PipelineIR

class lsst.pipe.base.pipelineIR.PipelineIR(loaded_yaml: Dict[str, Any])

Bases: object

Intermediate representation of a pipeline definition

Parameters:

loaded_yaml : dict

A dictionary which matches the structure that would be produced by a yaml reader which parses a pipeline definition document

Raises:

ValueError

Raised if:

• a pipeline is declared without a description;
• no tasks are declared in a pipeline, and no pipelines are to be inherited;
• more than one instrument is specified;
• more than one inherited pipeline share a label.

Methods Summary

from_string(pipeline_string)	Create a PipelineIR object from a string formatted like a pipeline document
from_uri(uri, urllib.parse.ParseResult, …)	Create a PipelineIR object from the document specified by the input uri.
merge_pipelines(pipelines)	Merge one or more other PipelineIR objects into this object.
reorder_tasks(task_labels)	Changes the order tasks are stored internally.
subset_from_labels(labelSpecifier)	Subset a pipelineIR to contain only labels specified in labelSpecifier.
to_primitives()	Convert to a representation used in yaml serialization
write_to_uri(uri, urllib.parse.ParseResult, …)	Serialize this PipelineIR object into a yaml formatted string and write the output to a file at the specified uri.

Methods Documentation

classmethod from_string(pipeline_string: str) → lsst.pipe.base.pipelineIR.PipelineIR

Create a PipelineIR object from a string formatted like a pipeline document

Parameters:

pipeline_string : str

A string that is formatted according like a pipeline document

classmethod from_uri(uri: Union[str, urllib.parse.ParseResult, lsst.resources._resourcePath.ResourcePath, pathlib.Path]) → lsst.pipe.base.pipelineIR.PipelineIR

Create a PipelineIR object from the document specified by the input uri.

Parameters:

uri: convertible to `ResourcePath`

Location of document to use in creating a PipelineIR object.

Returns:

pipelineIR : PipelineIR

The loaded pipeline

merge_pipelines(pipelines: Iterable[lsst.pipe.base.pipelineIR.PipelineIR]) → None

Merge one or more other PipelineIR objects into this object.

Parameters:

pipelines : Iterable of PipelineIR objects

An Iterable that contains one or more PipelineIR objects to merge into this object.

Raises:

ValueError

Raised if there is a conflict in instrument specifications. Raised if a task label appears in more than one of the input PipelineIR objects which are to be merged. Raised if a labeled subset appears in more than one of the input PipelineIR objects which are to be merged, and with any subset existing in this object.

reorder_tasks(task_labels: List[str]) → None

Changes the order tasks are stored internally. Useful for determining the order things will appear in the serialized (or printed) form.

Parameters:

task_labels : list of str

A list corresponding to all the labels in the pipeline inserted in the order the tasks are to be stored.

Raises:

KeyError

Raised if labels are supplied that are not in the pipeline, or if not all labels in the pipeline were supplied in task_labels input.

subset_from_labels(labelSpecifier: Set[str]) → lsst.pipe.base.pipelineIR.PipelineIR

Subset a pipelineIR to contain only labels specified in labelSpecifier.

Parameters:

labelSpecifier : set of str

Set containing labels that describes how to subset a pipeline.

Returns:

pipeline : PipelineIR

A new pipelineIR object that is a subset of the old pipelineIR

Raises:

ValueError

Raised if there is an issue with specified labels

Notes

This method attempts to prune any contracts that contain labels which are not in the declared subset of labels. This pruning is done using a string based matching due to the nature of contracts and may prune more than it should. Any labeled subsets defined that no longer have all members of the subset present in the pipeline will be removed from the resulting pipeline.

to_primitives() → Dict[str, Any]

Convert to a representation used in yaml serialization

write_to_uri(uri: Union[str, urllib.parse.ParseResult, lsst.resources._resourcePath.ResourcePath, pathlib.Path]) → None

Serialize this PipelineIR object into a yaml formatted string and write the output to a file at the specified uri.

Parameters:

uri: convertible to `ResourcePath`

Location of document to write a PipelineIR object.