update cassettes AllUsage

Author: Laure-di

cmd/scw/testdata/test-all-usage-inference-deployment-create-usage.golden

@@ -14,13 +14,15 @@ ARGS:
   [tags.{index}]                                           List of tags to apply to the deployment
   [min-size]                                               Defines the minimum size of the pool
   [max-size]                                               Defines the maximum size of the pool
+  [endpoints.{index}.is-public=false]                      Will configure your public endpoint if true
   [endpoints.{index}.private-network.private-network-id]   
   [endpoints.{index}.disable-auth=false]                   Disable the authentication on the endpoint.
   [quantization.bits]                                      The number of bits each model parameter should be quantized to. The quantization method is chosen based on this value.
   [region=fr-par]                                          Region to target. If none is passed will use default region from the config (fr-par)
 
 FLAGS:
   -h, --help   help for create
+  -w, --wait   wait until the deployment is ready
 
 GLOBAL FLAGS:
   -c, --config string    The path to the config file

cmd/scw/testdata/test-all-usage-inference-deployment-delete-usage.golden

@@ -11,6 +11,7 @@ ARGS:
 
 FLAGS:
   -h, --help   help for delete
+  -w, --wait   wait until the deployment is ready
 
 GLOBAL FLAGS:
   -c, --config string    The path to the config file

internal/namespaces/inference/v1/custom.go

@@ -19,6 +19,7 @@ func GetCommands() *core.Commands {
 	human.RegisterMarshalerFunc(inference.Deployment{}, DeploymentMarshalerFunc)
 
 	cmds.MustFind("inference", "deployment", "create").Override(deploymentCreateBuilder)
+	cmds.MustFind("inference", "deployment", "delete").Override(deploymentDeleteBuilder)
 
 	return cmds
 }

internal/namespaces/inference/v1/custom_deployment.go

@@ -2,8 +2,10 @@ package inference
 
 import (
 	"context"
+	"errors"
+	"fmt"
+	"net/http"
 	"reflect"
-	"strings"
 	"time"
 
 	"github.com/fatih/color"
@@ -15,6 +17,8 @@ import (
 
 const (
 	deploymentActionTimeout = 60 * time.Minute
+	deploymentActionCreate  = 1
+	deploymentActionDelete  = 2
 )
 
 var deploymentStateMarshalSpecs = human.EnumMarshalSpecs{
@@ -43,40 +47,13 @@ func DeploymentMarshalerFunc(i any, opt *human.MarshalOpt) (string, error) {
 	return str, nil
 }
 
-var completeListNodeTypesCache *inference.ListNodeTypesResponse
+func deploymentDeleteBuilder(c *core.Command) *core.Command {
+	c.WaitFunc = waitForDeploymentFunc(deploymentActionDelete)
 
-func autocompleteDeploymentNodeType(
-	ctx context.Context,
-	prefix string,
-	request any,
-) core.AutocompleteSuggestions {
-	req := request.(*inference.CreateDeploymentRequest)
-	suggestions := core.AutocompleteSuggestions(nil)
-
-	client := core.ExtractClient(ctx)
-	api := inference.NewAPI(client)
-
-	if completeListNodeTypesCache == nil {
-		res, err := api.ListNodeTypes(&inference.ListNodeTypesRequest{
-			Region: req.Region,
-		})
-		if err != nil {
-			return nil
-		}
-		completeListNodeTypesCache = res
-	}
-
-	for _, nodeType := range completeListNodeTypesCache.NodeTypes {
-		if strings.HasPrefix(nodeType.Name, prefix) {
-			suggestions = append(suggestions, nodeType.Name)
-		}
-	}
-
-	return suggestions
+	return c
 }
 
 func deploymentCreateBuilder(c *core.Command) *core.Command {
-	c.ArgSpecs.GetByName("node-type-name").AutoCompleteFunc = autocompleteDeploymentNodeType
 	type llmInferenceEndpointSpecCustom struct {
 		*inference.EndpointSpec
 		IsPublic bool `json:"is-public"`
@@ -96,17 +73,8 @@ func deploymentCreateBuilder(c *core.Command) *core.Command {
 
 	c.ArgsType = reflect.TypeOf(llmInferenceCreateDeploymentRequestCustom{})
 
-	c.WaitFunc = func(ctx context.Context, _, respI any) (any, error) {
-		api := inference.NewAPI(core.ExtractClient(ctx))
+	c.WaitFunc = waitForDeploymentFunc(deploymentActionCreate)
 
-		return api.WaitForDeployment(&inference.WaitForDeploymentRequest{
-			DeploymentID:  respI.(*inference.Deployment).ID,
-			Region:        respI.(*inference.Deployment).Region,
-			Status:        respI.(*inference.Deployment).Status,
-			Timeout:       scw.TimeDurationPtr(deploymentActionTimeout),
-			RetryInterval: core.DefaultRetryInterval,
-		})
-	}
 	c.Interceptor = func(ctx context.Context, argsI any, runner core.CommandRunner) (any, error) {
 		deploymentCreateCustomRequest := argsI.(*llmInferenceCreateDeploymentRequestCustom)
 		deploymentRequest := deploymentCreateCustomRequest.CreateDeploymentRequest
@@ -145,3 +113,35 @@ func deploymentCreateBuilder(c *core.Command) *core.Command {
 
 	return c
 }
+
+func waitForDeploymentFunc(action int) core.WaitFunc {
+	return func(ctx context.Context, _, respI any) (any, error) {
+		deployment, err := inference.NewAPI(core.ExtractClient(ctx)).WaitForDeployment(&inference.WaitForDeploymentRequest{
+			DeploymentID:  respI.(*inference.Deployment).ID,
+			Region:        respI.(*inference.Deployment).Region,
+			Timeout:       scw.TimeDurationPtr(deploymentActionTimeout),
+			RetryInterval: core.DefaultRetryInterval,
+		})
+
+		switch action {
+		case deploymentActionCreate:
+			return deployment, err
+		case deploymentActionDelete:
+			if err != nil {
+				// if we get a 404 here, it means the resource was successfully deleted
+				notFoundError := &scw.ResourceNotFoundError{}
+				responseError := &scw.ResponseError{}
+				if errors.As(err, &responseError) &&
+					responseError.StatusCode == http.StatusNotFound ||
+					errors.As(err, &notFoundError) {
+					return fmt.Sprintf(
+						"Server %s successfully deleted.",
+						respI.(*inference.Deployment).ID,
+					), nil
+				}
+			}
+		}
+
+		return nil, err
+	}
+}

internal/namespaces/inference/v1/custom_deployment_test.go

@@ -51,16 +51,15 @@ func Test_CreateDeploymentPrivateEndpoint(t *testing.T) {
 		Commands:   cmds,
 		BeforeFunc: CreatePN(),
 		Cmd: fmt.Sprintf(
-			"scw inference deployment create model-id=%s node-type-name=%s accept-eula=true endpoints.0.private-network.private-network-id={{ .PN.ID }}",
+			"scw inference deployment create model-id=%s node-type-name=H100-SXM-2 accept-eula=true endpoints.0.private-network.private-network-id={{ .PN.ID }} --wait",
 			ModelID,
-			NodeTypeName,
 		),
 		Check: core.TestCheckCombine(
 			core.TestCheckGolden(),
 		),
 		AfterFunc: core.AfterFuncCombine(
+			core.ExecAfterCmd("scw inference deployment delete {{ .CmdResult.ID }} --wait"),
 			DeletePrivateNetwork(),
-			DeleteDeployment(),
 		),
 	}))
 }

internal/namespaces/inference/v1/helper_test.go

@@ -27,7 +27,3 @@ func CreatePN() core.BeforeFunc {
 func DeletePrivateNetwork() core.AfterFunc {
 	return core.ExecAfterCmd("scw vpc private-network delete {{ .PN.ID }}")
 }
-
-func DeleteDeployment() core.AfterFunc {
-	return core.ExecAfterCmd("scw inference deployment delete {{ .DEPLOYMENT.ID }}")
-}