chore: /v1/inspect/routes to return only valid non-deprecated routes by default

src/llama_stack/apis/inspect/inspect.py

@@ -4,7 +4,7 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
-from typing import Protocol, runtime_checkable
+from typing import Literal, Protocol, runtime_checkable
 
 from pydantic import BaseModel
 
@@ -12,6 +12,9 @@ from llama_stack.apis.version import LLAMA_STACK_API_V1
 from llama_stack.providers.datatypes import HealthStatus
 from llama_stack.schema_utils import json_schema_type, webmethod
 
+# Valid API level values for filtering routes
+ApiLevel = Literal["v1", "v1alpha", "v1beta", "deprecated"]
+
 
 @json_schema_type
 class RouteInfo(BaseModel):
@@ -64,11 +67,13 @@ class Inspect(Protocol):
     """
 
     @webmethod(route="/inspect/routes", method="GET", level=LLAMA_STACK_API_V1)
-    async def list_routes(self) -> ListRoutesResponse:
+    async def list_routes(self, api_level: ApiLevel | None = None) -> ListRoutesResponse:
         """List routes.
 
         List all available API routes with their methods and implementing providers.
 
+        :param api_level: Optional filter for API level. Can be 'v1', 'v1alpha', 'v1beta', or 'deprecated'.
+                         If not specified, returns only non-deprecated v1 APIs and providers.
         :returns: Response containing information about all available routes.
         """
         ...

src/llama_stack/core/inspect.py

@@ -39,9 +39,21 @@ class DistributionInspectImpl(Inspect):
     async def initialize(self) -> None:
         pass
 
-    async def list_routes(self) -> ListRoutesResponse:
+    async def list_routes(self, api_level: str | None = None) -> ListRoutesResponse:
         run_config: StackRunConfig = self.config.run_config
 
+        # Helper function to determine if a route should be included based on api_level filter
+        def should_include_route(webmethod) -> bool:
+            if api_level is None:
+                # Default: only non-deprecated v1 APIs
+                return not webmethod.deprecated and webmethod.level == "v1"
+            elif api_level == "deprecated":
+                # Include only deprecated routes
+                return webmethod.deprecated
+            else:
+                # Include routes matching the specified level (non-deprecated)
+                return not webmethod.deprecated and webmethod.level == api_level
+
         ret = []
         external_apis = load_external_apis(run_config)
         all_endpoints = get_all_api_routes(external_apis)
@@ -55,8 +67,8 @@ class DistributionInspectImpl(Inspect):
                             method=next(iter([m for m in e.methods if m != "HEAD"])),
                             provider_types=[],  # These APIs don't have "real" providers - they're internal to the stack
                         )
-                        for e, _ in endpoints
+                        for e, webmethod in endpoints
-                        if e.methods is not None
+                        if e.methods is not None and should_include_route(webmethod)
                     ]
                 )
             else:
@@ -69,8 +81,8 @@ class DistributionInspectImpl(Inspect):
                                 method=next(iter([m for m in e.methods if m != "HEAD"])),
                                 provider_types=[p.provider_type for p in providers],
                             )
-                            for e, _ in endpoints
+                            for e, webmethod in endpoints
-                            if e.methods is not None
+                            if e.methods is not None and should_include_route(webmethod)
                         ]
                     )