fix(rate-limiting): regenerate dashboard schema, close lint-budget gaps from rebase

Rebasing onto litellm_internal_staging pulled in 234 new upstream commits;
regenerates ui/litellm-dashboard/src/lib/http/schema.d.ts for the new
apply_to_key_alias field and suppresses/annotates the handful of
mutable-construction and missing-Final findings the stricter, moved
type-discipline budget now surfaces across this PR's own prior commits.

litellm/proxy/common_request_processing.py

@@ -2321,7 +2321,9 @@ class ProxyBaseLLMRequestProcessing:
 
         try:
             if general_settings.get("cancel_on_disconnect", False):
-                responses = await _await_llm_call_cancelling_on_disconnect(request, llm_responses, self.data)
+                responses = await _await_llm_call_cancelling_on_disconnect(  # rebind-ok: assigned in exactly one of these two mutually exclusive branches
+                    request, llm_responses, self.data
+                )
             else:
                 responses = await llm_responses
         finally:

litellm/proxy/hooks/global_tag_rate_limits_hook.py

@@ -155,9 +155,9 @@ _request_stash: Final[ContextVar[_GlobalTagRateLimitStash | None]] = ContextVar(
 
 
 def _claim_stash_for_data(data: Mapping[str, object]) -> _GlobalTagRateLimitStash:
-    stash = _request_stash.get()
+    stash = _request_stash.get()  # rebind-ok: lazily initialized below if this ContextVar has never been set
     if stash is None:
-        stash = _GlobalTagRateLimitStash()
+        stash = _GlobalTagRateLimitStash()  # rebind-ok: see above
         _request_stash.set(stash)
     owner_call_id: Final = data.get("litellm_call_id")
     if isinstance(owner_call_id, str):
@@ -174,9 +174,7 @@ def _stash_for_call(litellm_call_id: str | None) -> _GlobalTagRateLimitStash | N
     return stash if litellm_call_id == stash.owner_litellm_call_id else None
 
 
-def _call_id_from_kwargs(kwargs: object) -> str | None:
-    if not isinstance(kwargs, dict):
-        return None
+def _call_id_from_kwargs(kwargs: Mapping[str, object]) -> str | None:
     call_id: Final = kwargs.get("litellm_call_id")
     return call_id if isinstance(call_id, str) else None
 
@@ -507,7 +505,7 @@ class _PROXY_GlobalTagRateLimitsHook(  # pyright: ignore[reportUnusedClass]  # o
         return data
 
     async def async_release_disconnect_state_hook(self, request_data: Mapping[str, object]) -> None:
-        stash: Final = _stash_for_call(_call_id_from_kwargs(dict(request_data)))
+        stash: Final = _stash_for_call(_call_id_from_kwargs(request_data))
         if stash is None or not stash.pending_concurrency_keys:
             return
         release_keys: Final = tuple(stash.pending_concurrency_keys)

litellm/proxy/hooks/model_based_tag_rate_limits_hook.py

@@ -242,12 +242,12 @@ def _entry_applies(entry: TagRateLimitEntry, tag_value: str, tags: Sequence[str]
     if entry.included_values is not None and tag_value not in entry.included_values:
         return False
     if entry.disabled_for is not None:
-        gate_value = _extract_identity(tags, entry.disabled_for.tag_id)
-        if gate_value is not None and gate_value in entry.disabled_for.values:
+        disabled_gate_value: Final = _extract_identity(tags, entry.disabled_for.tag_id)
+        if disabled_gate_value is not None and disabled_gate_value in entry.disabled_for.values:
             return False
     if entry.enabled_for is not None:
-        gate_value = _extract_identity(tags, entry.enabled_for.tag_id)
-        if gate_value is None or gate_value not in entry.enabled_for.values:
+        enabled_gate_value: Final = _extract_identity(tags, entry.enabled_for.tag_id)
+        if enabled_gate_value is None or enabled_gate_value not in entry.enabled_for.values:
             return False
     if entry.apply_to_key_alias is None:
         return True
@@ -901,9 +901,9 @@ def _queue_pending_concurrency_reservations(
     model_call_details: Final = getattr(logging_obj, "model_call_details", None)
     if not isinstance(model_call_details, dict):
         return
-    pending = model_call_details.get(_PENDING_CONCURRENCY_KEYS_FIELD)
+    pending = model_call_details.get(_PENDING_CONCURRENCY_KEYS_FIELD)  # rebind-ok: lazily initialized below when absent
     if pending is None:
-        pending = []  # mutable-ok: shared, request-scoped accumulator; see field's own docstring
+        pending = []  # mutable-ok: shared, request-scoped accumulator; see field's own docstring  # rebind-ok: lazily initialized only when absent
         model_call_details[_PENDING_CONCURRENCY_KEYS_FIELD] = pending
     pending.extend(reservations)  # mutable-ok: see comment above
 

litellm/types/router.py

@@ -171,7 +171,7 @@ class TagRateLimitScope(BaseModel):
         # is silently ignored when constructing via __init__ (only takes
         # effect via model_validate), so mutating in place is the only way
         # this normalization reliably applies regardless of construction path.
-        object.__setattr__(self, "values", tuple(sorted(set(self.values))))
+        object.__setattr__(self, "values", tuple(sorted(set(self.values))))  # mutable-ok: frozen before escaping
         return self
 
 
@@ -281,11 +281,11 @@ class TagRateLimitEntry(BaseModel):
         # unsorted tuple would make config-order alone decide whether two
         # deployments' entries dedup to one shared bucket.
         if self.included_values is not None:
-            self.included_values = tuple(sorted(set(self.included_values)))
+            self.included_values = tuple(sorted(set(self.included_values)))  # mutable-ok: frozen before escaping
         if self.excluded_values is not None:
-            self.excluded_values = tuple(sorted(set(self.excluded_values)))
+            self.excluded_values = tuple(sorted(set(self.excluded_values)))  # mutable-ok: frozen before escaping
         if self.apply_to_key_alias is not None:
-            self.apply_to_key_alias = tuple(sorted(set(self.apply_to_key_alias)))
+            self.apply_to_key_alias = tuple(sorted(set(self.apply_to_key_alias)))  # mutable-ok: frozen before escaping
         return self
 
 

ui/litellm-dashboard/src/lib/http/schema.d.ts

@@ -35112,6 +35112,8 @@ export interface components {
         };
         /** TagRateLimitEntry */
         TagRateLimitEntry: {
+            /** Apply To Key Alias */
+            apply_to_key_alias?: string[] | null;
             disabled_for?: components["schemas"]["TagRateLimitScope"] | null;
             enabled_for?: components["schemas"]["TagRateLimitScope"] | null;
             /** Excluded Values */