phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-07-26 22:19:49 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat(registry): more flexible model lookup (#2859)

Browse source 
This PR updates model registration and lookup behavior to be slightly
more general / flexible. See
https://github.com/meta-llama/llama-stack/issues/2843 for more details.

Note that this change is backwards compatible given the design of the
`lookup_model()` method.

## Test Plan

Added unit tests
This commit is contained in:
Ashwin Bharambe 2025-07-22 15:22:48 -07:00 • committed by GitHub
parent 9736f096f6
commit 3b83032555
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
15 changed files with 265 additions and 75 deletions
Download patch file Download diff file Expand all files Collapse all files

6
llama_stack/distribution/routing_tables/toolgroups.py
View file

@ -30,7 +30,7 @@ class ToolGroupsRoutingTable(CommonRoutingTableImpl, ToolGroups):
tool_to_toolgroup: dict[str, str] = {}
# overridden
def get_provider_impl(self, routing_key: str, provider_id: str | None = None) -> Any:
async def get_provider_impl(self, routing_key: str, provider_id: str | None = None) -> Any:
# we don't index tools in the registry anymore, but only keep a cache of them by toolgroup_id
# TODO: we may want to invalidate the cache (for a given toolgroup_id) every once in a while?
@ -40,7 +40,7 @@ class ToolGroupsRoutingTable(CommonRoutingTableImpl, ToolGroups):
if routing_key in self.tool_to_toolgroup:
routing_key = self.tool_to_toolgroup[routing_key]
return super().get_provider_impl(routing_key, provider_id)
return await super().get_provider_impl(routing_key, provider_id)
async def list_tools(self, toolgroup_id: str | None = None) -> ListToolsResponse:
if toolgroup_id:
@ -59,7 +59,7 @@ class ToolGroupsRoutingTable(CommonRoutingTableImpl, ToolGroups):
return ListToolsResponse(data=all_tools)
async def _index_tools(self, toolgroup: ToolGroup):
provider_impl = super().get_provider_impl(toolgroup.identifier, toolgroup.provider_id)
provider_impl = await super().get_provider_impl(toolgroup.identifier, toolgroup.provider_id)
tooldefs_response = await provider_impl.list_runtime_tools(toolgroup.identifier, toolgroup.mcp_endpoint)
# TODO: kill this Tool vs ToolDef distinction