Buckets:
AnonymousSpottedToad/candidate_bucket / preliminary /mappings /python /limit_request_rate /avinash-mall_CognitiveMemoryLayer.json
| { | |
| "role_mappings": [ | |
| { | |
| "role": "Entity", | |
| "status": "present", | |
| "symbol": "request", | |
| "file": "packages/py-cml/src/cml/transport/http.py", | |
| "lines": "211", | |
| "notes": "In-repo py-cml SDK client: HTTPTransport.request (http.py:211) sends requests to the CML API with the X-API-Key header built in _build_headers (http.py:147), which is exactly the identifier RateLimitMiddleware.dispatch keys on (src/api/middleware.py:97). An async twin exists in AsyncHTTPTransport." | |
| }, | |
| { | |
| "role": "Enforcer", | |
| "status": "present", | |
| "symbol": "dispatch", | |
| "file": "src/api/middleware.py", | |
| "lines": "90", | |
| "notes": "RateLimitMiddleware.dispatch intercepts every request, derives the id (sha256 of X-API-Key, else client IP; middleware.py:111-114), calls _check_rate_limit (middleware.py:116) and returns HTTP 429 when blocked, else call_next. Registered via app.add_middleware(RateLimitMiddleware, ...) at src/api/app.py:138, but only when auth.rate_limit_requests_per_minute > 0, which defaults to 0 (off by default; mapped per rule 5 as the only matching flow). Health/OPTIONS, 'test-key' and admin key are exempt." | |
| }, | |
| { | |
| "role": "Limiter", | |
| "status": "present", | |
| "symbol": "_check_rate_limit_redis", | |
| "file": "src/api/middleware.py", | |
| "lines": "141", | |
| "notes": "Reached from dispatch via _check_rate_limit (middleware.py:116 -> 130 -> 132), which picks Redis when app.state.db.redis exists. _check_rate_limit_redis increments the per-key counter in a 60s window and returns False when count > requests_per_minute (middleware.py:149). Fallback _check_rate_limit_in_memory (middleware.py:157) does the same with an in-process TTLCache when Redis is unavailable." | |
| }, | |
| { | |
| "role": "HistoryProvider", | |
| "status": "external", | |
| "symbol": "incr", | |
| "file": "src/api/middleware.py", | |
| "lines": "145", | |
| "notes": "Called by the Limiter _check_rate_limit_redis: Redis (redis.asyncio client from src/storage/connection.py:124) stores the request counter under 'ratelimit:<key>' via pipeline incr/expire (middleware.py:145-146), returning the count (history) and recording the request; decr on block (middleware.py:150). Fallback history is the in-memory TTLCache self._buckets (middleware.py:83)." | |
| }, | |
| { | |
| "role": "PolicyProvider", | |
| "status": "present", | |
| "symbol": "rate_limit_requests_per_minute", | |
| "file": "src/core/config.py", | |
| "lines": "116", | |
| "notes": "No method provides the policy; the limit is the AuthSettings field rate_limit_requests_per_minute (env AUTH__RATE_LIMIT_REQUESTS_PER_MINUTE, default 0 = disabled), read at src/api/app.py:136 and passed to RateLimitMiddleware.__init__ (middleware.py:80) as self.requests_per_minute, used by the Limiter at middleware.py:149 together with the fixed 60s window _RATE_LIMIT_WINDOW_SECONDS (middleware.py:16). Single global policy, not per-id." | |
| }, | |
| { | |
| "role": "System", | |
| "status": "present", | |
| "symbol": "write_memory", | |
| "file": "src/api/routes.py", | |
| "lines": "70", | |
| "notes": "After dispatch passes (call_next at middleware.py:124), requests reach the /api/v1 routers included at src/api/app.py:141; e.g. POST /api/v1/memory/write handled by write_memory (routes.py:70), which the SDK Entity calls. All API routes except /api/v1/health are equally guarded." | |
| } | |
| ] | |
| } | |
Xet Storage Details
- Size:
- 3.41 kB
- Xet hash:
- 17645e04cca73523b4c263d50949303a015e0afe65efd73c34ae46b08918244b
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.