HIGH PRIORITY: Fix for 2.19 model-explainability test failures (#439)

* fix: lmeval test format

* fix: add securityContext to minio container

Author: sheltoncyril

tests/model_explainability/lm_eval/conftest.py

@@ -27,7 +27,10 @@ def lmevaljob_hf(
         model_args=[{"name": "pretrained", "value": "google/flan-t5-base"}],
         task_list={
             "taskRecipes": [
-                {"card": {"name": "cards.wnli"}, "template": "templates.classification.multi_class.relation.default"}
+                {
+                    "card": {"name": "cards.wnli"},
+                    "template": {"name": "templates.classification.multi_class.relation.default"},
+                }
             ]
         },
         log_samples=True,

tests/model_explainability/trustyai_service/conftest.py

@@ -117,6 +117,12 @@ def minio_pod(admin_client: DynamicClient, model_namespace: Namespace) -> Genera
                 "image": "quay.io/trustyai_testing/modelmesh-minio-examples"
                 "@sha256:d2ccbe92abf9aa5085b594b2cae6c65de2bf06306c30ff5207956eb949bb49da",
                 "name": MINIO,
+                "securityContext": {
+                    "allowPrivilegeEscalation": False,
+                    "capabilities": {"drop": ["ALL"]},
+                    "runAsNonRoot": True,
+                    "seccompProfile": {"type": "RuntimeDefault"},
+                },
             }
         ],
         label={"app": "minio", "maistra.io/expose-route": "true"},