lsst.pipe.base

The lsst.pipe.base module provides base classes for the task framework. Tasks package the algorithmic units of the LSST Science Pipelines. You can create, configure, and run tasks with their Python APIs. Some tasks, called command-line tasks, are also packaged into data processing pipelines that you can run from the command line.

Changes

• pipe_base v23.0.1 2022-02-02
• pipe_base v23.0.0 2021-12-10
• pipe_base v22.0 (2021-04-01)

Using lsst.pipe.base

Overview

• Overview of the task framework

Using command-line tasks

• Using Butler data repositories and reruns with command-line tasks
• Specifying data IDs with command-line tasks
• Configuring command-line tasks
• Retargeting subtasks of command-line tasks
• Working with provenance checks in command-line tasks
• Logging with command-line tasks
• Parallel processing with command-line tasks
• Command-line task argument reference

Developing tasks and command-line tasks

• Creating a task
• Creating a command-line task
• Testing a pipeline task
• Creating a PipelineTask

Developing Pipelines

• Creating a Pipeline

Contributing

lsst.pipe.base is developed at https://github.com/lsst/pipe_base. You can find Jira issues for this module under the pipe_base component.

Python API reference

lsst.pipe.base Package

Functions

buildExecutionButler(butler, graph, …)	buildExecutionButler is a function that is responsible for exporting input QuantumGraphs into a new minimal Butler which only contains datasets specified by the QuantumGraph.
getTaskLogger([name, logger])	Get a logger compatible with LSST usage.
iterConnections(connections, connectionType, …)	Creates an iterator over the selected connections type which yields all the defined connections of that type.
logInfo(obj, prefix[, logLevel, metadata, …])	Log timer information to obj.metadata and obj.log.
timeMethod(*args, **kwargs)	Decorator to measure duration of a method.

Classes

AdjustQuantumHelper(inputs, …)	Helper class for calling PipelineTaskConnections.adjustQuantum.
ArgumentParser(name[, usage])	Argument parser for command-line tasks that is based on argparse.ArgumentParser.
ButlerInitializedTaskRunner(TaskClass, parsedCmd)	A TaskRunner for CmdLineTasks that require a butler keyword argument to be passed to their constructor.
ButlerQuantumContext(butler, quantum)	A Butler-like class specialized for a single quantum
CmdLineTask([config, name, parentTask, log])	Base class for command-line tasks: tasks that may be executed from the command-line.
ConfigDatasetType(name)	Dataset type specified by a config parameter.
ConfigFileAction(option_strings, dest[, …])	argparse action to load config overrides from one or more files.
ConfigValueAction(option_strings, dest[, …])	argparse action callback to override config parameters using name=value pairs from the command-line.
DataIdContainer([level])	Container for data IDs and associated data references.
DatasetArgument([name, help, default])	Dataset type specified by a command-line argument.
DeferredDatasetRef	Class which denotes that a datasetRef should be treated as deferred when interacting with the butler
GraphBuilder(registry[, skipExistingIn, …])	GraphBuilder class is responsible for building task execution graph from a Pipeline.
IncompatibleGraphError	Exception class to indicate that a lookup by NodeId is impossible due to incompatibilities
InputOnlyArgumentParser(name[, usage])	ArgumentParser for command-line tasks that don’t write any output.
InputQuantizedConnection(**kwargs)
InvalidQuantumError	Exception that may be raised by PipelineTasks (and code they delegate to) in order to indicate logic bug or configuration problem.
LabelSpecifier(labels, begin, end)	A structure to specify a subset of labels to load
LegacyTaskRunner(TaskClass, parsedCmd[, …])	A TaskRunner for CmdLineTasks which calls the Task’s run method on a dataRef rather than the runDataRef method.
NoWorkFound	An exception raised when a Quantum should not exist because there is no work for it to do.
NodeId(number, buildId)	Deprecated, this class is used with QuantumGraph save formats of 1 and 2 when unpicking objects and must be retained until those formats are considered unloadable.
OutputQuantizedConnection(**kwargs)
Pipeline(description)	A Pipeline is a representation of a series of tasks to run, and the configuration for those tasks.
PipelineDatasetTypes(initInputs, …)	An immutable struct that classifies the dataset types used in a Pipeline.
PipelineTask(*[, config, log, initInputs])	Base class for all pipeline tasks.
PipelineTaskConfig	Configuration class for PipelineTask
PipelineTaskConnections(*, config)	PipelineTaskConnections is a class used to declare desired IO when a PipelineTask is run by an activator
QuantumGraph(quanta, …)	QuantumGraph is a directed acyclic graph of QuantumNode objects
QuantumNode(quantum, taskDef, nodeId)	This class represents a node in the quantum graph.
RepeatableQuantumError	Exception that may be raised by PipelineTasks (and code they delegate to) in order to indicate that a repeatable problem that will not be addressed by retries.
ResourceConfig	Configuration for resource requirements.
ScalarError	Exception raised when dataset type is configured as scalar but there are multiple data IDs in a Quantum for that dataset.
Struct(**keyArgs)	A container to which you can add fields as attributes.
Task([config, name, parentTask, log])	Base class for data processing tasks.
TaskDatasetTypes(initInputs, initOutputs, …)	An immutable struct that extracts and classifies the dataset types used by a PipelineTask
TaskDef([taskName, config, taskClass, label])	TaskDef is a collection of information about task needed by Pipeline.
TaskError	Use to report errors for which a traceback is not useful.
TaskFactory	Abstract base class for task factory.
TaskMetadata	Dict-like object for storing task metadata.
TaskRunner(TaskClass, parsedCmd[, …])	Run a command-line task, using multiprocessing if requested.

Class Inheritance Diagram

lsst.pipe.base.testUtils Module

Functions

assertValidInitOutput(task)	Test that a constructed task conforms to its own init-connections.
assertValidOutput(task, result)	Test that the output of a call to run conforms to its own connections.
getInitInputs(butler, config)	Return the initInputs object that would have been passed to a PipelineTask constructor.
lintConnections(connections, *[, …])	Inspect a connections class for common errors.
makeQuantum(task, butler, dataId, ioDataIds)	Create a Quantum for a particular data ID(s).
runTestQuantum(task, butler, quantum[, mockRun])	Run a PipelineTask on a Quantum.

lsst.pipe.base.connectionTypes Module

Classes

InitInput(name, storageClass, doc, multiple)
InitOutput(name, storageClass, doc, multiple)
Input(name, storageClass, doc, multiple, …)
PrerequisiteInput(name, storageClass, doc, …)	Class used for declaring PipelineTask prerequisite connections
Output(name, storageClass, doc, multiple, …)
BaseConnection(name, storageClass, doc, multiple)	Base class used for declaring PipelineTask connections

Class Inheritance Diagram

lsst.pipe.base.pipelineIR Module

Classes

ConfigIR(python, dataId, file, rest)	Intermediate representation of configurations read from a pipeline yaml file.
ContractError	An exception that is raised when a pipeline contract is not satisfied
ContractIR(contract, msg)	Intermediate representation of configuration contracts read from a pipeline yaml file.
ImportIR(location, include, exclude, …)	An intermediate representation of imported pipelines
PipelineIR(loaded_yaml)	Intermediate representation of a pipeline definition
TaskIR(label, klass, config)	Intermediate representation of tasks read from a pipeline yaml file.
LabeledSubset(label, subset, description)	Intermediate representation of named subset of task labels read from a pipeline yaml file.

Class Inheritance Diagram