v0.5.28 Latest Latest

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

Go to latest
Published: Mar 4, 2021 License: Apache-2.0 Imports: 3 Imported by: 0




This section is empty.


This section is empty.


This section is empty.


type ExecutionError

type ExecutionError struct {
	// Core error structure
	// Indicates if this error is recoverable
	IsRecoverable bool

Indicates any error in executing the task

type InputFilePaths

type InputFilePaths interface {
	// The inputs file path, minus the protobuf file name.
	GetInputPrefixPath() storage.DataReference
	// Gets a path for where the protobuf encoded inputs of type `core.LiteralMap` can be found. The returned value is an URN in the configured storage backend
	GetInputPath() storage.DataReference

If using Files for IO with tasks, then the input will be written to this path All the files are always created in a sandbox per execution

type InputReader

type InputReader interface {
	// Get the inputs for this task as a literal map, an error is returned only in case of systemic errors.
	// No outputs or void is indicated using *core.LiteralMap -> nil
	Get(ctx context.Context) (*core.LiteralMap, error)

InputReader provides a method to access the inputs for a task execution within the plugin's Task Context

type OutputFilePaths

type OutputFilePaths interface {
	// RawOutputPaths are available with OutputFilePaths

	// A path to a directory or prefix that contains all execution metadata for this execution
	GetOutputPrefixPath() storage.DataReference
	// A fully qualified path (URN) to where the framework expects the output to exist in the configured storage backend
	GetOutputPath() storage.DataReference
	// A Fully qualified path (URN) where the error information should be placed as a protobuf core.ErrorDocument. It is not directly
	// used by the framework, but could be used in the future
	GetErrorPath() storage.DataReference

All paths where various meta outputs produced by the task can be placed, such that the framework can directly access them. All paths are represented using storage.DataReference -> an URN for the configured storage backend

type OutputReader

type OutputReader interface {
	// Returns true if an error was detected when reading the output and false if no error was detected
	IsError(ctx context.Context) (bool, error)
	// Returns the error as type ExecutionError
	ReadError(ctx context.Context) (ExecutionError, error)
	// Returns true if the outputs are using the OutputFilePaths specified files. If so it allows the system to
	// optimize the reads of the files
	IsFile(ctx context.Context) bool
	// Returns true if the output exists false otherwise
	Exists(ctx context.Context) (bool, error)
	// Returns the output -> *core.LiteralMap (nil if void), *ExecutionError if user error when reading the output and error to indicate system problems
	Read(ctx context.Context) (*core.LiteralMap, *ExecutionError, error)


type OutputWriter

type OutputWriter interface {
	// Once the task completes, use this method to indicate the output accessor to the framework
	Put(ctx context.Context, reader OutputReader) error

Framework Output writing interface.

type RawOutputPaths added in v0.3.15

type RawOutputPaths interface {
	// This is prefix (blob store prefix or directory) where all data produced can be stored.
	GetRawOutputPrefix() storage.DataReference

RawOutputPaths is the actual path where the data produced by a task can be placed. It is completely optional. The advantage of using this path is to provide exactly once semantics. It is guaranteed that this path is unique for every new execution of a task (across retries etc) and is constant for a specific execution. As of 02/20/2020 Flytekit generates this path randomly for S3. This structure proposes migration of this logic to FlytePluginMachinery so that it can be used more universally outside of Flytekit.


Path Synopsis

Jump to

Keyboard shortcuts

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