Pipeline component

YAML the Kubeflow Pipelines' component.yaml files which describe a pipeline components. Component consists of input/output definitions and the description of the implementation which can either be a containerized command line program or an interconnected graph of tasks. Documentation: https://cloud-pipelines.github.io/links/component_authoring_documentation

File match component.yaml kfp_component.yaml kfp_component.json
Schema URL https://catalog.lintel.tools/schemas/schemastore/pipeline-component/latest.json
Source https://raw.githubusercontent.com/Cloud-Pipelines/component_spec_schema/stable/component_spec.json_schema.json

Validate with Lintel

npx @lintel/lintel check

All of

1. ComponentSpec object

Definitions

TypeSpecType string | object
InputSpec object

Describes the component input specification

name string required
type string | object
description string
default string
optional boolean
Default: false
annotations object
OutputSpec object

Describes the component output specification

name string required
type string | object
description string
annotations object
InputValuePlaceholder object

Represents the command-line argument placeholder that will be replaced at run-time by the input argument value.

inputValue string required

Name of the input.

InputPathPlaceholder object

Represents the command-line argument placeholder that will be replaced at run-time by a local file path pointing to a file containing the input argument value.

inputPath string required

Name of the input.

OutputPathPlaceholder object

Represents the command-line argument placeholder that will be replaced at run-time by a local file path pointing to a file where the program should write its output data.

outputPath string required

Name of the output.

StringOrPlaceholder string | InputValuePlaceholder | InputPathPlaceholder | OutputPathPlaceholder | ConcatPlaceholder | IfPlaceholder
ConcatPlaceholder object

Represents the command-line argument placeholder that will be replaced at run-time by the concatenated values of its items.

concat StringOrPlaceholder[] required

Items to concatenate

IsPresentPlaceholder object

Represents the command-line argument placeholder that will be replaced at run-time by a boolean value specifying whether the caller has passed an argument for the specified optional input.

isPresent string required

Name of the input.

IfConditionArgumentType IsPresentPlaceholder | boolean | string | InputValuePlaceholder
ListOfStringsOrPlaceholders StringOrPlaceholder[]
IfPlaceholder object

Represents the command-line argument placeholder that will be replaced at run-time by a boolean value specifying whether the caller has passed an argument for the specified optional input.

if object required
3 nested properties
cond IsPresentPlaceholder | boolean | string | InputValuePlaceholder required
then StringOrPlaceholder[] required
else StringOrPlaceholder[]
ContainerSpec object
image string | InputValuePlaceholder | InputPathPlaceholder | OutputPathPlaceholder | ConcatPlaceholder | IfPlaceholder required
command StringOrPlaceholder[]

Entrypoint array. Not executed within a shell. The docker image's ENTRYPOINT is used if this is not provided.

args StringOrPlaceholder[]

Arguments to the entrypoint. The docker image's CMD is used if this is not provided.

env Record<string, string | InputValuePlaceholder | InputPathPlaceholder | OutputPathPlaceholder | ConcatPlaceholder | IfPlaceholder>

List of environment variables to set in the container.

ContainerImplementation object

Represents the container component implementation.

container object required
4 nested properties
image string | InputValuePlaceholder | InputPathPlaceholder | OutputPathPlaceholder | ConcatPlaceholder | IfPlaceholder required
command StringOrPlaceholder[]

Entrypoint array. Not executed within a shell. The docker image's ENTRYPOINT is used if this is not provided.

args StringOrPlaceholder[]

Arguments to the entrypoint. The docker image's CMD is used if this is not provided.

env Record<string, string | InputValuePlaceholder | InputPathPlaceholder | OutputPathPlaceholder | ConcatPlaceholder | IfPlaceholder>

List of environment variables to set in the container.

ImplementationType ContainerImplementation | GraphImplementation
MetadataSpec object
annotations object
ComponentSpec object

Component specification. Describes the metadata (name, description, source), the interface (inputs and outputs) and the implementation of the component.

implementation ContainerImplementation | GraphImplementation required
name string
description string
inputs InputSpec[]
outputs OutputSpec[]
metadata object
1 nested properties
annotations object
ComponentReference object

Component reference. Contains information that can be used to locate and load a component by name, digest or URL

name string
digest string
tag string
url string
text string
spec object

Component specification. Describes the metadata (name, description, source), the interface (inputs and outputs) and the implementation of the component.

6 nested properties
implementation ContainerImplementation | GraphImplementation required
name string
description string
inputs InputSpec[]
outputs OutputSpec[]
metadata object
1 nested properties
annotations object
GraphInputArgument object

Represents the component argument value that comes from the graph component input.

graphInput object required

References the input of the graph/pipeline.

2 nested properties
inputName string required
type string | object
TaskOutputArgument object

Represents the component argument value that comes from the output of a sibling task.

taskOutput object required

References the output of a sibling task.

3 nested properties
taskId string required
outputName string required
type string | object
ArgumentType string | GraphInputArgument | TaskOutputArgument
TwoArgumentOperands object

Pair of operands for a binary operation.

op1 string | GraphInputArgument | TaskOutputArgument required
op2 string | GraphInputArgument | TaskOutputArgument required
TwoLogicalOperands object

Pair of operands for a binary logical operation.

op1 object | object | object | object | object | object | object | object | object required
op2 object | object | object | object | object | object | object | object | object required
PredicateType object | object | object | object | object | object | object | object | object
RetryStrategySpec object

Optional configuration that specifies how the task should be retried if it fails.

maxRetries integer
CachingStrategySpec object

Optional configuration that specifies how the task execution may be skipped if the output data exist in cache.

maxCacheStaleness string
format=duration
ExecutionOptionsSpec object

Optional configuration that specifies how the task should be executed. Can be used to set some platform-specific options.

retryStrategy object

Optional configuration that specifies how the task should be retried if it fails.

1 nested properties
maxRetries integer
cachingStrategy object

Optional configuration that specifies how the task execution may be skipped if the output data exist in cache.

1 nested properties
maxCacheStaleness string
format=duration
TaskSpec object

'Task specification. Task is a configured component - a component supplied with arguments and other applied configuration changes.

componentRef object required

Component reference. Contains information that can be used to locate and load a component by name, digest or URL

6 nested properties
name string
digest string
tag string
url string
text string
spec object

Component specification. Describes the metadata (name, description, source), the interface (inputs and outputs) and the implementation of the component.

6 nested properties
implementation ContainerImplementation | GraphImplementation required
name string
description string
inputs InputSpec[]
outputs OutputSpec[]
metadata object
arguments Record<string, string | GraphInputArgument | TaskOutputArgument>
isEnabled object | object | object | object | object | object | object | object | object
executionOptions object

Optional configuration that specifies how the task should be executed. Can be used to set some platform-specific options.

2 nested properties
retryStrategy object

Optional configuration that specifies how the task should be retried if it fails.

1 nested properties
maxRetries integer
cachingStrategy object

Optional configuration that specifies how the task execution may be skipped if the output data exist in cache.

1 nested properties
maxCacheStaleness string
format=duration
annotations object
GraphSpec object

Describes the graph component implementation. It represents a graph of component tasks connected to the upstream sources of data using the argument specifications. It also describes the sources of graph output values.

tasks Record<string, object> required
outputValues Record<string, object>
GraphImplementation object

Represents the graph component implementation.

graph object required

Describes the graph component implementation. It represents a graph of component tasks connected to the upstream sources of data using the argument specifications. It also describes the sources of graph output values.

2 nested properties
tasks Record<string, object> required
outputValues Record<string, object>
PipelineRunSpec object

The object that can be sent to the backend to start a new Run.

rootTask object required

'Task specification. Task is a configured component - a component supplied with arguments and other applied configuration changes.

5 nested properties
componentRef object required

Component reference. Contains information that can be used to locate and load a component by name, digest or URL

6 nested properties
name string
digest string
tag string
url string
text string
spec object

Component specification. Describes the metadata (name, description, source), the interface (inputs and outputs) and the implementation of the component.

arguments Record<string, string | GraphInputArgument | TaskOutputArgument>
isEnabled object | object | object | object | object | object | object | object | object
executionOptions object

Optional configuration that specifies how the task should be executed. Can be used to set some platform-specific options.

2 nested properties
retryStrategy object

Optional configuration that specifies how the task should be retried if it fails.

cachingStrategy object

Optional configuration that specifies how the task execution may be skipped if the output data exist in cache.

annotations object
onExitTask object

'Task specification. Task is a configured component - a component supplied with arguments and other applied configuration changes.

5 nested properties
componentRef object required

Component reference. Contains information that can be used to locate and load a component by name, digest or URL

6 nested properties
name string
digest string
tag string
url string
text string
spec object

Component specification. Describes the metadata (name, description, source), the interface (inputs and outputs) and the implementation of the component.

arguments Record<string, string | GraphInputArgument | TaskOutputArgument>
isEnabled object | object | object | object | object | object | object | object | object
executionOptions object

Optional configuration that specifies how the task should be executed. Can be used to set some platform-specific options.

2 nested properties
retryStrategy object

Optional configuration that specifies how the task should be retried if it fails.

cachingStrategy object

Optional configuration that specifies how the task execution may be skipped if the output data exist in cache.

annotations object