api/app/main.py

from contextlib import asynccontextmanager
import logging

from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware

from app.api.middleware.rate_limit import TierRateLimitMiddleware
from app.api.middleware.sanitizer import SanitizerMiddleware
from app.config.settings import settings

logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s %(levelname)s %(name)s: %(message)s",
)
logging.getLogger("sqlalchemy.engine").setLevel(logging.WARNING)
logging.getLogger("sqlalchemy.pool").setLevel(logging.WARNING)


async def _memory_audit_cron_tick() -> None:
    """Weekly cron: contradiction scan + label canonicalization for all users (Phase 7)."""
    import logging  # noqa: PLC0415
    _log = logging.getLogger(__name__)
    _log.info("memory audit cron tick: starting")
    try:
        from app.db import async_session  # noqa: PLC0415
        from app.core.memory_maintenance import audit_memory  # noqa: PLC0415
        from app.models import User  # noqa: PLC0415
        from sqlalchemy import select  # noqa: PLC0415

        async with async_session() as db:
            result = await db.execute(select(User.id))
            user_ids: list[str] = list(result.scalars().all())

        for uid in user_ids:
            try:
                async with async_session() as db:
                    await audit_memory(db, uid)
            except Exception as exc:
                _log.warning("memory audit cron tick: audit_memory failed user=%s: %s", uid, exc)

        _log.info("memory audit cron tick: done users=%d", len(user_ids))
    except Exception as exc:
        _log.warning("memory audit cron tick: failed: %s", exc)


async def _memory_cron_tick() -> None:
    """Hourly cron: drain Free-tier extraction queue + mine proactive patterns for Power+ users."""
    import logging  # noqa: PLC0415
    _log = logging.getLogger(__name__)
    _log.info("memory cron tick: starting")
    try:
        from app.db import async_session  # noqa: PLC0415
        from app.core.memory_maintenance import drain_extraction_queue, mine_proactive_patterns  # noqa: PLC0415
        from app.billing.tier_manager import tier_manager  # noqa: PLC0415
        from app.models import User  # noqa: PLC0415
        from sqlalchemy import select  # noqa: PLC0415

        async with async_session() as db:
            await drain_extraction_queue(db)

        # mine proactive patterns for every Power+ user
        async with async_session() as db:
            result = await db.execute(select(User.id))
            user_ids: list[str] = list(result.scalars().all())

        for uid in user_ids:
            try:
                async with async_session() as db:
                    tier = await tier_manager.get_tier(uid, db)
                    if tier_manager.check_feature(tier, "proactive_mining"):
                        await mine_proactive_patterns(db, uid)
            except Exception as exc:
                _log.warning("memory cron tick: mine_proactive_patterns failed user=%s: %s", uid, exc)

        _log.info("memory cron tick: done users=%d", len(user_ids))
    except Exception as exc:
        _log.warning("memory cron tick: failed: %s", exc)


async def _scout_cron_tick() -> None:
    """Every-15-min cron: poll enabled cloud scouts (cron-fallback; push is primary).

    Skips any scout whose ``last_run_at`` is within the last 5 minutes so
    a push notification and the fallback cron don't double-fire within the
    same window.
    """
    import logging  # noqa: PLC0415
    import uuid  # noqa: PLC0415
    from datetime import datetime, timezone  # noqa: PLC0415

    _log = logging.getLogger(__name__)
    _log.info("scout cron tick: starting")
    try:
        from app.db import async_session  # noqa: PLC0415
        from app.models import CloudScoutConfig  # noqa: PLC0415
        from app.scouts.engine import ScoutEngine  # noqa: PLC0415
        from sqlalchemy import select  # noqa: PLC0415

        async with async_session() as session:
            scouts = (await session.execute(
                select(CloudScoutConfig).where(CloudScoutConfig.enabled == True)  # noqa: E712
            )).scalars().all()

        engine = ScoutEngine()
        triggered = 0
        for scout in scouts:
            # Rate-limit guard: push is primary; skip if ran within 5 minutes.
            if scout.last_run_at:
                elapsed = (datetime.now(tz=timezone.utc) - scout.last_run_at).total_seconds()
                if elapsed < 300:
                    continue
            try:
                await engine.trigger_scout(uuid.UUID(str(scout.id)))
                triggered += 1
            except Exception as exc:
                _log.warning("scout cron tick: trigger failed scout=%s: %s", scout.id, exc)

        _log.info("scout cron tick: done triggered=%d total=%d", triggered, len(scouts))
    except Exception as exc:
        _log.warning("scout cron tick: failed: %s", exc)


async def _scout_watch_renewal_tick() -> None:
    """Every-24-hour cron: re-issue Gmail users.watch for scouts expiring within 24h.

    Handles missing or misconfigured connectors gracefully — logs and continues.
    """
    import logging  # noqa: PLC0415
    from datetime import datetime, timedelta, timezone  # noqa: PLC0415

    _log = logging.getLogger(__name__)
    _log.info("scout watch renewal tick: starting")
    try:
        from app.db import async_session  # noqa: PLC0415
        from app.models import CloudScoutConfig  # noqa: PLC0415
        from app.scouts.connectors.registry import get_connector  # noqa: PLC0415
        from sqlalchemy import select  # noqa: PLC0415

        threshold = datetime.now(tz=timezone.utc) + timedelta(hours=24)
        renewed = 0
        async with async_session() as session:
            scouts = (await session.execute(
                select(CloudScoutConfig).where(
                    CloudScoutConfig.enabled == True,  # noqa: E712
                    CloudScoutConfig.provider == "gmail",
                    CloudScoutConfig.gmail_watch_expires_at <= threshold,
                )
            )).scalars().all()

            for scout in scouts:
                try:
                    connector = get_connector("gmail")
                    await connector.renew_watch(scout)
                    renewed += 1
                except Exception:
                    _log.exception("scout watch renewal tick: renew failed scout=%s", scout.id)

            await session.commit()

        _log.info("scout watch renewal tick: done renewed=%d", renewed)
    except Exception as exc:
        _log.warning("scout watch renewal tick: failed: %s", exc)


@asynccontextmanager
async def lifespan(app: FastAPI):
    # Startup: register source connectors.
    from app.scouts.connectors.gmail import GmailConnector  # noqa: PLC0415
    from app.scouts.connectors.registry import register_connector  # noqa: PLC0415
    register_connector(GmailConnector())

    # Startup: ensure agent tool modules are loaded.
    import app.agents  # noqa: F401

    scheduler = None
    if settings.SCHEDULER_ENABLED:
        from apscheduler.schedulers.asyncio import AsyncIOScheduler  # noqa: PLC0415

        scheduler = AsyncIOScheduler()
        scheduler.add_job(_memory_cron_tick, "interval", hours=1, id="memory_cron")
        scheduler.add_job(_memory_audit_cron_tick, "interval", weeks=1, id="memory_audit_cron")
        scheduler.add_job(
            _scout_cron_tick, "interval", minutes=15,
            id="scout_cron_tick", replace_existing=True,
        )
        scheduler.add_job(
            _scout_watch_renewal_tick, "interval", hours=24,
            id="scout_watch_renewal_tick", replace_existing=True,
        )
        scheduler.start()
        logging.getLogger(__name__).info("memory cron scheduler started (interval=1h)")

    yield

    if scheduler is not None:
        scheduler.shutdown(wait=False)

    # Shutdown: dispose SQLAlchemy connection pool
    from app.db import engine
    await engine.dispose()


def create_app() -> FastAPI:
    app = FastAPI(
        title="AdiuvAI Cloud API",
        version="0.1.0",
        docs_url="/docs" if settings.ENV == "dev" else None,
        redoc_url=None,
        lifespan=lifespan,
    )

    app.add_middleware(
        CORSMiddleware,
        allow_origins=settings.CORS_ORIGINS,
        allow_credentials=True,
        allow_methods=["*"],
        allow_headers=["*"],
    )
    # Middleware stack (Starlette inserts at position 0, so last-added = outermost).
    # Request flow:  TierRateLimit → Sanitizer → CORS → Router
    # Response flow: Router → CORS → Sanitizer → TierRateLimit
    app.add_middleware(SanitizerMiddleware)
    app.add_middleware(TierRateLimitMiddleware)

    from app.api.routes import scouts, auth, billing, chat, device_ws, memory, scout_webhooks

    app.include_router(auth.router,           prefix="/api/v1")
    app.include_router(chat.router,           prefix="/api/v1")
    app.include_router(billing.router,        prefix="/api/v1")
    app.include_router(scouts.router,         prefix="/api/v1")
    app.include_router(scout_webhooks.router, prefix="/api/v1")
    app.include_router(device_ws.router,      prefix="/api/v1")
    app.include_router(memory.router,         prefix="/api/v1")

    @app.get("/api/v1/health", tags=["health"])
    async def health() -> dict:
        return {"status": "ok", "version": app.version}

    return app


app = create_app()