renderer

Documentation

Index

• func ToYaml(v interface{}) (string, error)
• func YamlToObject(yaml string) (objs []runtime.Object, err error)
• type DefaultEnhancer
  • func (de *DefaultEnhancer) Apply(templates map[string]string, metadata Metadata) (objsToAdd []runtime.Object, err error)
• type Engine
  • func New() *Engine
  • func (e *Engine) Render(tplName string, tpl string, vals map[string]interface{}) (string, error)
  • func (e Engine) Template(name string) *template.Template
• type Enhancer
• type Metadata

Functions

func ToYaml

func ToYaml(v interface{}) (string, error)

ToYaml takes any value, and returns its YAML representation as a string.

func YamlToObject

func YamlToObject(yaml string) (objs []runtime.Object, err error)

YamlToObject parses a list of runtime.Objects from the provided yaml If the type is not known in the scheme, it tries to parse it as Unstructured We used to use 'apimachiner/pkg/util/yaml' for splitting the input into multiple yamls, however under the covers it uses bufio.NewScanner with token defaults with no option to modify. see: https://github.com/kubernetes/apimachinery/blob/release-1.6/pkg/util/yaml/decoder.go#L94 The YAML input can be too large for the default scan token size used by these packages. For more detail read: https://github.com/kudobuilder/kudo/pull/1400 TODO(av) could we use something else than a global scheme here? Should we somehow inject it?

Types

type DefaultEnhancer

type DefaultEnhancer struct {
	Scheme    *runtime.Scheme
	Discovery discovery.CachedDiscoveryInterface
}

DefaultEnhancer is implementation of Enhancer that applies the defined conventions by directly editing runtime.Objects (Unstructured).

func (*DefaultEnhancer) Apply

func (de *DefaultEnhancer) Apply(templates map[string]string, metadata Metadata) (objsToAdd []runtime.Object, err error)

Apply accepts templates to be rendered in kubernetes and enhances them with our own KUDO conventions These include the way we name our objects and what labels we apply to them

type Engine

type Engine struct {
	FuncMap template.FuncMap
}

Engine is the control struct for parsing and templating Kubernetes resources in an ordered fashion

func New

func New() *Engine

New creates an engine with a default function map, using a modified Sprig func map. Because these templates are rendered by the operator, we delete any functions that potentially access the environment the controller is running in.

func (*Engine) Render

func (e *Engine) Render(tplName string, tpl string, vals map[string]interface{}) (string, error)

Render creates a fully rendered template based on a set of values. It parses these in strict mode, returning errors when keys are missing.

func (Engine) Template

func (e Engine) Template(name string) *template.Template

Template provides access to the engines template engine.

type Enhancer

type Enhancer interface {
	Apply(templates map[string]string, metadata Metadata) ([]runtime.Object, error)
}

Enhancer takes your kubernetes template and kudo related Metadata and applies them to all resources in form of labels and annotations it also takes care of setting an owner of all the resources to the provided object

type Metadata

type Metadata struct {
	engine.Metadata

	PlanName  string
	PlanUID   types.UID
	PhaseName string
	StepName  string
	TaskName  string
}

Metadata contains Metadata along with specific fields associated with current plan being executed like current plan, phase, step or task names.