[Fix] Restore data_parallel_rank alias on native /generate (dp-aware gateway routing is silently dropped) (#33565)

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

python/sglang/srt/managers/io_struct.py

@@ -263,6 +263,11 @@ class GenerateReqInput:
 
     # For DP routing — external router assigns a specific DP worker
     routed_dp_rank: Optional[int] = None
+    # Deprecated alias for `routed_dp_rank`, still accepted because
+    # sgl-model-gateway's dp-aware mode injects this spelling into every
+    # request it forwards (DPAwareWorker::prepare_request), and the OpenAI
+    # entrypoints and Engine.generate() accept it as well.
+    data_parallel_rank: Optional[int] = None
     # For PD disagg — hint telling decode which prefill DP worker has the KV cache
     disagg_prefill_dp_rank: Optional[int] = None
     # Routing key for routing-key schedule policy
@@ -365,6 +370,18 @@ class GenerateReqInput:
             ValueError: If inputs are not properly specified (e.g., none or all of
                        text, input_ids, input_embeds are provided)
         """
+        if self.data_parallel_rank is not None:
+            import warnings
+
+            warnings.warn(
+                "'data_parallel_rank' is deprecated, use 'routed_dp_rank' instead.",
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            if self.routed_dp_rank is None:
+                self.routed_dp_rank = self.data_parallel_rank
+            self.data_parallel_rank = None
+
         self._validate_inputs()
         self._determine_batch_size()
         if self.session_id is not None and self.session_params is not None:

test/registered/unit/managers/test_io_struct.py

@@ -703,6 +703,30 @@ class TestGenerateReqInputNormalization(CustomTestCase):
             )
             req.normalize_batch_and_arguments()
 
+    def test_data_parallel_rank_alias_maps_to_routed_dp_rank(self):
+        req = GenerateReqInput(text="Hello", sampling_params={}, data_parallel_rank=2)
+        req.normalize_batch_and_arguments()
+        self.assertEqual(req.routed_dp_rank, 2)
+        self.assertIsNone(req.data_parallel_rank)
+
+    def test_data_parallel_rank_alias_does_not_override_routed_dp_rank(self):
+        req = GenerateReqInput(
+            text="Hello", sampling_params={}, data_parallel_rank=2, routed_dp_rank=1
+        )
+        req.normalize_batch_and_arguments()
+        self.assertEqual(req.routed_dp_rank, 1)
+
+    def test_data_parallel_rank_alias_propagates_to_batch_items(self):
+        req = GenerateReqInput(
+            text=["Hello", "World"],
+            sampling_params=[{}, {}],
+            rid=["id1", "id2"],
+            data_parallel_rank=3,
+        )
+        req.normalize_batch_and_arguments()
+        self.assertEqual(req[0].routed_dp_rank, 3)
+        self.assertEqual(req[1].routed_dp_rank, 3)
+
 
 class TestEmbeddingReqInputGetItem(CustomTestCase):
     """Test EmbeddingReqInput.__getitem__."""