Support v0.10.1 (#2584)

### What this PR does / why we need it?
This patch also supports v0.10.1

### Does this PR introduce _any_ user-facing change?
No

### How was this patch tested?
- CI passed
- test 0.10.1: https://github.com/vllm-project/vllm-ascend/pull/2583
- vLLM version: v0.10.1.1
- vLLM main:
321938e9ac

Signed-off-by: Yikun Jiang <yikunkero@gmail.com>

tests/ut/core/test_scheduler.py

@@ -21,7 +21,7 @@ from tests.ut.base import TestBase
 from vllm_ascend.core.scheduler import AscendScheduler
 from vllm_ascend.utils import vllm_version_is
 
-if not vllm_version_is("0.10.1.1"):
+if not (vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1")):
     from vllm.v1.outputs import DraftTokenIds
 else:
     DraftTokenIds = None
@@ -78,7 +78,7 @@ def make_output(scheduler):
     }
     sampled_token_ids = [[1000]] * len(scheduler.running)
     logprobs = None
-    if vllm_version_is("0.10.1.1"):
+    if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
         modelrunner_output = ModelRunnerOutput(
             req_ids=req_ids,
             req_id_to_index=req_id_to_index,
@@ -297,7 +297,7 @@ class TestAscendScheduler(TestBase):
             scheduler.running.append(req)
             req.status = RequestStatus.RUNNING
 
-        if vllm_version_is("0.10.1.1"):
+        if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
             scheduler_output = SchedulerOutput(
                 scheduled_new_reqs=[],
                 scheduled_cached_reqs=[],
@@ -384,7 +384,7 @@ class TestAscendScheduler(TestBase):
             scheduler.running.append(req)
             req.status = RequestStatus.RUNNING
 
-        if vllm_version_is("0.10.1.1"):
+        if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
             scheduler_output = SchedulerOutput(
                 scheduled_new_reqs=[],
                 scheduled_cached_reqs=[],
@@ -468,7 +468,7 @@ class TestAscendScheduler(TestBase):
             scheduler.running.append(req)
             req.status = RequestStatus.RUNNING
 
-        if vllm_version_is("0.10.1.1"):
+        if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
             scheduler_output = SchedulerOutput(
                 scheduled_new_reqs=[],
                 scheduled_cached_reqs=[],
@@ -549,7 +549,7 @@ class TestAscendScheduler(TestBase):
         scheduler.requests[requests[0].request_id] = requests[0]
         scheduler.running.append(requests[0])
 
-        if vllm_version_is("0.10.1.1"):
+        if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
             scheduler_output = SchedulerOutput(
                 scheduled_new_reqs=[],
                 scheduled_cached_reqs=[],
@@ -645,7 +645,7 @@ class TestAscendScheduler(TestBase):
                 512)
 
             # Model output of the first request.
-            if vllm_version_is("0.10.1.1"):
+            if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
                 model_runner_output = ModelRunnerOutput(
                     req_ids=[requests[0].request_id],
                     req_id_to_index={requests[0].request_id: 0},
@@ -671,7 +671,7 @@ class TestAscendScheduler(TestBase):
             # request is still running.
             scheduler.schedule()
             # Model output of the second request.
-            if vllm_version_is("0.10.1.1"):
+            if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
                 model_runner_output = ModelRunnerOutput(
                     req_ids=[requests[1].request_id],
                     req_id_to_index={requests[1].request_id: 0},
@@ -739,7 +739,7 @@ class TestAscendScheduler(TestBase):
                 req_id = requests[i].request_id
                 self.assertEqual(output.num_scheduled_tokens[req_id], 1)
                 self.assertNotIn(req_id, output.scheduled_spec_decode_tokens)
-            if vllm_version_is("0.10.1.1"):
+            if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
                 model_runner_output = ModelRunnerOutput(
                     req_ids=req_ids,
                     req_id_to_index=req_to_index,
@@ -760,7 +760,7 @@ class TestAscendScheduler(TestBase):
 
             engine_core_outputs = scheduler.update_from_output(
                 output, model_runner_output)
-            if not vllm_version_is("0.10.1.1"):
+            if not (vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1")):
                 scheduler.update_draft_token_ids(draft_token_ids)
 
             for i in range(len(requests)):
@@ -797,7 +797,7 @@ class TestAscendScheduler(TestBase):
                 else:
                     self.assertNotIn(req_id,
                                      output.scheduled_spec_decode_tokens)
-            if vllm_version_is("0.10.1.1"):
+            if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
                 model_runner_output = ModelRunnerOutput(
                     req_ids=req_ids,
                     req_id_to_index=req_to_index,

tests/ut/kv_connector/utils.py

@@ -200,7 +200,7 @@ def create_model_runner_output(
     kv_connector_output = KVConnectorOutput(finished_sending=finished_sending,
                                             finished_recving=finished_recving)
     extra_args = {"kv_connector_output": kv_connector_output}
-    if vllm_version_is("0.10.1.1"):
+    if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
         model_runner_output = ModelRunnerOutput(
             req_ids=req_ids,
             req_id_to_index=req_id_to_index,