[CI] Add Triton Ascend in CI (#4921)

Add triton-ascend in UT and e2e

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c
---------
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>

tests/ut/conftest.py

@@ -15,10 +15,23 @@
 # limitations under the License.
 # This file is a part of the vllm-ascend project.
 #
+import sys
+from unittest.mock import MagicMock
 
 from vllm_ascend.utils import adapt_patch  # noqa E402
 from vllm_ascend.utils import register_ascend_customop
 
+# triton and torch_npu is not available in the environment, so we need to mock them
+sys.modules['torch_npu'].npu.current_device = MagicMock(return_value=0)
+sys.modules['torch_npu._inductor'] = MagicMock()
+
+triton_runtime = MagicMock()
+triton_runtime.driver.active.utils.get_device_properties.return_value = {
+    'num_aic': 8,
+    'num_vectorcore': 8,
+}
+sys.modules['triton.runtime'] = triton_runtime
+
 adapt_patch()
 adapt_patch(True)