Address comment

internal/modelcontroller/adapters.go

@@ -16,7 +16,7 @@ const (
 )
 
 // reconcileAdapters ensures that the specified adapters are loaded in the model server pods.
-// Loaded adapters are identified by the presence of a label with the adapter name and the hash
+// Loaded adapters are identified by the presence of a Pod label with the adapter name and the hash
 // of the adapter URL.
 // At request-time, the endpoint resolver will inspect these labels to determine which adapters
 // are loaded in the pod.

internal/modelcontroller/engine_vllm.go

@@ -1,11 +1,9 @@
 package modelcontroller
 
 import (
-	"context"
 	"sort"
 
 	kubeaiv1 "github.com/substratusai/kubeai/api/v1"
-	v1 "github.com/substratusai/kubeai/api/v1"
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/util/intstr"
@@ -28,17 +26,17 @@ func (r *ModelReconciler) vLLMPodForModel(m *kubeaiv1.Model, c ModelConfig) *cor
 		"--model=" + vllmModelFlag,
 		"--served-model-name=" + m.Name,
 	}
-	if m.Spec.Adapters != nil {
-		args = append(args, "--enable-lora")
-	}
 	args = append(args, m.Spec.Args...)
 
-	env := []corev1.EnvVar{
-		{
+	env := []corev1.EnvVar{}
+
+	if m.Spec.Adapters != nil {
+		args = append(args, "--enable-lora")
+		env = append(env, corev1.EnvVar{
 			// https://docs.vllm.ai/en/latest/models/lora.html#dynamically-serving-lora-adapters
 			Name:  "VLLM_ALLOW_RUNTIME_LORA_UPDATING",
 			Value: "True",
-		},
+		})
 	}
 
 	var envKeys []string
@@ -151,7 +149,3 @@ func (r *ModelReconciler) vLLMPodForModel(m *kubeaiv1.Model, c ModelConfig) *cor
 
 	return pod
 }
-
-func (r *ModelReconciler) vLLMLoadAdapter(ctx context.Context, pod *corev1.Pod, adapter v1.Adapter) error {
-	return nil
-}