scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-08 03:08:45 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 115

fix(proxy): resolve pre-existing basedpyright reportArgumentType/reportCallIssue errors

Browse source 
model_based_tag_rate_limits_hook.py had 3 real type errors present since its
introduction, invisible in scoped per-file checks but visible once codebase-wide
upstream drift finally pushed the totals over ceiling: sorted()/groupby() keyed
by a raw dict lookup returning object rather than a provably orderable type, and
a tuple passed where async_increment_tokens_with_ttl_preservation expects a
list. Adds a small typed _model_name_of accessor for the first two, and drops
an unneeded tuple() conversion for the third since the value was already a
list. Ran make lint-budget-update per repo convention.
This commit is contained in:
Deepanshu 2026-08-27 12:15:19 -04:00
parent b98195e3a4
commit 563b3d6adf
2 changed files with 13 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
litellm/proxy/hooks/model_based_tag_rate_limits_hook.py
View file

@ -175,6 +175,15 @@ def _deployment_id(deployment: Mapping[str, object]) -> str | None:
return (deployment.get("model_info") or _EMPTY_MAPPING).get("id")
def _model_name_of(deployment: Mapping[str, object]) -> str:
"""`model_name` is a required field on every deployment dict; `str(...)`
(rather than a bare index) gives `sorted`/`groupby`'s key functions a
provably orderable, hashable return type without an unsafe cast -- every
real deployment's value here is already a string, so this is a no-op
coercion in practice."""
return str(deployment["model_name"])
def _extract_team_id(request_kwargs: Mapping[str, object], metadata_variable_name: str) -> str | None:
"""Reads `user_api_key_team_id` from only the one field
`get_metadata_variable_name_from_kwargs` names as authoritative for this
@ -444,7 +453,7 @@ def _build_limits_index(model_list: Sequence[Mapping[str, object]]) -> _LimitsIn
would have seen them in without the sort, which is what keeps this safe
(that relative order decides first-seen signature order downstream).
"""
sorted_by_model_name: Final = sorted(model_list, key=lambda deployment: deployment["model_name"])
sorted_by_model_name: Final = sorted(model_list, key=_model_name_of)
by_model_name: Final[Mapping[str, tuple[_ConfiguredLimit, ...]]] = MappingProxyType(
{
model_name: (
@ -461,9 +470,7 @@ def _build_limits_index(model_list: Sequence[Mapping[str, object]]) -> _LimitsIn
if (team_scope := next((key[0] for dep in group if (key := _team_alias_key(dep))), None)) is not None
else configured
)
for model_name, deployment_group in groupby(
sorted_by_model_name, key=lambda deployment: deployment["model_name"]
)
for model_name, deployment_group in groupby(sorted_by_model_name, key=_model_name_of)
for group in (tuple(deployment_group),)
if (configured := tuple(limit for unit in _LIMIT_UNITS for limit in _build_group_limits(group, unit)))
}
@ -1727,7 +1734,7 @@ class _PROXY_ModelBasedTagRateLimitsHook( # pyright: ignore[reportUnusedClass]
partition = await self._partition_for(partition_key) # not Final: rebound each loop iteration
accounting_task = asyncio.create_task( # not Final: rebound each loop iteration
partition.v3.async_increment_tokens_with_ttl_preservation(
pipeline_operations=tuple(group_operations),
pipeline_operations=group_operations,
parent_otel_span=parent_otel_span,
)
)

2
type-discipline-budget.json
View file

@ -27,7 +27,7 @@
"limit": 0
},
"LIT010": {
"limit": 16616
"limit": 16615
},
"LIT011": {
"limit": 5583