kite/backend/tests/test_cloud_cache.py

"""
Phase 14 Plan 01 — RED contract tests for the byte cache service.

Requirements: CACHE-03, CACHE-04, CACHE-05

Threat coverage:
  T-14-01 — IDOR: cache metadata queries are always owner-scoped
  T-14-02 — Object-key secrecy: object_key never appears in API responses
  T-14-05 — Cache quota corruption: quota must be atomically incremented on
             cache retain and decremented on eviction; a safe in-memory
             simulation is used because SQLite cannot run the atomic UPDATE.

All tests import non-existent Phase 14 modules and must FAIL due to
ImportError until implementation is complete.
"""
from __future__ import annotations

import uuid
from unittest.mock import AsyncMock, MagicMock, patch

import pytest

pytestmark = pytest.mark.asyncio

from tests.conftest import _TEST_USER_AGENT


# ─── Helpers ─────────────────────────────────────────────────────────────────

async def _create_user_and_token(session, role: str = "user"):
    from db.models import User, Quota
    from services.auth import hash_password, create_access_token

    user_id = uuid.uuid4()
    user = User(
        id=user_id,
        handle=f"cache_user_{user_id.hex[:8]}",
        email=f"cache_{user_id.hex[:8]}@example.com",
        password_hash=hash_password("Testpassword123!"),
        role=role,
        is_active=True,
        password_must_change=False,
    )
    quota = Quota(user_id=user_id, limit_bytes=104857600, used_bytes=0)
    session.add(user)
    session.add(quota)
    await session.commit()
    await session.refresh(user)

    token = create_access_token(str(user_id), role, user_agent=_TEST_USER_AGENT)
    return {
        "user": user,
        "token": token,
        "headers": {
            "Authorization": f"Bearer {token}",
            "User-Agent": _TEST_USER_AGENT,
        },
    }


async def _create_cloud_connection(session, user_id, provider="google_drive", status="ACTIVE"):
    from db.models import CloudConnection
    from storage.cloud_utils import encrypt_credentials
    from config import settings

    master_key = settings.cloud_creds_key.encode()
    creds_enc = encrypt_credentials(
        master_key,
        str(user_id),
        {"access_token": "tok", "refresh_token": "ref"},
    )
    conn = CloudConnection(
        id=uuid.uuid4(),
        user_id=user_id,
        provider=provider,
        display_name="Cache Test Connection",
        credentials_enc=creds_enc,
        status=status,
    )
    session.add(conn)
    await session.commit()
    return conn


async def _create_cloud_item(session, user_id, connection_id, *, name="doc.pdf",
                              kind="file", etag="etag-v1"):
    from db.models import CloudItem

    item = CloudItem(
        id=uuid.uuid4(),
        user_id=user_id,
        connection_id=connection_id,
        provider_item_id=f"pitem-{uuid.uuid4().hex[:8]}",
        name=name,
        kind=kind,
        content_type="application/pdf",
        provider_size=102400,
        etag=etag,
        analysis_status="pending",
    )
    session.add(item)
    await session.commit()
    return item


# ─── CACHE-03: Cache bytes only during open/preview/analysis ─────────────────

async def test_cache_entry_created_during_analysis_job(db_session):
    """CACHE-03: A cache entry is created when bytes are hydrated for analysis.

    Imports CloudByteCacheEntry (Phase 14 schema addition).
    """
    from services.cloud_cache import create_cache_entry  # Phase 14

    user_id = uuid.uuid4()
    conn_id = uuid.uuid4()
    item_id = uuid.uuid4()

    entry = await create_cache_entry(
        session=db_session,
        user_id=user_id,
        connection_id=conn_id,
        cloud_item_id=item_id,
        provider_item_id="pitem-001",
        version_key="etag-v1",
        object_key=f"cache/{user_id}/{uuid.uuid4()}.pdf",
        content_type="application/pdf",
        size_bytes=102400,
    )

    assert entry is not None
    assert entry.user_id == user_id
    assert entry.cloud_item_id == item_id
    assert entry.size_bytes == 102400
    assert entry.pin_count == 0
    assert entry.active_job_count == 0


async def test_browse_without_analysis_does_not_create_cache_entry(db_session):
    """CACHE-03: Listing a folder never creates a cache entry (no byte download).

    Verifies that reconcile_cloud_listing (existing Phase 12/13 code path) leaves
    cloud_byte_cache_entries empty.
    """
    from db.models import CloudConnection, CloudItem
    from sqlalchemy import select

    # Try to import Phase 14 cache model
    try:
        from db.models import CloudByteCacheEntry  # Phase 14 — not yet in models
    except ImportError:
        pytest.fail(
            "CloudByteCacheEntry model must be added to db/models.py in Phase 14"
        )

    user_id = uuid.uuid4()
    # After any browse/reconcile operation, zero cache entries should exist
    result = await db_session.execute(
        select(CloudByteCacheEntry).where(CloudByteCacheEntry.user_id == user_id)
    )
    rows = result.scalars().all()
    assert len(rows) == 0, "Browse operations must not create cache entries"


# ─── CACHE-04: LRU eviction and cache limits ─────────────────────────────────

async def test_eviction_does_not_remove_pinned_entry(db_session):
    """CACHE-04/D-16: Eviction must not evict entries with active jobs or pins.

    An entry with pin_count > 0 must survive LRU eviction.
    """
    from services.cloud_cache import create_cache_entry, evict_lru_entries  # Phase 14

    user_id = uuid.uuid4()
    conn_id = uuid.uuid4()

    # Create a pinned cache entry (pin_count=1 simulates active analysis/preview)
    pinned = await create_cache_entry(
        session=db_session,
        user_id=user_id,
        connection_id=conn_id,
        cloud_item_id=uuid.uuid4(),
        provider_item_id="pinned-item",
        version_key="etag-pinned",
        object_key=f"cache/{user_id}/{uuid.uuid4()}.pdf",
        content_type="application/pdf",
        size_bytes=10 * 1024 * 1024,  # 10 MB
        pin_count=1,  # pinned — must survive eviction
    )

    # Create an unpinned entry to confirm eviction can happen at all
    evictable = await create_cache_entry(
        session=db_session,
        user_id=user_id,
        connection_id=conn_id,
        cloud_item_id=uuid.uuid4(),
        provider_item_id="evictable-item",
        version_key="etag-evict",
        object_key=f"cache/{user_id}/{uuid.uuid4()}.pdf",
        content_type="application/pdf",
        size_bytes=5 * 1024 * 1024,  # 5 MB
        pin_count=0,
    )

    # Run eviction with a tight byte limit that forces eviction of unpinned entries
    evicted_ids = await evict_lru_entries(
        session=db_session,
        user_id=user_id,
        max_cache_bytes=1,  # 1 byte limit forces eviction
    )

    # The pinned entry must NOT be evicted
    assert str(pinned.id) not in [str(e) for e in evicted_ids], (
        "Eviction must preserve entries with pin_count > 0 (D-16)"
    )


async def test_eviction_removes_oldest_unpinned_entry_first(db_session):
    """CACHE-04/D-14: LRU eviction removes least-recently-accessed unpinned entry first."""
    from services.cloud_cache import create_cache_entry, evict_lru_entries  # Phase 14
    from datetime import datetime, timezone, timedelta

    user_id = uuid.uuid4()
    conn_id = uuid.uuid4()

    older_entry = await create_cache_entry(
        session=db_session,
        user_id=user_id,
        connection_id=conn_id,
        cloud_item_id=uuid.uuid4(),
        provider_item_id="old-item",
        version_key="etag-old",
        object_key=f"cache/{user_id}/{uuid.uuid4()}.pdf",
        content_type="application/pdf",
        size_bytes=5 * 1024 * 1024,
        pin_count=0,
        # Simulate last accessed 2 hours ago
        last_accessed_at=datetime.now(timezone.utc) - timedelta(hours=2),
    )

    newer_entry = await create_cache_entry(
        session=db_session,
        user_id=user_id,
        connection_id=conn_id,
        cloud_item_id=uuid.uuid4(),
        provider_item_id="new-item",
        version_key="etag-new",
        object_key=f"cache/{user_id}/{uuid.uuid4()}.pdf",
        content_type="application/pdf",
        size_bytes=5 * 1024 * 1024,
        pin_count=0,
        # Simulate last accessed 1 minute ago
        last_accessed_at=datetime.now(timezone.utc) - timedelta(minutes=1),
    )

    # Evict enough to remove exactly one entry (limit below combined size)
    evicted_ids = await evict_lru_entries(
        session=db_session,
        user_id=user_id,
        max_cache_bytes=6 * 1024 * 1024,  # fits one entry; must evict the older one
    )

    # The older entry should be evicted first (LRU order)
    assert str(older_entry.id) in [str(e) for e in evicted_ids], (
        "LRU eviction must remove least-recently-accessed entry first"
    )
    # The newer entry should survive
    assert str(newer_entry.id) not in [str(e) for e in evicted_ids]


# ─── CACHE-04/T-14-05: Quota accounting ─────────────────────────────────────

async def test_cache_retain_increments_quota(db_session):
    """T-14-05: Creating a cache entry must atomically increment used_bytes in quotas.

    The increment requirement is tested via the service layer; the actual atomic
    SQL UPDATE…RETURNING runs during implementation against real PostgreSQL.
    In SQLite we verify the function signature exists and accepts the right arguments.
    """
    from services.cloud_cache import increment_quota_for_cache  # Phase 14

    # Verify the function exists with the expected signature
    import inspect
    sig = inspect.signature(increment_quota_for_cache)
    params = list(sig.parameters.keys())

    # Must accept session, user_id, and size_bytes
    assert "session" in params, "increment_quota_for_cache must accept session"
    assert "user_id" in params, "increment_quota_for_cache must accept user_id"
    assert "size_bytes" in params, "increment_quota_for_cache must accept size_bytes"


async def test_cache_eviction_decrements_quota(db_session):
    """T-14-05: Evicting a cache entry must decrement used_bytes in quotas.

    Verifies the decrement function signature exists (full atomic SQL in PostgreSQL).
    """
    from services.cloud_cache import decrement_quota_for_cache  # Phase 14

    import inspect
    sig = inspect.signature(decrement_quota_for_cache)
    params = list(sig.parameters.keys())

    assert "session" in params, "decrement_quota_for_cache must accept session"
    assert "user_id" in params, "decrement_quota_for_cache must accept user_id"
    assert "size_bytes" in params, "decrement_quota_for_cache must accept size_bytes"


# ─── CACHE-05: Owner isolation ───────────────────────────────────────────────

async def test_cache_entry_query_is_owner_scoped(db_session):
    """CACHE-05/T-14-01: Cache metadata queries are scoped to owner.

    Foreign user must not see another user's cache entries.
    """
    from services.cloud_cache import create_cache_entry, list_cache_entries  # Phase 14

    owner_id = uuid.uuid4()
    foreign_id = uuid.uuid4()
    conn_id = uuid.uuid4()

    # Create entry for owner
    await create_cache_entry(
        session=db_session,
        user_id=owner_id,
        connection_id=conn_id,
        cloud_item_id=uuid.uuid4(),
        provider_item_id="owner-item",
        version_key="etag-v1",
        object_key=f"cache/{owner_id}/{uuid.uuid4()}.pdf",
        content_type="application/pdf",
        size_bytes=1024,
    )

    # Foreign user should see zero entries
    foreign_entries = await list_cache_entries(session=db_session, user_id=foreign_id)
    assert len(foreign_entries) == 0, (
        "Cache entries must be invisible to non-owner users (CACHE-05)"
    )

    # Owner should see their own entries
    owner_entries = await list_cache_entries(session=db_session, user_id=owner_id)
    assert len(owner_entries) == 1


async def test_cache_api_excludes_object_key_from_response(async_client, db_session):
    """T-14-02: Cache settings/status API response never includes the MinIO object key.

    Object keys are internal and must not appear in any API response.
    """
    from api.cloud.analysis import router as analysis_router  # Phase 14

    ctx = await _create_user_and_token(db_session)

    # GET cache settings/status endpoint
    resp = await async_client.get(
        "/api/cloud/analysis/cache",
        headers=ctx["headers"],
    )

    # May return 200 (empty state) or 404/422 if the endpoint exists
    assert resp.status_code != 500, "Cache status endpoint must not 500"

    if resp.status_code == 200:
        body_text = resp.text
        # T-14-02: object_key must not appear in any response
        assert "object_key" not in body_text, (
            "MinIO object_key must not appear in cache API responses (T-14-02)"
        )


async def test_admin_cannot_read_user_cache_status(async_client, db_session):
    """T-14-01/CACHE-05: Admin cannot access user's cache metadata via the API."""
    from api.cloud.analysis import router as analysis_router  # Phase 14

    user_ctx = await _create_user_and_token(db_session, role="user")
    admin_ctx = await _create_user_and_token(db_session, role="admin")

    # Admin tries to read the user's cache
    resp = await async_client.get(
        "/api/cloud/analysis/cache",
        headers=admin_ctx["headers"],
    )

    # Admin is blocked from cloud analysis routes entirely (same as browse/mutations)
    assert resp.status_code in (403, 404), (
        f"Admin must not read cache metadata — got {resp.status_code}"
    )


# ─── Version-key helper ───────────────────────────────────────────────────────

async def test_version_key_prefers_version_over_etag():
    """ANALYZE-06/D-19: Version key computation prefers 'version' field over etag.

    Precedence: version > etag > metadata fingerprint > content hash.
    Content hash must never be a pre-download requirement.
    """
    from services.cloud_cache import compute_version_key  # Phase 14

    # version takes priority
    key_with_version = compute_version_key(
        provider_item_id="item-001",
        version="v3",
        etag="etag-stale",
        size=1024,
        modified_at="2026-06-01T00:00:00Z",
        content_type="application/pdf",
    )
    key_from_etag_only = compute_version_key(
        provider_item_id="item-001",
        version=None,
        etag="etag-stable",
        size=1024,
        modified_at="2026-06-01T00:00:00Z",
        content_type="application/pdf",
    )
    key_from_fingerprint = compute_version_key(
        provider_item_id="item-001",
        version=None,
        etag=None,
        size=1024,
        modified_at="2026-06-01T00:00:00Z",
        content_type="application/pdf",
    )

    # Keys derived from different inputs must differ
    assert key_with_version != key_from_etag_only, (
        "version-derived key must differ from etag-only key"
    )
    assert key_from_etag_only != key_from_fingerprint, (
        "etag-derived key must differ from metadata-fingerprint key"
    )

    # Same inputs must produce stable keys (deterministic)
    key_repeat = compute_version_key(
        provider_item_id="item-001",
        version="v3",
        etag="etag-stale",
        size=1024,
        modified_at="2026-06-01T00:00:00Z",
        content_type="application/pdf",
    )
    assert key_with_version == key_repeat, "compute_version_key must be deterministic"


async def test_version_key_fallback_to_fingerprint_without_etag():
    """ANALYZE-06/D-19: Version key falls back to metadata fingerprint when etag absent."""
    from services.cloud_cache import compute_version_key  # Phase 14

    key = compute_version_key(
        provider_item_id="item-fingerprint",
        version=None,
        etag=None,
        size=2048,
        modified_at="2026-06-15T10:00:00Z",
        content_type="text/plain",
    )

    assert key is not None
    assert isinstance(key, str)
    assert len(key) > 0

    # Same metadata must produce same fingerprint key
    key_repeat = compute_version_key(
        provider_item_id="item-fingerprint",
        version=None,
        etag=None,
        size=2048,
        modified_at="2026-06-15T10:00:00Z",
        content_type="text/plain",
    )
    assert key == key_repeat


async def test_content_hash_not_required_during_version_key_computation():
    """ANALYZE-06/D-20: Content hash is not computed during version key resolution.

    Passing content_hash=None must not cause a byte-fetch fallback.
    """
    from services.cloud_cache import compute_version_key  # Phase 14

    # Providing no content_hash still produces a valid key via metadata fingerprint
    key = compute_version_key(
        provider_item_id="item-no-hash",
        version=None,
        etag=None,
        size=512,
        modified_at="2026-06-20T12:00:00Z",
        content_type="application/pdf",
        content_hash=None,  # Must be accepted without triggering any download
    )

    assert key is not None
    assert isinstance(key, str)