v1alpha1

package
v0.12.0-rc1 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Jan 27, 2024 License: Apache-2.0 Imports: 18 Imported by: 13

Documentation

Overview

Package v1alpha1 contains API Schema definitions for the serving v1alpha1 API group

Package v1alpha1 contains API Schema definitions for the serving v1alpha1 API group +k8s:openapi-gen=true +kubebuilder:object:generate=true +k8s:conversion-gen=kserve/pkg/apis/serving +k8s:defaulter-gen=TypeMeta +groupName=serving.kserve.io

Index

Constants

View Source
const (
	// InvalidGraphNameFormatError defines the error message for invalid inference graph name
	InvalidGraphNameFormatError = "" /* 231-byte string literal not displayed */
	// RootNodeNotFoundError defines the error message for root node not found
	RootNodeNotFoundError = "root node not found, InferenceGraph needs a node with name 'root' as the root node of the graph"
	// WeightNotProvidedError defines the error message for traffic weight is nil for inference step
	WeightNotProvidedError = "InferenceGraph[%s] Node[%s] Route[%s] missing the 'Weight'"
	// InvalidWeightError defines the error message for sum of traffic weight is not 100
	InvalidWeightError = "InferenceGraph[%s] Node[%s] splitter node: the sum of traffic weights for all routing targets should be 100"
	// DuplicateStepNameError defines the error message for more than one step contains same name
	DuplicateStepNameError = "Node \"%s\" of InferenceGraph \"%s\" contains more than one step with name \"%s\""
	// TargetNotProvidedError defines the error message for inference graph target not specified
	TargetNotProvidedError = "Step %d (\"%s\") in node \"%s\" of InferenceGraph \"%s\" does not specify an inference target"
	// InvalidTargetError defines the error message for inference graph target specifies more than one of nodeName, serviceName, serviceUrl
	InvalidTargetError = "Step %d (\"%s\") in node \"%s\" of InferenceGraph \"%s\" specifies more than one of nodeName, serviceName, serviceUrl"
)
View Source
const (
	// InferenceServiceReady is set when inference service reported readiness
	InferenceServiceReady apis.ConditionType = "InferenceServiceReady"
	// MemoryResourceAvailable is set when inference service reported resources availability
	MemoryResourceAvailable apis.ConditionType = "MemoryResourceAvailable"
	// IsMMSPredictor is set when inference service predictor is set to multi-model serving
	IsMMSPredictor apis.ConditionType = "IsMMSPredictor"
)

ConditionType represents a Service condition value

View Source
const (
	CommaSpaceSeparator                 = ", "
	TmNameFmt                    string = "[a-zA-Z0-9_-]+"
	InvalidTmNameFormatError            = "" /* 180-byte string literal not displayed */
	InvalidStorageUriFormatError        = "" /* 144-byte string literal not displayed */
	InvalidTmMemoryModification         = "the Trained Model \"%s\" memory field is immutable. The memory was \"%s\" but it is updated to \"%s\""
)

regular expressions for validation of isvc name

View Source
const (
	// GraphNameFmt regular expressions for validation of isvc name
	GraphNameFmt string = "[a-z]([-a-z0-9]*[a-z0-9])?"
)
View Source
const (
	// GraphRootNodeName is the root node name.
	GraphRootNodeName string = "root"
)

Variables

View Source
var (

	// regular expressions for validation of tm name
	TmRegexp = regexp.MustCompile("^" + TmNameFmt + "$")
	// protocols that are accepted by storage uri
	StorageUriProtocols = strings.Join(storage.GetAllProtocol(), CommaSpaceSeparator)
)
View Source
var (
	// APIVersion is the current API version used to register these objects
	APIVersion = "v1alpha1"

	// SchemeGroupVersion is group version used to register these objects
	SchemeGroupVersion = schema.GroupVersion{Group: constants.KServeAPIGroupName, Version: APIVersion}

	// SchemeBuilder is used to add go types to the GroupVersionKind scheme
	SchemeBuilder = &scheme.Builder{GroupVersion: SchemeGroupVersion}

	// AddToScheme is required by pkg/client/...
	AddToScheme = SchemeBuilder.AddToScheme
)
View Source
var (

	//GraphRegexp regular expressions for validation of graph name
	GraphRegexp = regexp.MustCompile("^" + GraphNameFmt + "$")
)

Functions

func Resource

func Resource(resource string) schema.GroupResource

Resource is required by pkg/client/listers/...

Types

type BuiltInAdapter added in v0.8.0

type BuiltInAdapter struct {
	// ServerType must be one of the supported built-in types such as "triton" or "mlserver",
	// and the runtime's container must have the same name
	ServerType ServerType `json:"serverType,omitempty"`
	// Port which the runtime server listens for model management requests
	RuntimeManagementPort int `json:"runtimeManagementPort,omitempty"`
	// Fixed memory overhead to subtract from runtime container's memory allocation to determine model capacity
	MemBufferBytes int `json:"memBufferBytes,omitempty"`
	// Timeout for model loading operations in milliseconds
	ModelLoadingTimeoutMillis int `json:"modelLoadingTimeoutMillis,omitempty"`
	// Environment variables used to control other aspects of the built-in adapter's behaviour (uncommon)
	Env []corev1.EnvVar `json:"env,omitempty"`
}

+k8s:openapi-gen=true

func (*BuiltInAdapter) DeepCopy added in v0.8.0

func (in *BuiltInAdapter) DeepCopy() *BuiltInAdapter

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new BuiltInAdapter.

func (*BuiltInAdapter) DeepCopyInto added in v0.8.0

func (in *BuiltInAdapter) DeepCopyInto(out *BuiltInAdapter)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type ClusterServingRuntime added in v0.8.0

type ClusterServingRuntime struct {
	metav1.TypeMeta   `json:",inline"`
	metav1.ObjectMeta `json:"metadata,omitempty"`

	Spec   ServingRuntimeSpec   `json:"spec,omitempty"`
	Status ServingRuntimeStatus `json:"status,omitempty"`
}

ClusterServingRuntime is the Schema for the servingruntimes API +k8s:openapi-gen=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object +genclient +kubebuilder:object:root=true +kubebuilder:resource:scope="Cluster" +kubebuilder:printcolumn:name="Disabled",type="boolean",JSONPath=".spec.disabled" +kubebuilder:printcolumn:name="ModelType",type="string",JSONPath=".spec.supportedModelFormats[*].name" +kubebuilder:printcolumn:name="Containers",type="string",JSONPath=".spec.containers[*].name" +kubebuilder:printcolumn:name="Age",type="date",JSONPath=".metadata.creationTimestamp"

func (*ClusterServingRuntime) DeepCopy added in v0.8.0

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ClusterServingRuntime.

func (*ClusterServingRuntime) DeepCopyInto added in v0.8.0

func (in *ClusterServingRuntime) DeepCopyInto(out *ClusterServingRuntime)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*ClusterServingRuntime) DeepCopyObject added in v0.8.0

func (in *ClusterServingRuntime) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

type ClusterServingRuntimeList added in v0.8.0

type ClusterServingRuntimeList struct {
	metav1.TypeMeta `json:",inline"`
	metav1.ListMeta `json:"metadata,omitempty"`
	Items           []ClusterServingRuntime `json:"items"`
}

ClusterServingRuntimeList contains a list of ServingRuntime +k8s:openapi-gen=true +kubebuilder:object:root=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object

func (*ClusterServingRuntimeList) DeepCopy added in v0.8.0

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ClusterServingRuntimeList.

func (*ClusterServingRuntimeList) DeepCopyInto added in v0.8.0

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*ClusterServingRuntimeList) DeepCopyObject added in v0.8.0

func (in *ClusterServingRuntimeList) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

type ClusterStorageContainer added in v0.11.1

type ClusterStorageContainer struct {
	metav1.TypeMeta   `json:",inline"`
	metav1.ObjectMeta `json:"metadata,omitempty"`

	Spec StorageContainerSpec `json:"spec,omitempty"`

	// +optional
	Disabled *bool `json:"disabled,omitempty"`
}

+k8s:openapi-gen=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object +genclient +kubebuilder:object:root=true +kubebuilder:resource:scope="Cluster"

func (*ClusterStorageContainer) DeepCopy added in v0.11.1

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ClusterStorageContainer.

func (*ClusterStorageContainer) DeepCopyInto added in v0.11.1

func (in *ClusterStorageContainer) DeepCopyInto(out *ClusterStorageContainer)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*ClusterStorageContainer) DeepCopyObject added in v0.11.1

func (in *ClusterStorageContainer) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

func (*ClusterStorageContainer) IsDisabled added in v0.11.1

func (sc *ClusterStorageContainer) IsDisabled() bool

type ClusterStorageContainerList added in v0.11.1

type ClusterStorageContainerList struct {
	metav1.TypeMeta `json:",inline"`
	metav1.ListMeta `json:"metadata,omitempty"`
	Items           []ClusterStorageContainer `json:"items" validate:"required"`
}

+k8s:openapi-gen=true +kubebuilder:object:root=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object

func (*ClusterStorageContainerList) DeepCopy added in v0.11.1

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ClusterStorageContainerList.

func (*ClusterStorageContainerList) DeepCopyInto added in v0.11.1

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*ClusterStorageContainerList) DeepCopyObject added in v0.11.1

func (in *ClusterStorageContainerList) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

type InferenceGraph added in v0.9.0

type InferenceGraph struct {
	metav1.TypeMeta   `json:",inline"`
	metav1.ObjectMeta `json:"metadata,omitempty"`
	Spec              InferenceGraphSpec   `json:"spec,omitempty"`
	Status            InferenceGraphStatus `json:"status,omitempty"`
}

InferenceGraph is the Schema for the InferenceGraph API for multiple models +k8s:openapi-gen=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object +genclient +kubebuilder:object:root=true +kubebuilder:subresource:status +kubebuilder:printcolumn:name="URL",type="string",JSONPath=".status.url" +kubebuilder:printcolumn:name="Ready",type="string",JSONPath=".status.conditions[?(@.type=='Ready')].status" +kubebuilder:printcolumn:name="Age",type="date",JSONPath=".metadata.creationTimestamp" +kubebuilder:resource:path=inferencegraphs,shortName=ig,singular=inferencegraph

func (*InferenceGraph) DeepCopy added in v0.9.0

func (in *InferenceGraph) DeepCopy() *InferenceGraph

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new InferenceGraph.

func (*InferenceGraph) DeepCopyInto added in v0.9.0

func (in *InferenceGraph) DeepCopyInto(out *InferenceGraph)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*InferenceGraph) DeepCopyObject added in v0.9.0

func (in *InferenceGraph) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

func (*InferenceGraph) ValidateCreate added in v0.9.0

func (ig *InferenceGraph) ValidateCreate() (admission.Warnings, error)

ValidateCreate implements webhook.Validator so a webhook will be registered for the type

func (*InferenceGraph) ValidateDelete added in v0.9.0

func (ig *InferenceGraph) ValidateDelete() (admission.Warnings, error)

ValidateDelete implements webhook.Validator so a webhook will be registered for the type

func (*InferenceGraph) ValidateUpdate added in v0.9.0

func (ig *InferenceGraph) ValidateUpdate(old runtime.Object) (admission.Warnings, error)

ValidateUpdate implements webhook.Validator so a webhook will be registered for the type

type InferenceGraphList added in v0.9.0

type InferenceGraphList struct {
	metav1.TypeMeta `json:",inline"`
	metav1.ListMeta `json:"metadata,omitempty"`
	// +listType=set
	Items []InferenceGraph `json:"items"`
}

InferenceGraphList contains a list of InferenceGraph +k8s:openapi-gen=true +kubebuilder:object:root=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object

func (*InferenceGraphList) DeepCopy added in v0.9.0

func (in *InferenceGraphList) DeepCopy() *InferenceGraphList

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new InferenceGraphList.

func (*InferenceGraphList) DeepCopyInto added in v0.9.0

func (in *InferenceGraphList) DeepCopyInto(out *InferenceGraphList)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*InferenceGraphList) DeepCopyObject added in v0.9.0

func (in *InferenceGraphList) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

type InferenceGraphSpec added in v0.9.0

type InferenceGraphSpec struct {
	// Map of InferenceGraph router nodes
	// Each node defines the router which can be different routing types
	Nodes map[string]InferenceRouter `json:"nodes"`
	// +optional
	Resources corev1.ResourceRequirements `json:"resources,omitempty"`
	// +optional
	Affinity *corev1.Affinity `json:"affinity,omitempty" protobuf:"bytes,18,opt,name=affinity"`
	// TimeoutSeconds specifies the number of seconds to wait before timing out a request to the component.
	// +optional
	TimeoutSeconds *int64 `json:"timeout,omitempty"`
	// Minimum number of replicas, defaults to 1 but can be set to 0 to enable scale-to-zero.
	// +optional
	MinReplicas *int `json:"minReplicas,omitempty"`
	// Maximum number of replicas for autoscaling.
	// +optional
	MaxReplicas int `json:"maxReplicas,omitempty"`
	// ScaleTarget specifies the integer target value of the metric type the Autoscaler watches for.
	// concurrency and rps targets are supported by Knative Pod Autoscaler
	//(https://knative.dev/docs/serving/autoscaling/autoscaling-targets/).
	// +optional
	ScaleTarget *int `json:"scaleTarget,omitempty"`
	// ScaleMetric defines the scaling metric type watched by autoscaler
	// possible values are concurrency, rps, cpu, memory. concurrency, rps are supported via
	// Knative Pod Autoscaler(https://knative.dev/docs/serving/autoscaling/autoscaling-metrics).
	// +optional
	ScaleMetric *ScaleMetric `json:"scaleMetric,omitempty"`
}

InferenceGraphSpec defines the InferenceGraph spec +k8s:openapi-gen=true

func (*InferenceGraphSpec) DeepCopy added in v0.9.0

func (in *InferenceGraphSpec) DeepCopy() *InferenceGraphSpec

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new InferenceGraphSpec.

func (*InferenceGraphSpec) DeepCopyInto added in v0.9.0

func (in *InferenceGraphSpec) DeepCopyInto(out *InferenceGraphSpec)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type InferenceGraphStatus added in v0.9.0

type InferenceGraphStatus struct {
	// Conditions for InferenceGraph
	duckv1.Status `json:",inline"`
	// Url for the InferenceGraph
	// +optional
	URL *apis.URL `json:"url,omitempty"`
}

InferenceGraphStatus defines the InferenceGraph conditions and status +k8s:openapi-gen=true

func (*InferenceGraphStatus) DeepCopy added in v0.9.0

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new InferenceGraphStatus.

func (*InferenceGraphStatus) DeepCopyInto added in v0.9.0

func (in *InferenceGraphStatus) DeepCopyInto(out *InferenceGraphStatus)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type InferenceRouter added in v0.9.0

type InferenceRouter struct {
	// RouterType
	//
	// - `Sequence:` chain multiple inference steps with input/output from previous step
	//
	// - `Splitter:` randomly routes to the target service according to the weight
	//
	// - `Ensemble:` routes the request to multiple models and then merge the responses
	//
	// - `Switch:` routes the request to one of the steps based on condition
	//
	RouterType InferenceRouterType `json:"routerType"`

	// Steps defines destinations for the current router node
	// +optional
	Steps []InferenceStep `json:"steps,omitempty"`
}

+k8s:openapi-gen=true InferenceRouter defines the router for each InferenceGraph node with one or multiple steps

```yaml kind: InferenceGraph metadata:

name: canary-route

spec:

nodes:
  root:
    routerType: Splitter
    routes:
    - service: mymodel1
      weight: 20
    - service: mymodel2
      weight: 80

```

```yaml kind: InferenceGraph metadata:

name: abtest

spec:

nodes:
  mymodel:
    routerType: Switch
    routes:
    - service: mymodel1
      condition: "{ .input.userId == 1 }"
    - service: mymodel2
      condition: "{ .input.userId == 2 }"

```

Scoring a case using a model ensemble consists of scoring it using each model separately, then combining the results into a single scoring result using one of the pre-defined combination methods.

Tree Ensemble constitutes a case where simple algorithms for combining results of either classification or regression trees are well known. Multiple classification trees, for example, are commonly combined using a "majority-vote" method. Multiple regression trees are often combined using various averaging techniques. e.g tagging models with segment identifiers and weights to be used for their combination in these ways. ```yaml kind: InferenceGraph metadata:

name: ensemble

spec:

nodes:
  root:
    routerType: Sequence
    routes:
    - service: feast
    - nodeName: ensembleModel
      data: $response
  ensembleModel:
    routerType: Ensemble
    routes:
    - service: sklearn-model
    - service: xgboost-model

```

Scoring a case using a sequence, or chain of models allows the output of one model to be passed in as input to the subsequent models. ```yaml kind: InferenceGraph metadata:

name: model-chainer

spec:

nodes:
  root:
    routerType: Sequence
    routes:
    - service: mymodel-s1
    - service: mymodel-s2
      data: $response
    - service: mymodel-s3
      data: $response

```

In the flow described below, the pre_processing node base64 encodes the image and passes it to two model nodes in the flow. The encoded data is available to both these nodes for classification. The second node i.e. dog-breed-classification takes the original input from the pre_processing node along-with the response from the cat-dog-classification node to do further classification of the dog breed if required. ```yaml kind: InferenceGraph metadata:

name: dog-breed-classification

spec:

nodes:
  root:
    routerType: Sequence
    routes:
    - service: cat-dog-classifier
    - nodeName: breed-classifier
      data: $request
  breed-classifier:
    routerType: Switch
    routes:
    - service: dog-breed-classifier
      condition: { .predictions.class == "dog" }
    - service: cat-breed-classifier
      condition: { .predictions.class == "cat" }

```

func (*InferenceRouter) DeepCopy added in v0.9.0

func (in *InferenceRouter) DeepCopy() *InferenceRouter

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new InferenceRouter.

func (*InferenceRouter) DeepCopyInto added in v0.9.0

func (in *InferenceRouter) DeepCopyInto(out *InferenceRouter)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type InferenceRouterType added in v0.9.0

type InferenceRouterType string

InferenceRouterType constant for inference routing types +k8s:openapi-gen=true +kubebuilder:validation:Enum=Sequence;Splitter;Ensemble;Switch

const (
	// Sequence Default type only route to one destination
	Sequence InferenceRouterType = "Sequence"

	// Splitter router randomly routes the requests to the named service according to the weight
	Splitter InferenceRouterType = "Splitter"

	// Ensemble router routes the requests to multiple models and then merge the responses
	Ensemble InferenceRouterType = "Ensemble"

	// Switch routes the request to the model based on certain condition
	Switch InferenceRouterType = "Switch"
)

InferenceRouterType Enum

type InferenceStep added in v0.9.0

type InferenceStep struct {
	// Unique name for the step within this node
	// +optional
	StepName string `json:"name,omitempty"`

	// Node or service used to process this step
	InferenceTarget `json:",inline"`

	// request data sent to the next route with input/output from the previous step
	// $request
	// $response.predictions
	// +optional
	Data string `json:"data,omitempty"`

	// the weight for split of the traffic, only used for Split Router
	// when weight is specified all the routing targets should be sum to 100
	// +optional
	Weight *int64 `json:"weight,omitempty"`

	// routing based on the condition
	// +optional
	Condition string `json:"condition,omitempty"`

	// to decide whether a step is a hard or a soft dependency in the Inference Graph
	// +optional
	Dependency InferenceStepDependencyType `json:"dependency,omitempty"`
}

InferenceStep defines the inference target of the current step with condition, weights and data. +k8s:openapi-gen=true

func (*InferenceStep) DeepCopy added in v0.9.0

func (in *InferenceStep) DeepCopy() *InferenceStep

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new InferenceStep.

func (*InferenceStep) DeepCopyInto added in v0.9.0

func (in *InferenceStep) DeepCopyInto(out *InferenceStep)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type InferenceStepDependencyType added in v0.11.1

type InferenceStepDependencyType string

InferenceStepDependencyType constant for inference step dependency +k8s:openapi-gen=true +kubebuilder:validation:Enum=Soft;Hard

const (
	// Soft
	Soft InferenceStepDependencyType = "Soft"

	// Hard
	Hard InferenceStepDependencyType = "Hard"
)

StepDependency Enum

type InferenceTarget added in v0.9.0

type InferenceTarget struct {
	// The node name for routing as next step
	// +optional
	NodeName string `json:"nodeName,omitempty"`

	// named reference for InferenceService
	ServiceName string `json:"serviceName,omitempty"`

	// InferenceService URL, mutually exclusive with ServiceName
	// +optional
	ServiceURL string `json:"serviceUrl,omitempty"`
}

+k8s:openapi-gen=true Exactly one InferenceTarget field must be specified

func (*InferenceTarget) DeepCopy added in v0.9.0

func (in *InferenceTarget) DeepCopy() *InferenceTarget

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new InferenceTarget.

func (*InferenceTarget) DeepCopyInto added in v0.9.0

func (in *InferenceTarget) DeepCopyInto(out *InferenceTarget)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type ModelSpec

type ModelSpec struct {
	// Storage URI for the model repository
	StorageURI string `json:"storageUri"`
	// Machine Learning <framework name>
	// The values could be: "tensorflow","pytorch","sklearn","onnx","xgboost", "myawesomeinternalframework" etc.
	Framework string `json:"framework"`
	// Maximum memory this model will consume, this field is used to decide if a model server has enough memory to load this model.
	Memory resource.Quantity `json:"memory"`
}

ModelSpec describes a TrainedModel +k8s:openapi-gen=true

func (*ModelSpec) DeepCopy

func (in *ModelSpec) DeepCopy() *ModelSpec

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ModelSpec.

func (*ModelSpec) DeepCopyInto

func (in *ModelSpec) DeepCopyInto(out *ModelSpec)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type ScaleMetric added in v0.12.0

type ScaleMetric string

ScaleMetric enum +kubebuilder:validation:Enum=cpu;memory;concurrency;rps

type ServerType added in v0.8.0

type ServerType string

ServerType constant for specifying the runtime name +k8s:openapi-gen=true

const (
	// Model server is Triton
	Triton ServerType = "triton"
	// Model server is MLServer
	MLServer ServerType = "mlserver"
	// Model server is OpenVino Model Server
	OVMS ServerType = "ovms"
)

Built-in ServerTypes (others may be supported)

type ServingRuntime added in v0.8.0

type ServingRuntime struct {
	metav1.TypeMeta   `json:",inline"`
	metav1.ObjectMeta `json:"metadata,omitempty"`

	Spec   ServingRuntimeSpec   `json:"spec,omitempty"`
	Status ServingRuntimeStatus `json:"status,omitempty"`
}

ServingRuntime is the Schema for the servingruntimes API +k8s:openapi-gen=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object +genclient +kubebuilder:object:root=true +kubebuilder:printcolumn:name="Disabled",type="boolean",JSONPath=".spec.disabled" +kubebuilder:printcolumn:name="ModelType",type="string",JSONPath=".spec.supportedModelFormats[*].name" +kubebuilder:printcolumn:name="Containers",type="string",JSONPath=".spec.containers[*].name" +kubebuilder:printcolumn:name="Age",type="date",JSONPath=".metadata.creationTimestamp"

func (*ServingRuntime) DeepCopy added in v0.8.0

func (in *ServingRuntime) DeepCopy() *ServingRuntime

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ServingRuntime.

func (*ServingRuntime) DeepCopyInto added in v0.8.0

func (in *ServingRuntime) DeepCopyInto(out *ServingRuntime)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*ServingRuntime) DeepCopyObject added in v0.8.0

func (in *ServingRuntime) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

type ServingRuntimeList added in v0.8.0

type ServingRuntimeList struct {
	metav1.TypeMeta `json:",inline"`
	metav1.ListMeta `json:"metadata,omitempty"`
	Items           []ServingRuntime `json:"items"`
}

ServingRuntimeList contains a list of ServingRuntime +k8s:openapi-gen=true +kubebuilder:object:root=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object

func (*ServingRuntimeList) DeepCopy added in v0.8.0

func (in *ServingRuntimeList) DeepCopy() *ServingRuntimeList

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ServingRuntimeList.

func (*ServingRuntimeList) DeepCopyInto added in v0.8.0

func (in *ServingRuntimeList) DeepCopyInto(out *ServingRuntimeList)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*ServingRuntimeList) DeepCopyObject added in v0.8.0

func (in *ServingRuntimeList) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

type ServingRuntimePodSpec added in v0.8.0

type ServingRuntimePodSpec struct {
	// List of containers belonging to the pod.
	// Containers cannot currently be added or removed.
	// There must be at least one container in a Pod.
	// Cannot be updated.
	// +patchMergeKey=name
	// +patchStrategy=merge
	Containers []corev1.Container `json:"containers" patchStrategy:"merge" patchMergeKey:"name" validate:"required"`

	// List of volumes that can be mounted by containers belonging to the pod.
	// More info: https://kubernetes.io/docs/concepts/storage/volumes
	// +optional
	// +patchMergeKey=name
	// +patchStrategy=merge,retainKeys
	Volumes []corev1.Volume `json:"volumes,omitempty" patchStrategy:"merge,retainKeys" patchMergeKey:"name" protobuf:"bytes,1,rep,name=volumes"`

	// NodeSelector is a selector which must be true for the pod to fit on a node.
	// Selector which must match a node's labels for the pod to be scheduled on that node.
	// More info: https://kubernetes.io/docs/concepts/configuration/assign-pod-node/
	// +optional
	NodeSelector map[string]string `json:"nodeSelector,omitempty"`

	// If specified, the pod's scheduling constraints
	// +optional
	Affinity *corev1.Affinity `json:"affinity,omitempty"`

	// If specified, the pod's tolerations.
	// +optional
	Tolerations []corev1.Toleration `json:"tolerations,omitempty"`

	// Labels that will be add to the pod.
	// More info: http://kubernetes.io/docs/user-guide/labels
	// +optional
	Labels map[string]string `json:"labels,omitempty"`

	// Annotations that will be add to the pod.
	// More info: http://kubernetes.io/docs/user-guide/annotations
	// +optional
	Annotations map[string]string `json:"annotations,omitempty"`
	// ImagePullSecrets is an optional list of references to secrets in the same namespace to use for pulling any of the images used by this PodSpec.
	// If specified, these secrets will be passed to individual puller implementations for them to use. For example,
	// in the case of docker, only DockerConfig type secrets are honored.
	// More info: https://kubernetes.io/docs/concepts/containers/images#specifying-imagepullsecrets-on-a-pod
	// +optional
	// +patchMergeKey=name
	// +patchStrategy=merge
	ImagePullSecrets []corev1.LocalObjectReference `json:"imagePullSecrets,omitempty" patchStrategy:"merge" patchMergeKey:"name" protobuf:"bytes,15,rep,name=imagePullSecrets"`
}

+k8s:openapi-gen=true

func (*ServingRuntimePodSpec) DeepCopy added in v0.8.0

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ServingRuntimePodSpec.

func (*ServingRuntimePodSpec) DeepCopyInto added in v0.8.0

func (in *ServingRuntimePodSpec) DeepCopyInto(out *ServingRuntimePodSpec)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type ServingRuntimeSpec added in v0.8.0

type ServingRuntimeSpec struct {
	// Model formats and version supported by this runtime
	SupportedModelFormats []SupportedModelFormat `json:"supportedModelFormats,omitempty"`

	// Whether this ServingRuntime is intended for multi-model usage or not.
	// +optional
	MultiModel *bool `json:"multiModel,omitempty"`

	// Set to true to disable use of this runtime
	// +optional
	Disabled *bool `json:"disabled,omitempty"`

	// Supported protocol versions (i.e. v1 or v2 or grpc-v1 or grpc-v2)
	// +optional
	ProtocolVersions []constants.InferenceServiceProtocol `json:"protocolVersions,omitempty"`

	ServingRuntimePodSpec `json:",inline"`

	// Grpc endpoint for internal model-management (implementing mmesh.ModelRuntime gRPC service)
	// Assumed to be single-model runtime if omitted
	// +optional
	GrpcMultiModelManagementEndpoint *string `json:"grpcEndpoint,omitempty"`

	// Grpc endpoint for inferencing
	// +optional
	GrpcDataEndpoint *string `json:"grpcDataEndpoint,omitempty"`
	// HTTP endpoint for inferencing
	// +optional
	HTTPDataEndpoint *string `json:"httpDataEndpoint,omitempty"`

	// Configure the number of replicas in the Deployment generated by this ServingRuntime
	// If specified, this overrides the podsPerRuntime configuration value
	// +optional
	Replicas *uint16 `json:"replicas,omitempty"`

	// Configuration for this runtime's use of the storage helper (model puller)
	// It is enabled unless explicitly disabled
	// +optional
	StorageHelper *StorageHelper `json:"storageHelper,omitempty"`

	// Provide the details about built-in runtime adapter
	// +optional
	BuiltInAdapter *BuiltInAdapter `json:"builtInAdapter,omitempty"`
}

ServingRuntimeSpec defines the desired state of ServingRuntime. This spec is currently provisional and are subject to change as details regarding single-model serving and multi-model serving are hammered out. +k8s:openapi-gen=true

func (*ServingRuntimeSpec) DeepCopy added in v0.8.0

func (in *ServingRuntimeSpec) DeepCopy() *ServingRuntimeSpec

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ServingRuntimeSpec.

func (*ServingRuntimeSpec) DeepCopyInto added in v0.8.0

func (in *ServingRuntimeSpec) DeepCopyInto(out *ServingRuntimeSpec)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*ServingRuntimeSpec) GetPriority added in v0.11.1

func (srSpec *ServingRuntimeSpec) GetPriority(modelName string) *int32

GetPriority returns the priority of the specified model. It returns nil if priority is not set or the model is not found.

func (*ServingRuntimeSpec) IsDisabled added in v0.8.0

func (srSpec *ServingRuntimeSpec) IsDisabled() bool

func (*ServingRuntimeSpec) IsMultiModelRuntime added in v0.8.0

func (srSpec *ServingRuntimeSpec) IsMultiModelRuntime() bool

func (*ServingRuntimeSpec) IsProtocolVersionSupported added in v0.9.0

func (srSpec *ServingRuntimeSpec) IsProtocolVersionSupported(modelProtocolVersion constants.InferenceServiceProtocol) bool

type ServingRuntimeStatus added in v0.8.0

type ServingRuntimeStatus struct {
}

ServingRuntimeStatus defines the observed state of ServingRuntime +k8s:openapi-gen=true

func (*ServingRuntimeStatus) DeepCopy added in v0.8.0

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ServingRuntimeStatus.

func (*ServingRuntimeStatus) DeepCopyInto added in v0.8.0

func (in *ServingRuntimeStatus) DeepCopyInto(out *ServingRuntimeStatus)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type StorageContainerSpec added in v0.11.1

type StorageContainerSpec struct {
	// Container spec for the storage initializer init container
	Container corev1.Container `json:"container" validate:"required"`

	// List of URI formats that this container supports
	SupportedUriFormats []SupportedUriFormat `json:"supportedUriFormats" validate:"required"`
}

StorageContainerSpec defines the container spec for the storage initializer init container, and the protocols it supports. +k8s:openapi-gen=true

func (*StorageContainerSpec) DeepCopy added in v0.11.1

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new StorageContainerSpec.

func (*StorageContainerSpec) DeepCopyInto added in v0.11.1

func (in *StorageContainerSpec) DeepCopyInto(out *StorageContainerSpec)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*StorageContainerSpec) IsStorageUriSupported added in v0.11.1

func (spec *StorageContainerSpec) IsStorageUriSupported(storageUri string) (bool, error)

type StorageHelper added in v0.8.0

type StorageHelper struct {
	// +optional
	Disabled bool `json:"disabled,omitempty"`
}

+k8s:openapi-gen=true

func (*StorageHelper) DeepCopy added in v0.8.0

func (in *StorageHelper) DeepCopy() *StorageHelper

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new StorageHelper.

func (*StorageHelper) DeepCopyInto added in v0.8.0

func (in *StorageHelper) DeepCopyInto(out *StorageHelper)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type SupportedModelFormat added in v0.8.0

type SupportedModelFormat struct {
	// Name of the model format.
	// +required
	Name string `json:"name"`
	// Version of the model format.
	// Used in validating that a predictor is supported by a runtime.
	// Can be "major", "major.minor" or "major.minor.patch".
	// +optional
	Version *string `json:"version,omitempty"`
	// Set to true to allow the ServingRuntime to be used for automatic model placement if
	// this model format is specified with no explicit runtime.
	// +optional
	AutoSelect *bool `json:"autoSelect,omitempty"`

	// Priority of this serving runtime for auto selection.
	// This is used to select the serving runtime if more than one serving runtime supports the same model format.
	// The value should be greater than zero.  The higher the value, the higher the priority.
	// Priority is not considered if AutoSelect is either false or not specified.
	// Priority can be overridden by specifying the runtime in the InferenceService.
	// +optional
	Priority *int32 `json:"priority,omitempty"`
}

+k8s:openapi-gen=true

func (*SupportedModelFormat) DeepCopy added in v0.8.0

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new SupportedModelFormat.

func (*SupportedModelFormat) DeepCopyInto added in v0.8.0

func (in *SupportedModelFormat) DeepCopyInto(out *SupportedModelFormat)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*SupportedModelFormat) IsAutoSelectEnabled added in v0.11.1

func (m *SupportedModelFormat) IsAutoSelectEnabled() bool

type SupportedRuntime added in v0.9.0

type SupportedRuntime struct {
	Name string
	Spec ServingRuntimeSpec
}

SupportedRuntime is the schema for supported runtime result of automatic selection

func (*SupportedRuntime) DeepCopy added in v0.9.0

func (in *SupportedRuntime) DeepCopy() *SupportedRuntime

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new SupportedRuntime.

func (*SupportedRuntime) DeepCopyInto added in v0.9.0

func (in *SupportedRuntime) DeepCopyInto(out *SupportedRuntime)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type SupportedUriFormat added in v0.11.1

type SupportedUriFormat struct {
	Prefix string `json:"prefix,omitempty"`
	Regex  string `json:"regex,omitempty"`
}

SupportedUriFormat can be either prefix or regex. Todo: Add validation that only one of them is set. +k8s:openapi-gen=true

func (*SupportedUriFormat) DeepCopy added in v0.11.1

func (in *SupportedUriFormat) DeepCopy() *SupportedUriFormat

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new SupportedUriFormat.

func (*SupportedUriFormat) DeepCopyInto added in v0.11.1

func (in *SupportedUriFormat) DeepCopyInto(out *SupportedUriFormat)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type TrainedModel added in v0.7.0

type TrainedModel struct {
	metav1.TypeMeta   `json:",inline"`
	metav1.ObjectMeta `json:"metadata,omitempty"`
	Spec              TrainedModelSpec   `json:"spec,omitempty"`
	Status            TrainedModelStatus `json:"status,omitempty"`
}

TrainedModel is the Schema for the TrainedModel API +k8s:openapi-gen=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object +genclient +kubebuilder:object:root=true +kubebuilder:subresource:status +kubebuilder:printcolumn:name="URL",type="string",JSONPath=".status.url" +kubebuilder:printcolumn:name="Ready",type="string",JSONPath=".status.conditions[?(@.type=='Ready')].status" +kubebuilder:printcolumn:name="Age",type="date",JSONPath=".metadata.creationTimestamp" +kubebuilder:resource:path=trainedmodels,shortName=tm,singular=trainedmodel

func (*TrainedModel) DeepCopy added in v0.7.0

func (in *TrainedModel) DeepCopy() *TrainedModel

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new TrainedModel.

func (*TrainedModel) DeepCopyInto added in v0.7.0

func (in *TrainedModel) DeepCopyInto(out *TrainedModel)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*TrainedModel) DeepCopyObject added in v0.7.0

func (in *TrainedModel) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

func (*TrainedModel) ValidateCreate added in v0.7.0

func (tm *TrainedModel) ValidateCreate() (admission.Warnings, error)

ValidateCreate implements webhook.Validator so a webhook will be registered for the type

func (*TrainedModel) ValidateDelete added in v0.7.0

func (tm *TrainedModel) ValidateDelete() (admission.Warnings, error)

ValidateDelete implements webhook.Validator so a webhook will be registered for the type

func (*TrainedModel) ValidateUpdate added in v0.7.0

func (tm *TrainedModel) ValidateUpdate(old runtime.Object) (admission.Warnings, error)

ValidateUpdate implements webhook.Validator so a webhook will be registered for the type

type TrainedModelList added in v0.7.0

type TrainedModelList struct {
	metav1.TypeMeta `json:",inline"`
	metav1.ListMeta `json:"metadata,omitempty"`
	// +listType=set
	Items []TrainedModel `json:"items"`
}

TrainedModelList contains a list of TrainedModel +k8s:openapi-gen=true +kubebuilder:object:root=true +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object

func (*TrainedModelList) DeepCopy added in v0.7.0

func (in *TrainedModelList) DeepCopy() *TrainedModelList

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new TrainedModelList.

func (*TrainedModelList) DeepCopyInto added in v0.7.0

func (in *TrainedModelList) DeepCopyInto(out *TrainedModelList)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*TrainedModelList) DeepCopyObject added in v0.7.0

func (in *TrainedModelList) DeepCopyObject() runtime.Object

DeepCopyObject is an autogenerated deepcopy function, copying the receiver, creating a new runtime.Object.

func (*TrainedModelList) TotalRequestedMemory added in v0.7.0

func (tms *TrainedModelList) TotalRequestedMemory() resource.Quantity

type TrainedModelSpec added in v0.7.0

type TrainedModelSpec struct {
	// parent inference service to deploy to
	// +required
	InferenceService string `json:"inferenceService"`
	// Predictor model spec
	// +required
	Model ModelSpec `json:"model"`
}

TrainedModelSpec defines the TrainedModel spec +k8s:openapi-gen=true

func (*TrainedModelSpec) DeepCopy added in v0.7.0

func (in *TrainedModelSpec) DeepCopy() *TrainedModelSpec

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new TrainedModelSpec.

func (*TrainedModelSpec) DeepCopyInto added in v0.7.0

func (in *TrainedModelSpec) DeepCopyInto(out *TrainedModelSpec)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

type TrainedModelStatus added in v0.7.0

type TrainedModelStatus struct {
	// Conditions for trained model
	duckv1.Status `json:",inline"`
	// URL holds the url that will distribute traffic over the provided traffic targets.
	// For v1: http[s]://{route-name}.{route-namespace}.{cluster-level-suffix}/v1/models/<trainedmodel>:predict
	// For v2: http[s]://{route-name}.{route-namespace}.{cluster-level-suffix}/v2/models/<trainedmodel>/infer
	URL *apis.URL `json:"url,omitempty"`
	// Addressable endpoint for the deployed trained model
	// http://<inferenceservice.metadata.name>/v1/models/<trainedmodel>.metadata.name
	Address *duckv1.Addressable `json:"address,omitempty"`
}

TrainedModelStatus defines the observed state of TrainedModel

func (*TrainedModelStatus) DeepCopy added in v0.7.0

func (in *TrainedModelStatus) DeepCopy() *TrainedModelStatus

DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new TrainedModelStatus.

func (*TrainedModelStatus) DeepCopyInto added in v0.7.0

func (in *TrainedModelStatus) DeepCopyInto(out *TrainedModelStatus)

DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.

func (*TrainedModelStatus) GetCondition added in v0.7.0

func (ss *TrainedModelStatus) GetCondition(t apis.ConditionType) *apis.Condition

GetCondition returns the condition by name.

func (*TrainedModelStatus) InitializeConditions added in v0.7.0

func (ss *TrainedModelStatus) InitializeConditions()

func (*TrainedModelStatus) IsConditionReady added in v0.7.0

func (ss *TrainedModelStatus) IsConditionReady(t apis.ConditionType) bool

IsConditionReady returns the readiness for a given condition

func (*TrainedModelStatus) IsReady added in v0.7.0

func (ss *TrainedModelStatus) IsReady() bool

IsReady returns if the service is ready to serve the requested configuration.

func (*TrainedModelStatus) SetCondition added in v0.7.0

func (ss *TrainedModelStatus) SetCondition(conditionType apis.ConditionType, condition *apis.Condition)

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL