[CI]Fix broken CI (#1773)

This PR fixed the broken CI. It require
vllm-project/vllm#20900 merged first.

- vLLM version: v0.9.2
- vLLM main:
vllm-project/vllm@e8cc53a

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>

Author: wangxiyuan

tests/conftest.py

@@ -47,6 +47,7 @@
 from vllm_ascend.utils import adapt_patch  # noqa E402
 
 adapt_patch(True)
+adapt_patch(False)
 
 from vllm.distributed.parallel_state import (  # noqa E402
     destroy_distributed_environment, destroy_model_parallel)

tests/e2e/singlecard/core/ascend_scheduler/test_ascend_scheduler.py

@@ -16,6 +16,7 @@
 from vllm.v1.structured_output import StructuredOutputManager
 
 from vllm_ascend.core.scheduler import AscendScheduler
+from vllm_ascend.utils import vllm_version_is
 
 EOS_TOKEN_ID = 50256
 
@@ -303,6 +304,8 @@ def test_stop_via_update_from_output():
         req.num_computed_tokens = req.num_tokens
         scheduler.requests[req.request_id] = req
         scheduler.running.append(req)
+        if not vllm_version_is("0.9.2"):
+            req.status = RequestStatus.RUNNING
 
     scheduler_output = SchedulerOutput(scheduled_new_reqs=[],
                                        scheduled_cached_reqs=[],
@@ -355,6 +358,8 @@ def test_stop_via_update_from_output():
         req.num_computed_tokens = req.num_tokens
         scheduler.requests[req.request_id] = req
         scheduler.running.append(req)
+        if not vllm_version_is("0.9.2"):
+            req.status = RequestStatus.RUNNING
 
     scheduler_output = SchedulerOutput(scheduled_new_reqs=[],
                                        scheduled_cached_reqs=[],
@@ -405,6 +410,8 @@ def test_stop_via_update_from_output():
         req.num_computed_tokens = req.num_tokens
         scheduler.requests[req.request_id] = req
         scheduler.running.append(req)
+        if not vllm_version_is("0.9.2"):
+            req.status = RequestStatus.RUNNING
 
     scheduler_output = SchedulerOutput(scheduled_new_reqs=[],
                                        scheduled_cached_reqs=[],

tests/e2e/singlecard/test_prompt_embedding.py

@@ -31,6 +31,7 @@
 from vllm.v1.structured_output import StructuredOutputManager
 
 from vllm_ascend.core.scheduler import AscendScheduler
+from vllm_ascend.utils import vllm_version_is
 
 EOS_TOKEN_ID = 50256
 
@@ -213,6 +214,8 @@ def test_stop_via_update_from_output():
         scheduler.requests[req.request_id] = req
         scheduler.running.append(req)
         scheduler.scheduled_req_ids.add(req.request_id)
+        if not vllm_version_is("0.9.2"):
+            req.status = RequestStatus.RUNNING
 
     scheduler_output = SchedulerOutput(scheduled_new_reqs=[],
                                        scheduled_cached_reqs=[],
@@ -263,6 +266,8 @@ def test_stop_via_update_from_output():
         scheduler.requests[req.request_id] = req
         scheduler.running.append(req)
         scheduler.scheduled_req_ids.add(req.request_id)
+        if not vllm_version_is("0.9.2"):
+            req.status = RequestStatus.RUNNING
 
     scheduler_output = SchedulerOutput(scheduled_new_reqs=[],
                                        scheduled_cached_reqs=[],
@@ -311,6 +316,8 @@ def test_stop_via_update_from_output():
         scheduler.requests[req.request_id] = req
         scheduler.running.append(req)
         scheduler.scheduled_req_ids.add(req.request_id)
+        if not vllm_version_is("0.9.2"):
+            req.status = RequestStatus.RUNNING
 
     scheduler_output = SchedulerOutput(scheduled_new_reqs=[],
                                        scheduled_cached_reqs=[],