Add documentloader CRD and add audio handler

[nkwangleiGIT]

1. Add documentloader CRD to handle uploaded files from user(read file from s3)
2. Add audio handler using whisper, so we can transcribe audio to txt and then chat with audio content

[Abirdcfly]

we should split resource get into func (p *Executor) Init, like

arcadia/pkg/appruntime/knowledgebase/knowledgebase.go

Lines 41 to 48 in 352145b

	func (k *Knowledgebase) Init(ctx context.Context, cli client.Client, _ map[string]any) error {
	instance := &v1alpha1.KnowledgeBase{}
	if err := cli.Get(ctx, types.NamespacedName{Namespace: k.RefNamespace(), Name: k.Ref.Name}, instance); err != nil {
	return fmt.Errorf("can't find the knowledgebase in cluster: %w", err)
	}
	k.Instance = instance
	return nil
	}

And add func (p *Executor) Ready() like

arcadia/pkg/appruntime/knowledgebase/knowledgebase.go

Lines 54 to 56 in 352145b

	func (k *Knowledgebase) Ready() (isReady bool, msg string) {
	return k.Instance.Status.IsReadyOrGetReadyMessage()
	}

method for app health check.

Since the state update has not been implemented yet, we can implement this along with the state update in one pr.

[nkwangleiGIT]

let's add it to make documentloader always ready, and check if we need to add some logic later.

[Abirdcfly]

we should use klog.FromContext(ctx).xxx to replace klog, We add a requestid to each request in the logs to make debugging easier, in

arcadia/apiserver/pkg/requestid/requestid.go

Lines 29 to 36 in 352145b

	func addRequestIDToLog(c *gin.Context, id string) {
	ctx := c.Request.Context()
	logger := klog.FromContext(ctx)
	newLogger := logger.WithValues("requestID", id)
	newLogger.Info("new request")
	c.Request = c.Request.WithContext(klog.NewContext(ctx, newLogger))
	c.Next()

But it looks like this is an issue elsewhere as well, and we can leave that for a subsequent pr along with an update.

[nkwangleiGIT]

yes, klog it more easy to use

[Abirdcfly]

Need to double check that since we are using group name arcadia.kubeagi.k8s.com.cn, should we move these files to api/base?

[nkwangleiGIT]

I think about this before, but documentloader is a Node that should run with app runtime, so place it to app-node.
api/base seems just resources that does not have any Node run logic, so don't use it although they have the same group. Let's keep it for now?