> bool:
    """True when platform_id is a shared provider scope row."""
    return platform_id.startswith(PROVIDER_SCOPE_PREFIX)


def provider_id_from_scope(scope_id: str) -> str:
    """Extract provider id from @provider:scrapingbee."""
    return scope_id.removeprefix(PROVIDER_SCOPE_PREFIX)


def platforms_for_provider(provider: str) -> list[str]:
    """Retail platform IDs that can use this provider."""
    return [
        pid for pid, meta in PLATFORM_API_META.items()
        if provider in meta.get("providers", [])
    ]


def _env_credential_for_provider(provider: str) -> EffectiveCredential | None:
    """First matching env fallback for any platform using this provider."""
    for pid in platforms_for_provider(provider):
        meta = PLATFORM_API_META.get(pid, {})
        if meta.get("default_provider") != provider:
            continue
        cred = _env_credential(pid)
        if cred:
            return cred
    return None


def _env_credential(platform_id: str) -> EffectiveCredential | None:
    """Build credential from environment when configured."""
    meta = PLATFORM_API_META.get(platform_id)
    if not meta or not meta.get("env_key"):
        return None
    env_name = meta["env_key"]
    cfg = get_settings()
    raw = (getattr(cfg, env_name, None) or "").strip()
    if not raw:
        return None
    provider = meta["default_provider"]
    if not provider:
        return None
    return EffectiveCredential(
        platform_id=platform_id,
        provider=provider,
        api_key=raw,
        api_secret=None,
        source="environment",
    )


def _row_to_effective(row) -> EffectiveCredential:
    """Decrypt DB row into runtime credential."""
    secret = None
    if row.api_secret_encrypted:
        secret = decrypt_secret(row.api_secret_encrypted)
    return EffectiveCredential(
        platform_id=row.platform_id,
        provider=row.provider,
        api_key=decrypt_secret(row.api_key_encrypted),
        api_secret=secret,
        source="superadmin",
    )


async def _cache_get(redis, platform_id: str) -> EffectiveCredential | None:
    if redis is None:
        return None
    try:
        raw = await redis.get(_CACHE_PREFIX + platform_id)
        if not raw:
            return None
        data = json.loads(raw)
        if data.get("empty"):
            return None
        return EffectiveCredential(**data)
    except Exception:
        return None


async def _cache_set(redis, platform_id: str, cred: EffectiveCredential | None) -> None:
    if redis is None:
        return
    try:
        key = _CACHE_PREFIX + platform_id
        if cred is None:
            await redis.set(key, json.dumps({"empty": True}), ex=_CACHE_TTL)
        else:
            await redis.set(
                key,
                json.dumps({
                    "platform_id": cred.platform_id,
                    "provider": cred.provider,
                    "api_key": cred.api_key,
                    "api_secret": cred.api_secret,
                    "source": cred.source,
                }),
                ex=_CACHE_TTL,
            )
    except Exception:
        pass


async def invalidate_credential_cache(redis, platform_id: str | None = None) -> None:
    """Clear cached credentials after SuperAdmin writes (R-179)."""
    if redis is None:
        return
    try:
        if platform_id:
            await redis.delete(_CACHE_PREFIX + platform_id)
            if is_provider_scope(platform_id):
                for pid in platforms_for_provider(provider_id_from_scope(platform_id)):
                    await redis.delete(_CACHE_PREFIX + pid)
        else:
            keys = await redis.keys(_CACHE_PREFIX + "*")
            if keys:
                await redis.delete(*keys)
    except Exception:
        pass


async def resolve_platform_credential(
    platform_id: str,
    *,
    db: AsyncSession | None = None,
    redis=None,
) -> EffectiveCredential | None:
    """Resolve credential: platform override → provider scope → environment → None."""
    cached = await _cache_get(redis, platform_id)
    if cached is not None or (redis and await redis.get(_CACHE_PREFIX + platform_id)):
        return cached

    cred: EffectiveCredential | None = None
    if db is not None:
        repo = PlatformApiCredentialRepository(db)
        row = await repo.get_enabled(platform_id)
        if row and not is_provider_scope(row.platform_id):
            cred = _row_to_effective(row)
        if cred is None:
            meta = PLATFORM_API_META.get(platform_id)
            default_provider = meta.get("default_provider") if meta else None
            if default_provider:
                scope_row = await repo.get_enabled(provider_scope_id(default_provider))
                if scope_row:
                    cred = _row_to_effective(scope_row)

    if cred is None:
        cred = _env_credential(platform_id)

    await _cache_set(redis, platform_id, cred)
    return cred


async def load_all_credentials(
    db: AsyncSession | None = None,
    redis=None,
) -> dict[str, EffectiveCredential]:
    """Load credentials for all scored platforms."""
    from app.services.platform_presence import SCORED_PLATFORMS

    out: dict[str, EffectiveCredential] = {}
    for cfg in SCORED_PLATFORMS:
        pid = cfg["platform_id"]
        cred = await resolve_platform_credential(pid, db=db, redis=redis)
        if cred:
            out[pid] = cred
    return out


def _extract_asin(url: str) -> str | None:
    m = re.search(r"/(?:dp|gp/product|gp/aw/d)/([A-Z0-9]{10})", url, re.I)
    return m.group(1).upper() if m else None


async def _fetch_rainforest_price(
    client: httpx.AsyncClient,
    cred: EffectiveCredential,
    listing_url: str,
) -> PlatformPrice | None:
    asin = _extract_asin(listing_url)
    if not asin:
        return None
    try:
        r = await client.get(
            "https://api.rainforestapi.com/request",
            params={
                "api_key": cred.api_key,
                "type": "product",
                "amazon_domain": "amazon.com",
                "asin": asin,
            },
            timeout=15.0,
        )
        if r.status_code != 200:
            return None
        data = r.json()
        winner = data.get("product", {}).get("buybox_winner") or data.get("buybox_winner") or {}
        price = winner.get("price") or {}
        amount = price.get("value") or price.get("raw")
        if amount is None:
            return None
        amount_f = float(str(amount).replace(",", "").replace("$", ""))
        currency = str(price.get("currency") or "USD").upper()
        return PlatformPrice(
            formatted=format_price(amount_f, currency),
            amount=amount_f,
            currency=currency,
            source="api:rainforest",
        )
    except Exception as exc:
        logger.debug("Rainforest price failed", error=str(exc)[:120])
    return None


async def _fetch_scrapingbee_html(
    client: httpx.AsyncClient,
    cred: EffectiveCredential,
    listing_url: str,
) -> str | None:
    try:
        r = await client.get(
            "https://app.scrapingbee.com/api/v1",
            params={
                "api_key": cred.api_key,
                "url": listing_url,
                "premium_proxy": "true",
                "country_code": "us",
            },
            timeout=30.0,
        )
        if r.status_code == 200 and len(r.text) > 500:
            return r.text
    except Exception as exc:
        logger.debug("ScrapingBee fetch failed", error=str(exc)[:120])
    return None


async def _fetch_google_books_with_key(
    client: httpx.AsyncClient,
    cred: EffectiveCredential,
    listing_url: str,
    ctx: PriceContext | None,
) -> PlatformPrice | None:
    from app.services.platform_pricing import fetch_google_books_price

    volume_m = re.search(r"[?&]id=([^&]+)", listing_url)
    isbn = (ctx.isbn13 if ctx else None) or ""
    try:
        if volume_m:
            r = await client.get(
                f"https://www.googleapis.com/books/v1/volumes/{volume_m.group(1)}",
                params={"key": cred.api_key, "country": "US"},
                timeout=10.0,
            )
            if r.status_code == 200:
                from app.services.platform_pricing import _sale_info_price
                hit = _sale_info_price(r.json().get("saleInfo") or {})
                if hit:
                    hit.source = "api:google_books"
                    return hit
        if isbn:
            r = await client.get(
                "https://www.googleapis.com/books/v1/volumes",
                params={
                    "key": cred.api_key,
                    "q": f"isbn:{isbn}",
                    "maxResults": "5",
                    "country": "US",
                },
                timeout=10.0,
            )
            if r.status_code == 200:
                from app.services.platform_pricing import _sale_info_price
                for item in r.json().get("items", []):
                    hit = _sale_info_price(item.get("saleInfo") or {})
                    if hit:
                        hit.source = "api:google_books"
                        return hit
    except Exception as exc:
        logger.debug("Google Books keyed fetch failed", error=str(exc)[:120])

    return await fetch_google_books_price(client, listing_url, ctx)


async def fetch_price_via_credential(
    client: httpx.AsyncClient,
    cred: EffectiveCredential,
    platform_id: str,
    listing_url: str,
    ctx: PriceContext | None = None,
) -> PlatformPrice | None:
    """Fetch price using SuperAdmin/env credential (priority tier 1)."""
    if cred.provider == "rainforest":
        if platform_id != "amazon":
            return None
        return await _fetch_rainforest_price(client, cred, listing_url)

    if cred.provider == "google_books":
        if platform_id != "google_books":
            return None
        return await _fetch_google_books_with_key(client, cred, listing_url, ctx)

    if cred.provider == "scrapingbee":
        html = await _fetch_scrapingbee_html(client, cred, listing_url)
        if html:
            hit = extract_price_from_html(platform_id, html)
            if hit:
                hit.source = "api:scrapingbee"
            return hit

    return None


def build_credential_views(rows: list) -> list[PlatformApiCredentialView]:
    """Build SuperAdmin list view for all platforms."""
    by_id = {r.platform_id: r for r in rows}
    views: list[PlatformApiCredentialView] = []
    for pid, meta in PLATFORM_API_META.items():
        row = by_id.get(pid)
        env_cred = _env_credential(pid)
        if row:
            key_hint = mask_secret(decrypt_secret(row.api_key_encrypted))
            views.append(PlatformApiCredentialView(
                platform_id=pid,
                display_name=meta["display_name"],
                provider=row.provider,
                api_key_hint=key_hint,
                has_secret=bool(row.api_secret_encrypted),
                is_enabled=row.is_enabled,
                source="superadmin",
                available_providers=list(meta["providers"]),
            ))
        elif env_cred:
            views.append(PlatformApiCredentialView(
                platform_id=pid,
                display_name=meta["display_name"],
                provider=env_cred.provider,
                api_key_hint=mask_secret(env_cred.api_key),
                has_secret=False,
                is_enabled=True,
                source="environment",
                available_providers=list(meta["providers"]),
            ))
        else:
            views.append(PlatformApiCredentialView(
                platform_id=pid,
                display_name=meta["display_name"],
                provider=meta.get("default_provider"),
                api_key_hint=None,
                has_secret=False,
                is_enabled=False,
                source="none",
                available_providers=list(meta["providers"]),
            ))
    return views


def build_provider_views(rows: list) -> dict:
    """Build SuperAdmin provider-scoped credential views (one key per provider)."""
    by_id = {r.platform_id: r for r in rows}
    providers_out: list[dict] = []

    for provider_id, info in PROVIDER_DEFINITIONS.items():
        scope_id = provider_scope_id(provider_id)
        row = by_id.get(scope_id)
        env_cred = _env_credential_for_provider(provider_id)
        platform_labels = [
            {
                "platform_id": pid,
                "display_name": PLATFORM_API_META[pid]["display_name"],
            }
            for pid in platforms_for_provider(provider_id)
        ]

        if row and row.is_enabled:
            source = "superadmin"
            key_hint = mask_secret(decrypt_secret(row.api_key_encrypted))
            is_enabled = True
        elif env_cred:
            source = "environment"
            key_hint = mask_secret(env_cred.api_key)
            is_enabled = True
        else:
            source = "none"
            key_hint = None
            is_enabled = False

        providers_out.append({
            "provider_id": provider_id,
            "scope_id": scope_id,
            "display_name": info["display_name"],
            "description": info["description"],
            "platforms": platform_labels,
            "api_key_hint": key_hint,
            "is_enabled": is_enabled,
            "source": source,
        })

    without_api = [
        {
            "platform_id": pid,
            "display_name": meta["display_name"],
        }
        for pid, meta in PLATFORM_API_META.items()
        if not meta.get("providers")
    ]
    return {"providers": providers_out, "platforms_without_api": without_api}