fix(adaptive_router): P2 review items — @updatedAt + snapshot samples

- Mark last_updated_at (AdaptiveRouterState) and last_activity_at
  (AdaptiveRouterSession) with @updatedAt so Prisma refreshes the
  timestamps on every write. Without this the fields stayed frozen at
  INSERT time and the last_activity_at index was misleading for any
  future TTL/eviction logic. Applied to all three schema.prisma copies;
  no migration SQL change needed (Prisma @updatedAt is a client-side
  annotation that doesn't touch DDL).

- get_state_snapshot: report cell.total_samples instead of alpha+beta
  for the 'samples' field. The previous value inflated every cell by
  the COLD_START_MASS prior (e.g. showed 10.0 before any real traffic
  arrived), which confused operators reading /adaptive_router/.../state.
  Updated docs + the snapshot test to match.

Also fixes two pre-existing merge-break syntax errors in router.py
(missing ')' on the AdaptiveRouter TYPE_CHECKING import; truncated
async_pre_routing_hook dispatch call for the adaptive router branch)
that were masking the rest of the file from the interpreter.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

docs/my-website/docs/adaptive_router.md

@@ -131,13 +131,13 @@ Returns current quality estimates per model per request type. Useful for underst
           "request_type": "analytical_reasoning",
           "model": "fast",
           "quality_mean": 0.5,
-          "samples": 10.0
+          "samples": 0
         },
         {
           "request_type": "analytical_reasoning",
           "model": "smart",
           "quality_mean": 0.95,
-          "samples": 10.0
+          "samples": 0
         }
       ]
     }
@@ -145,7 +145,7 @@ Returns current quality estimates per model per request type. Useful for underst
 }
 ```
 
-`quality_mean` is the key number — it's the router's current estimate of how well that model handles that request type. `samples` counts how many real observations have moved the prior (starts at 10, the cold-start mass).
+`quality_mean` is the key number — it's the router's current estimate of how well that model handles that request type. `samples` counts how many real observations have moved the prior (starts at 0; the cold-start prior mass is excluded).
 
 ## Known limitations
 

litellm-proxy-extras/litellm_proxy_extras/schema.prisma

@@ -1232,7 +1232,7 @@ model LiteLLM_AdaptiveRouterState {
   alpha           Float
   beta            Float
   total_samples   Int      @default(0)
-  last_updated_at DateTime @default(now())
+  last_updated_at DateTime @default(now()) @updatedAt
 
   @@id([router_name, request_type, model_name])
 }
@@ -1261,7 +1261,7 @@ model LiteLLM_AdaptiveRouterSession {
   last_processed_turn    Int      @default(-1)
   clean_credit_awarded   Boolean  @default(false)
   terminal_status        Int?
-  last_activity_at       DateTime @default(now())
+  last_activity_at       DateTime @default(now()) @updatedAt
 
   @@id([session_id, router_name, model_name])
   @@index([last_activity_at])

litellm/proxy/schema.prisma

@@ -1232,7 +1232,7 @@ model LiteLLM_AdaptiveRouterState {
   alpha           Float
   beta            Float
   total_samples   Int      @default(0)
-  last_updated_at DateTime @default(now())
+  last_updated_at DateTime @default(now()) @updatedAt
 
   @@id([router_name, request_type, model_name])
 }
@@ -1261,7 +1261,7 @@ model LiteLLM_AdaptiveRouterSession {
   last_processed_turn    Int      @default(-1)
   clean_credit_awarded   Boolean  @default(false)
   terminal_status        Int?
-  last_activity_at       DateTime @default(now())
+  last_activity_at       DateTime @default(now()) @updatedAt
 
   @@id([session_id, router_name, model_name])
   @@index([last_activity_at])

litellm/router.py

@@ -202,6 +202,7 @@ if TYPE_CHECKING:
     )
     from litellm.router_strategy.adaptive_router.adaptive_router import (
         AdaptiveRouter,
+    )
     from litellm.router_strategy.quality_router.quality_router import (
         QualityRouter,
     )
@@ -9901,6 +9902,14 @@ class Router:
         adaptive_router = self.adaptive_routers.get(model)
         if adaptive_router is not None:
             return await adaptive_router.async_pre_routing_hook(
+                model=model,
+                request_kwargs=request_kwargs,
+                messages=messages,
+                input=input,
+                specific_deployment=specific_deployment,
+            )
+
+        #########################################################
         # Check if any quality-router should be used
         #########################################################
         if model in self.quality_routers:

litellm/router_strategy/adaptive_router/adaptive_router.py

@@ -261,7 +261,11 @@ class AdaptiveRouter:
                     "model": model,
                     "alpha": cell.alpha,
                     "beta": cell.beta,
-                    "samples": total,
+                    # Net observations that have moved the posterior, excluding
+                    # the cold-start prior mass. `alpha + beta` would show the
+                    # initial COLD_START_MASS (e.g. 10) before any real traffic
+                    # arrives, which confuses operators reading the endpoint.
+                    "samples": cell.total_samples,
                     "quality_mean": cell.alpha / total if total > 0 else 0.0,
                 }
             )

schema.prisma

@@ -1232,7 +1232,7 @@ model LiteLLM_AdaptiveRouterState {
   alpha           Float
   beta            Float
   total_samples   Int      @default(0)
-  last_updated_at DateTime @default(now())
+  last_updated_at DateTime @default(now()) @updatedAt
 
   @@id([router_name, request_type, model_name])
 }
@@ -1261,7 +1261,7 @@ model LiteLLM_AdaptiveRouterSession {
   last_processed_turn    Int      @default(-1)
   clean_credit_awarded   Boolean  @default(false)
   terminal_status        Int?
-  last_activity_at       DateTime @default(now())
+  last_activity_at       DateTime @default(now()) @updatedAt
 
   @@id([session_id, router_name, model_name])
   @@index([last_activity_at])

tests/test_litellm/router_strategy/adaptive_router/test_state_endpoint.py

@@ -89,7 +89,9 @@ async def test_get_state_snapshot_quality_mean_matches_alpha_over_total():
     )
     assert cell["alpha"] == expected.alpha
     assert cell["beta"] == expected.beta
-    assert cell["samples"] == expected.alpha + expected.beta
+    # `samples` reports net observations after subtracting the cold-start
+    # prior mass, so operators aren't misled by the initial value.
+    assert cell["samples"] == expected.total_samples
     assert cell["quality_mean"] == pytest.approx(expected_mean)