myYouTube/backend/app/services/google_oauth.py

import logging
import os
import threading
from datetime import datetime, timedelta, timezone

# Google frequently echoes back a scope string that's a superset/reordering
# of what we requested (e.g. we ask for "youtube", Google's token response
# also lists "youtube.readonly" since it's implied). oauthlib does an exact
# string comparison by default and raises on any mismatch -- this disables
# that overly strict check. Must be set before requests_oauthlib reads it.
os.environ.setdefault("OAUTHLIB_RELAX_TOKEN_SCOPE", "1")

import httpx
from google.auth.transport.requests import Request as GoogleAuthRequest
from google.oauth2.credentials import Credentials
from google_auth_oauthlib.flow import Flow
from sqlalchemy.orm import Session

from app.config import settings
from app.core.crypto import decrypt_token, encrypt_token
from app.models.oauth_credentials import SINGLETON_ID, OAuthCredentials

logger = logging.getLogger(__name__)

SCOPES = [
    # Full read/write scope, not just youtube.readonly: unsubscribing from a
    # channel (subscriptions.delete) requires write access. Deliberate
    # deviation from the original "minimal scope" TZ recommendation, per
    # explicit user request. Existing stored refresh tokens were granted
    # under the old readonly-only scope and won't cover this -- users must
    # reconnect once for this to take effect.
    "https://www.googleapis.com/auth/youtube",
    "openid",
    "https://www.googleapis.com/auth/userinfo.email",
    "https://www.googleapis.com/auth/userinfo.profile",
]

# Refresh proactively: never hand out a token that could die mid-request.
_ACCESS_TOKEN_REFRESH_BUFFER_SECONDS = 60

# Module-level access-token cache. Google access tokens live ~1h; previously
# every get_credentials() call (sync, unsubscribe, ...) paid for a full token
# exchange with Google. The lock keeps concurrent callers sharing one refresh
# instead of racing each other to the token endpoint.
_credentials_lock = threading.Lock()
_cached_credentials: Credentials | None = None


class OAuthNotConnected(Exception):
    pass


def _client_config() -> dict:
    return {
        "web": {
            "client_id": settings.google_client_id,
            "client_secret": settings.google_client_secret,
            "auth_uri": settings.google_auth_uri,
            "token_uri": settings.google_token_uri,
            "redirect_uris": [settings.google_redirect_uri],
        }
    }


def _build_flow(state: str | None = None) -> Flow:
    return Flow.from_client_config(
        _client_config(),
        scopes=SCOPES,
        state=state,
        redirect_uri=settings.google_redirect_uri,
    )


def build_authorization_url() -> tuple[str, str]:
    flow = _build_flow()
    auth_url, state = flow.authorization_url(
        access_type="offline",
        prompt="consent",
        include_granted_scopes="true",
    )
    return auth_url, state


def exchange_code(code: str, state: str) -> Credentials:
    flow = _build_flow(state=state)
    flow.fetch_token(code=code)
    return flow.credentials


def fetch_userinfo(access_token: str) -> dict:
    response = httpx.get(
        settings.google_userinfo_uri,
        headers={"Authorization": f"Bearer {access_token}"},
        timeout=settings.metube_request_timeout_seconds,
    )
    response.raise_for_status()
    return response.json()


def revoke_token(token: str) -> None:
    try:
        httpx.post(settings.google_revoke_uri, params={"token": token}, timeout=10)
    except Exception:
        logger.warning("Failed to revoke Google token", exc_info=True)


def store_credentials(db: Session, google_email: str, credentials: Credentials) -> None:
    global _cached_credentials
    encrypted = encrypt_token(credentials.refresh_token)

    row = db.get(OAuthCredentials, SINGLETON_ID)
    if row is None:
        row = OAuthCredentials(id=SINGLETON_ID, google_email=google_email, encrypted_refresh_token=encrypted)
        db.add(row)
    else:
        row.google_email = google_email
        row.encrypted_refresh_token = encrypted
    db.commit()

    # The exchanged credentials carry a fresh access token -- reuse them so
    # the immediately following syncs don't pay for a second Google request.
    with _credentials_lock:
        _cached_credentials = credentials


def clear_credentials(db: Session) -> None:
    global _cached_credentials
    row = db.get(OAuthCredentials, SINGLETON_ID)
    if row is not None:
        db.delete(row)
        db.commit()
    with _credentials_lock:
        _cached_credentials = None


def is_connected(db: Session) -> bool:
    return db.get(OAuthCredentials, SINGLETON_ID) is not None


def get_connected_email(db: Session) -> str | None:
    row = db.get(OAuthCredentials, SINGLETON_ID)
    return row.google_email if row else None


def _token_missing_or_expiring(credentials: Credentials) -> bool:
    if not credentials.token:
        return True
    expiry = credentials.expiry
    if expiry is None:
        # Unknown expiry -- be conservative and refresh.
        return True
    if expiry.tzinfo is None:
        # google-auth reports expiry as a naive UTC datetime.
        expiry = expiry.replace(tzinfo=timezone.utc)
    return expiry <= datetime.now(timezone.utc) + timedelta(seconds=_ACCESS_TOKEN_REFRESH_BUFFER_SECONDS)


def get_credentials(db: Session) -> Credentials:
    global _cached_credentials
    row = db.get(OAuthCredentials, SINGLETON_ID)
    if row is None:
        raise OAuthNotConnected("Google account is not connected")

    refresh_token = decrypt_token(row.encrypted_refresh_token)

    with _credentials_lock:
        credentials = _cached_credentials
        if credentials is None or credentials.refresh_token != refresh_token:
            # No cached token yet, or the account was reconnected with a new
            # refresh token. The refresh below does the initial exchange.
            credentials = Credentials(
                token=None,
                refresh_token=refresh_token,
                token_uri=settings.google_token_uri,
                client_id=settings.google_client_id,
                client_secret=settings.google_client_secret,
                scopes=SCOPES,
            )
            _cached_credentials = credentials

        if _token_missing_or_expiring(credentials):
            try:
                credentials.refresh(GoogleAuthRequest())
            except Exception:
                # Don't keep a broken cached object behind.
                _cached_credentials = None
                raise

        return credentials