- cache.py:38: Add explicit type annotation for redis.get() return value to resolve mypy no-any-return
- rate_limit.py: Remove duplicate forward-declaration block (dead code, mypy no-redef)
- conftest.py: Remove one excess blank line to satisfy ruff format check
All three fixes verified locally: ruff check ✅, ruff format ✅, mypy ✅
Co-Authored-By: Paperclip <noreply@paperclip.ing>
QA review of PR #39 (CAR-1121) identified three blocking issues; this
commit addresses all three plus the typecheck errors flagged as CI RED.
CAR-1077 (PR #39) changes:
- database.py: add pool_timeout=30 so the engine fails fast when the
connection pool is exhausted (defends against the "server closed
connection unexpectedly" pod failures).
- routes/health.py: /health now calls SELECT 1 through Depends(get_db)
and raises HTTPException(503) when the database is unreachable, so
Kubernetes readiness probes can correctly mark the pod unhealthy and
stop routing traffic to it. Logs the failure at exception level for
observability.
- Drop .mcp.json from this PR (root-level MCP server config, not
related to the pool fix; tracked separately).
CI typecheck fixes (pre-existing on dev, were failing mypy on PR #39):
- auth/passwords.py: cast bcrypt return values so mypy doesn't widen
to Any.
- config.py: silence the false-positive call-arg on Settings() — the
three required fields are populated from the environment by
pydantic-settings at runtime.
- cache.py: coerce the bytes/str union returned by the redis client
to the documented str | None return type.
- middleware/rate_limit.py: annotate the three module-level limiters
with the RateLimitBackend protocol, cast the redis zrange score to
float before arithmetic, and add max_requests/window_seconds to the
protocol so the response-header builder can read them.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
mypy complained: 'Unsupported operand types for - ("str" and "float")'
on rate_limit.py:87. redis-py's zrange withscores=True returns the
score as whatever the codec produces (often str), but we treat it as
a numeric millisecond timestamp. Cast to float before subtracting
the cutoff.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The data routes (purchases, alerts, stores, etc.) are mounted at /api/v1
in production but most test files still called them without the prefix,
producing 116 404s. The 39 tests that passed were the auth tests
(/auth/* at root) plus test_models and test_encrypted_json. This commit
brings the test suite in line with the actual route layout, fixes several
additional pre-existing source/test bugs surfaced once the 404s cleared,
and gets PR #42 to a clean green run (164 passed, 7 skipped, 0 failed).
Source fixes
- src/cartsnitch_api/auth/dependencies.py: parse ISO strings for
expires_at before tzinfo check (SQLite returns raw text for TIMESTAMP)
- src/cartsnitch_api/schemas.py: UserResponse.id is UUID, matching the
actual model type and avoiding ResponseValidationError on /auth/me
Test alignment
- tests/test_routes/*, tests/test_e2e/*: add /api/v1 prefix to all data
route calls (auth routes left alone — they live at root)
- tests/test_openapi.py: refresh EXPECTED_ROUTES to match the actual
OpenAPI spec (drop Better-Auth-only routes, add /api/v1 prefix,
update route count to 31)
Pre-existing test fixes
- tests/test_middleware/test_rate_limit.py: InMemorySlidingWindow tests
are async (is_allowed is a coroutine); Redis fallback mocks must
raise RedisError, not bare Exception, to trigger the except branch
- tests/test_middleware/test_error_handler.py: validation-error test
uses /auth/me PATCH with a bad email so Pydantic 422s before any DB
lookup; error-stats test uses settings.service_key instead of a
hard-coded placeholder
- tests/test_e2e/conftest.py: Coupon.valid_to is date.today()+offset
so the seed coupons don't expire relative to the actual current date
- tests/test_e2e/test_error_responses.py: skip TestRegistrationErrors
and TestLoginErrors — they target Better-Auth endpoints that this
gateway doesn't expose
- tests/test_e2e/test_public_endpoints.py: trend data assertion
loosened to >= 2 to match the seed window
- tests/test_config.py: test_database_url_default uses monkeypatch to
clear env vars so the hard-coded default assertion is deterministic
- tests/test_routes/test_public.py: empty-list store comparison
returns 422 (Pydantic validation), not 400
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Three categories of pre-existing CI failure on PR #42:
1. typecheck (mypy src/cartsnitch_api, 9 errors):
- src/cartsnitch_api/config.py:89 — Settings() needs required secret
args that only exist in env at runtime; suppress with
type: ignore[call-arg]
- src/cartsnitch_api/cache.py:38 — redis-py returns Any/bytes,
normalize to str before returning from get()
- src/cartsnitch_api/middleware/rate_limit.py:128,131,134 — three
limiter globals were inferred as RedisSlidingWindow on the if
branch then re-assigned InMemorySlidingWindow on else; declare
them as RateLimitBackend up front
- src/cartsnitch_api/middleware/rate_limit.py:181,187 —
RateLimitBackend Protocol didn't declare max_requests even
though both InMemorySlidingWindow and RedisSlidingWindow expose
it; add max_requests: int to the Protocol
2. test (FK constraint on purchases.user_id):
- tests/conftest.py:_create_test_user_and_session stored user_id
as 32-char hex; test_e2e conftest reads it via raw SQL and wraps
in uuid.UUID (36 chars) before passing to Purchase.user_id, so
the FK never matched. Switch back to str(uuid.uuid4()) (36 chars)
so the stored value and the FK bind value use the same format.
3. Verify lint + format clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Auto-fix F401 (unused imports) and I001 (unsorted imports) with ruff --fix
- Manually fix E501 (line too long) in alembic migrations and src/ models
- Run ruff format to ensure consistent formatting
Co-Authored-By: Paperclip <noreply@paperclip.ing>
fix: remove dead dispose_engine import from API main.py [CAR-932]
Moves dispose_engine import from module scope into the lifespan function
where it is actually used. Fixes ImportError crashing API pods.
Reviewed-by: cs_charlie (QA)
Approved-by: cs_savannah (CTO)
CI-override: pre-existing failures unrelated to this change
The top-level import of dispose_engine from cartsnitch_api.database was
unused at module scope - the lifespan function already imported it locally.
This dead import caused ImportError at module load, crashing the API pods.
Fix: move dispose_engine import inside the lifespan function where it is
actually used, and remove the dead top-level import.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Add dev.cartsnitch.com and uat.cartsnitch.com to the CORS origins list
to match the infra HTTPRoute domains and fix auth blocking on UAT.
Refs: CAR-992
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Add rate_limit_auth_requests (5/min) and rate_limit_auth_window_seconds (60) settings
- Add rate_limit_redis_enabled flag for opt-in Redis usage
- Refactor _SlidingWindowCounter into InMemorySlidingWindow class
- Add RedisSlidingWindow using sorted sets with fallback to in-memory
- Add third _auth_strict_limiter for POST /auth/* paths (5 req/min)
- Add protocol-based backend selection at module load time
- Update tests for auth strict limiter and Redis fallback behavior
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Add async Redis client using redis-py with connection pooling
- Implement get/set/delete with graceful degradation when unavailable
- Add TTL support (default 300s) via SETEX
- Add cache invalidation hooks for price and product changes
- Use pattern-based SCAN for bulk invalidation
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Add rate_limit_auth_requests (5/min) and rate_limit_auth_window_seconds (60)
settings to config.py
- Refactor rate_limit.py to use protocol/ABC pattern with InMemorySlidingWindow
and RedisSlidingWindow implementations
- Add RedisSlidingWindow using sorted sets for distributed rate limiting
- Add auth_strict_limiter for /auth/* POST endpoints (5 req/min per IP)
- Fall back to in-memory when Redis is unavailable
- Update tests to cover new functionality
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Add connection pool config to SQLAlchemy async engine (pool_size=10, max_overflow=20, pool_pre_ping, pool_recycle)
- Implement Redis connection pool in CacheClient with initialize/close lifecycle
- Wire lifespan startup/shutdown to initialize and dispose pools
- Add dispose_engine() for graceful DB pool cleanup on shutdown
Closes CAR-550
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Add days query param to GET /public/trends/{product_id} (ge=1, le=365)
- Add category query param to GET /public/store-comparison
- Add category and period query params to GET /public/inflation
- Add boundary and malicious input test cases
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Remove dangerous default values for jwt_secret_key, service_key, and
fernet_key. Add startup validation that raises RuntimeError if these
secrets are not set via environment variables or contain placeholder
values.
Add test fixture to provide explicit test values for these secrets,
ensuring existing tests continue to pass.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Better-Auth sets the session cookie as "token.sessionId". The DB stores
only the token part, so passing the full compound value caused 401s.
Splits on "." for both cookie and Bearer paths.
Tests added for compound cookie, raw token cookie (regression), and
compound Bearer token.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
API config.py now reads CARTSNITCH_DATABASE_URL first, falls back to
DATABASE_URL (which the infra K8s overlay sets for all pods), and finally
falls back to the hardcoded default. Also normalizes plain postgresql://
to postgresql+asyncpg:// for the asyncpg driver.
Fixes CAR-510.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Better-auth v1.5.6 stores raw 32-char tokens in sessions.token, not SHA-256
hashes. The SHA-256 fix from PR #136 causes all authenticated API calls to
return 401 because the UAT sessions table contains raw tokens.
- Remove hashlib from dependencies.py; compare tokens directly
- Remove hashlib from conftest.py; store raw tokens in test DB
- Remove hashlib from test_expired_session_rejected; use raw tokens
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Better-Auth v1.2+ stores SHA-256(raw_token) in the sessions.token
column. The cookie/Bearer header carries the raw token, so the API was
doing a plain-text lookup that would never match a hashed value —
causing all authenticated endpoints to return 401.
- Add hashlib import and hash token in _validate_session_token()
- Update conftest._create_test_user_and_session() to store hashed tokens
- Update test_expired_session_rejected() to store hashed tokens
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Better-auth sets the session cookie with the __Secure- prefix on HTTPS
deployments. The API was only reading the plain cookie name, causing all
authenticated calls to return 401 in dev/UAT/prod environments.
Check __Secure-better-auth.session_token first, fall back to
better-auth.session_token for HTTP local dev compatibility.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Fix email format in AuthService.get_email_in_address to use
receipts+{token}@receipts.cartsnitch.com (was broken: @email.cartsnitch.com)
- Remove dead EmailInAddressResponse class and GET /auth/me/email-in-address
endpoint from auth/routes.py (endpoint moved to routes/user.py)
- Add instructions field to EmailInAddressResponse schema
- Update routes/user.py to include instructions in the response
- Update test URLs from /auth/me/email-in-address to /api/v1/me/email-in-address
Co-authored-by: CartSnitch Engineer Bot <cartnoreply@cartsnitch.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
The GET /me/email-in-address endpoint was unreachable because the Gateway
HTTPRoute routes all /auth/* traffic to Better-Auth (port 3001), not the
API service. This change:
- Moves the endpoint from the /auth router to a new /api/v1/me/ router
- Adds EmailInAddressResponse schema and get_email_in_address service method
- Updates Settings.tsx to call /api/v1/me/email-in-address
Fixes CAR-445.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Revert auth/dependencies.py to cookie+Bearer dual auth with str user IDs
- Add GET /auth/me/email-in-address endpoint for receipt email routing
- Update User model: add email_inbound_token, change id/store_id/user_id to str
- Update AuthService and UserResponse to use str user IDs
- Update route count test: 33 -> 34 routes
- Restore e2e test for email-in-address endpoint
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Revert auth/dependencies.py, auth/routes.py, services/auth.py, schemas.py
to Better-Auth session-cookie auth (removed JWT register/login/refresh)
- Preserve GET /auth/me/email-in-address endpoint
- Fix UUIDString TypeDecorator: process_result_value returns uuid.UUID
(not str) so SQLAlchemy 2.0 sentinel tracking matches UUID-to-UUID
- Fix seed_data fixture: look up real user_id from session token via
sessions table; purchases now reference actual user FK
- Update purchase_data fixture to use session-cookie auth
- Update test_auth_endpoints, test_auth_validation to cookie-based tests
- Remove TestRegistrationErrors and TestLoginErrors (no longer applicable)
- Update test_openapi.py expected routes and count
- Update test_error_handler.py to use PATCH /auth/me validation
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Remove TimestampMixin (created_at/updated_at) from Purchase, PurchaseItem,
PriceHistory, Coupon, and ShrinkflationEvent models since their PostgreSQL
tables do not have those columns. This was causing 500 errors on
/api/v1/purchases and /api/v1/purchases/stats.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Better-Auth uses nanoid strings for user IDs, not UUIDs. Changed all
user_id parameter/return types in the API layer from UUID to str,
removed the obsolete UUID import where unused, and updated the
_validate_session_token return type accordingly.
Co-authored-by: CartSnitch Engineer Bot <cartnoreply@cartsnitch.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
Better-Auth v1.5.6 stores raw tokens in sessions.token, not SHA-256
hashes. The session cookie is signed (rawToken.hmacSignature), so
strip the HMAC signature suffix before querying the DB.
Fixes 401 errors on all data endpoints caused by the incorrect hash.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Restores sha256 import and token hashing in _validate_session_token.
Regression introduced when PR #95 (cookie name fix) was merged without
the hash fix from PR #93.
QA approved: CAR-324 (Checkout Charlie)
CTO approved: Paperclip (Savannah Savings)
Resolves CAR-323
cc @cpfarhood
Better-Auth automatically prefixes cookie names with __Secure- when serving
over HTTPS. The API gateway now tries __Secure-better-auth.session_token
first (HTTPS/deployed), falling back to better-auth.session_token (HTTP/local dev).
Fixes CAR-321.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Better-Auth v1.5.6+ stores session tokens as SHA-256 hashes in the
sessions table. The raw token from the cookie was being queried directly,
causing all authenticated /api/v1/* requests to return 401.
Fixes CAR-313.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Fixes CAR-161 UAT failure: k8s HTTPRoute forwards /api/* to the API
gateway without path rewriting, so requests arrive at FastAPI as
/api/v1/purchases, /api/v1/products, etc. FastAPI previously mounted
data routers at root, causing 404s on all /api/v1/* calls.
Keep health and auth routers at root (probes hit /health directly;
auth traffic is routed to the auth service via HTTPRoute).
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Savannah Savings
Barcode Betty
Barcode Betty
Pawla Abdul