Coverage for python/lsst/pipe/base/pipeline.py: 21%
439 statements
« prev ^ index » next coverage.py v7.2.7, created at 2023-06-11 02:00 -0700
« prev ^ index » next coverage.py v7.2.7, created at 2023-06-11 02:00 -0700
1# This file is part of pipe_base.
2#
3# Developed for the LSST Data Management System.
4# This product includes software developed by the LSST Project
5# (http://www.lsst.org).
6# See the COPYRIGHT file at the top-level directory of this distribution
7# for details of code ownership.
8#
9# This program is free software: you can redistribute it and/or modify
10# it under the terms of the GNU General Public License as published by
11# the Free Software Foundation, either version 3 of the License, or
12# (at your option) any later version.
13#
14# This program is distributed in the hope that it will be useful,
15# but WITHOUT ANY WARRANTY; without even the implied warranty of
16# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17# GNU General Public License for more details.
18#
19# You should have received a copy of the GNU General Public License
20# along with this program. If not, see <http://www.gnu.org/licenses/>.
21from __future__ import annotations
23"""Module defining Pipeline class and related methods.
24"""
26__all__ = ["Pipeline", "TaskDef", "TaskDatasetTypes", "PipelineDatasetTypes", "LabelSpecifier"]
28import copy
29import logging
30import re
31import urllib.parse
33# -------------------------------
34# Imports of standard modules --
35# -------------------------------
36from collections.abc import Callable, Generator, Iterable, Iterator, Mapping, Set
37from dataclasses import dataclass
38from types import MappingProxyType
39from typing import TYPE_CHECKING, ClassVar, cast
41# -----------------------------
42# Imports for other modules --
43from lsst.daf.butler import (
44 DataCoordinate,
45 DatasetType,
46 DimensionUniverse,
47 NamedValueSet,
48 Registry,
49 SkyPixDimension,
50)
51from lsst.resources import ResourcePath, ResourcePathExpression
52from lsst.utils import doImportType
53from lsst.utils.introspection import get_full_type_name
55from . import automatic_connection_constants as acc
56from . import pipelineIR, pipeTools
57from ._instrument import Instrument as PipeBaseInstrument
58from ._task_metadata import TaskMetadata
59from .config import PipelineTaskConfig
60from .connections import iterConnections
61from .connectionTypes import Input
62from .pipelineTask import PipelineTask
63from .task import _TASK_METADATA_TYPE
65if TYPE_CHECKING: # Imports needed only for type annotations; may be circular.
66 from lsst.obs.base import Instrument
67 from lsst.pex.config import Config
69# ----------------------------------
70# Local non-exported definitions --
71# ----------------------------------
73_LOG = logging.getLogger(__name__)
75# ------------------------
76# Exported definitions --
77# ------------------------
80@dataclass
81class LabelSpecifier:
82 """A structure to specify a subset of labels to load
84 This structure may contain a set of labels to be used in subsetting a
85 pipeline, or a beginning and end point. Beginning or end may be empty,
86 in which case the range will be a half open interval. Unlike python
87 iteration bounds, end bounds are *INCLUDED*. Note that range based
88 selection is not well defined for pipelines that are not linear in nature,
89 and correct behavior is not guaranteed, or may vary from run to run.
90 """
92 labels: set[str] | None = None
93 begin: str | None = None
94 end: str | None = None
96 def __post_init__(self) -> None:
97 if self.labels is not None and (self.begin or self.end):
98 raise ValueError(
99 "This struct can only be initialized with a labels set or a begin (and/or) end specifier"
100 )
103class TaskDef:
104 """TaskDef is a collection of information about task needed by Pipeline.
106 The information includes task name, configuration object and optional
107 task class. This class is just a collection of attributes and it exposes
108 all of them so that attributes could potentially be modified in place
109 (e.g. if configuration needs extra overrides).
111 Attributes
112 ----------
113 taskName : `str`, optional
114 The fully-qualified `PipelineTask` class name. If not provided,
115 ``taskClass`` must be.
116 config : `lsst.pipe.base.config.PipelineTaskConfig`, optional
117 Instance of the configuration class corresponding to this task class,
118 usually with all overrides applied. This config will be frozen. If
119 not provided, ``taskClass`` must be provided and
120 ``taskClass.ConfigClass()`` will be used.
121 taskClass : `type`, optional
122 `PipelineTask` class object; if provided and ``taskName`` is as well,
123 the caller guarantees that they are consistent. If not provided,
124 ``taskName`` is used to import the type.
125 label : `str`, optional
126 Task label, usually a short string unique in a pipeline. If not
127 provided, ``taskClass`` must be, and ``taskClass._DefaultName`` will
128 be used.
129 """
131 def __init__(
132 self,
133 taskName: str | None = None,
134 config: PipelineTaskConfig | None = None,
135 taskClass: type[PipelineTask] | None = None,
136 label: str | None = None,
137 ):
138 if taskName is None:
139 if taskClass is None:
140 raise ValueError("At least one of `taskName` and `taskClass` must be provided.")
141 taskName = get_full_type_name(taskClass)
142 elif taskClass is None:
143 taskClass = doImportType(taskName)
144 if config is None:
145 if taskClass is None:
146 raise ValueError("`taskClass` must be provided if `config` is not.")
147 config = taskClass.ConfigClass()
148 if label is None:
149 if taskClass is None:
150 raise ValueError("`taskClass` must be provided if `label` is not.")
151 label = taskClass._DefaultName
152 self.taskName = taskName
153 try:
154 config.validate()
155 except Exception:
156 _LOG.error("Configuration validation failed for task %s (%s)", label, taskName)
157 raise
158 config.freeze()
159 self.config = config
160 self.taskClass = taskClass
161 self.label = label
162 self.connections = config.connections.ConnectionsClass(config=config)
164 @property
165 def configDatasetName(self) -> str:
166 """Name of a dataset type for configuration of this task (`str`)"""
167 return acc.CONFIG_INIT_OUTPUT_TEMPLATE.format(label=self.label)
169 @property
170 def metadataDatasetName(self) -> str:
171 """Name of a dataset type for metadata of this task (`str`)"""
172 return self.makeMetadataDatasetName(self.label)
174 @classmethod
175 def makeMetadataDatasetName(cls, label: str) -> str:
176 """Construct the name of the dataset type for metadata for a task.
178 Parameters
179 ----------
180 label : `str`
181 Label for the task within its pipeline.
183 Returns
184 -------
185 name : `str`
186 Name of the task's metadata dataset type.
187 """
188 return acc.METADATA_OUTPUT_TEMPLATE.format(label=label)
190 @property
191 def logOutputDatasetName(self) -> str | None:
192 """Name of a dataset type for log output from this task, `None` if
193 logs are not to be saved (`str`)
194 """
195 if cast(PipelineTaskConfig, self.config).saveLogOutput:
196 return acc.LOG_OUTPUT_TEMPLATE.format(label=self.label)
197 else:
198 return None
200 def __str__(self) -> str:
201 rep = "TaskDef(" + self.taskName
202 if self.label:
203 rep += ", label=" + self.label
204 rep += ")"
205 return rep
207 def __eq__(self, other: object) -> bool:
208 if not isinstance(other, TaskDef):
209 return False
210 # This does not consider equality of configs when determining equality
211 # as config equality is a difficult thing to define. Should be updated
212 # after DM-27847
213 return self.taskClass == other.taskClass and self.label == other.label
215 def __hash__(self) -> int:
216 return hash((self.taskClass, self.label))
218 @classmethod
219 def _unreduce(cls, taskName: str, config: PipelineTaskConfig, label: str) -> TaskDef:
220 """Unpickle pickle. Custom callable for unpickling.
222 All arguments are forwarded directly to the constructor; this
223 trampoline is only needed because ``__reduce__`` callables can't be
224 called with keyword arguments.
225 """
226 return cls(taskName=taskName, config=config, label=label)
228 def __reduce__(self) -> tuple[Callable[[str, PipelineTaskConfig, str], TaskDef], tuple[str, Config, str]]:
229 return (self._unreduce, (self.taskName, self.config, self.label))
232class Pipeline:
233 """A `Pipeline` is a representation of a series of tasks to run, and the
234 configuration for those tasks.
236 Parameters
237 ----------
238 description : `str`
239 A description of that this pipeline does.
240 """
242 def __init__(self, description: str):
243 pipeline_dict = {"description": description, "tasks": {}}
244 self._pipelineIR = pipelineIR.PipelineIR(pipeline_dict)
246 @classmethod
247 def fromFile(cls, filename: str) -> Pipeline:
248 """Load a pipeline defined in a pipeline yaml file.
250 Parameters
251 ----------
252 filename: `str`
253 A path that points to a pipeline defined in yaml format. This
254 filename may also supply additional labels to be used in
255 subsetting the loaded Pipeline. These labels are separated from
256 the path by a ``#``, and may be specified as a comma separated
257 list, or a range denoted as beginning..end. Beginning or end may
258 be empty, in which case the range will be a half open interval.
259 Unlike python iteration bounds, end bounds are *INCLUDED*. Note
260 that range based selection is not well defined for pipelines that
261 are not linear in nature, and correct behavior is not guaranteed,
262 or may vary from run to run.
264 Returns
265 -------
266 pipeline: `Pipeline`
267 The pipeline loaded from specified location with appropriate (if
268 any) subsetting.
270 Notes
271 -----
272 This method attempts to prune any contracts that contain labels which
273 are not in the declared subset of labels. This pruning is done using a
274 string based matching due to the nature of contracts and may prune more
275 than it should.
276 """
277 return cls.from_uri(filename)
279 @classmethod
280 def from_uri(cls, uri: ResourcePathExpression) -> Pipeline:
281 """Load a pipeline defined in a pipeline yaml file at a location
282 specified by a URI.
284 Parameters
285 ----------
286 uri : convertible to `~lsst.resources.ResourcePath`
287 If a string is supplied this should be a URI path that points to a
288 pipeline defined in yaml format, either as a direct path to the
289 yaml file, or as a directory containing a ``pipeline.yaml`` file
290 the form used by `write_to_uri` with ``expand=True``). This uri may
291 also supply additional labels to be used in subsetting the loaded
292 `Pipeline`. These labels are separated from the path by a ``#``,
293 and may be specified as a comma separated list, or a range denoted
294 as beginning..end. Beginning or end may be empty, in which case the
295 range will be a half open interval. Unlike python iteration bounds,
296 end bounds are *INCLUDED*. Note that range based selection is not
297 well defined for pipelines that are not linear in nature, and
298 correct behavior is not guaranteed, or may vary from run to run.
299 The same specifiers can be used with a
300 `~lsst.resources.ResourcePath` object, by being the sole contents
301 in the fragments attribute.
303 Returns
304 -------
305 pipeline : `Pipeline`
306 The pipeline loaded from specified location with appropriate (if
307 any) subsetting.
309 Notes
310 -----
311 This method attempts to prune any contracts that contain labels which
312 are not in the declared subset of labels. This pruning is done using a
313 string based matching due to the nature of contracts and may prune more
314 than it should.
315 """
316 # Split up the uri and any labels that were supplied
317 uri, label_specifier = cls._parse_file_specifier(uri)
318 pipeline: Pipeline = cls.fromIR(pipelineIR.PipelineIR.from_uri(uri))
320 # If there are labels supplied, only keep those
321 if label_specifier is not None:
322 pipeline = pipeline.subsetFromLabels(label_specifier)
323 return pipeline
325 def subsetFromLabels(self, labelSpecifier: LabelSpecifier) -> Pipeline:
326 """Subset a pipeline to contain only labels specified in labelSpecifier
328 Parameters
329 ----------
330 labelSpecifier : `labelSpecifier`
331 Object containing labels that describes how to subset a pipeline.
333 Returns
334 -------
335 pipeline : `Pipeline`
336 A new pipeline object that is a subset of the old pipeline
338 Raises
339 ------
340 ValueError
341 Raised if there is an issue with specified labels
343 Notes
344 -----
345 This method attempts to prune any contracts that contain labels which
346 are not in the declared subset of labels. This pruning is done using a
347 string based matching due to the nature of contracts and may prune more
348 than it should.
349 """
350 # Labels supplied as a set
351 if labelSpecifier.labels:
352 labelSet = labelSpecifier.labels
353 # Labels supplied as a range, first create a list of all the labels
354 # in the pipeline sorted according to task dependency. Then only
355 # keep labels that lie between the supplied bounds
356 else:
357 # Create a copy of the pipeline to use when assessing the label
358 # ordering. Use a dict for fast searching while preserving order.
359 # Remove contracts so they do not fail in the expansion step. This
360 # is needed because a user may only configure the tasks they intend
361 # to run, which may cause some contracts to fail if they will later
362 # be dropped
363 pipeline = copy.deepcopy(self)
364 pipeline._pipelineIR.contracts = []
365 labels = {taskdef.label: True for taskdef in pipeline.toExpandedPipeline()}
367 # Verify the bounds are in the labels
368 if labelSpecifier.begin is not None:
369 if labelSpecifier.begin not in labels:
370 raise ValueError(
371 f"Beginning of range subset, {labelSpecifier.begin}, not found in pipeline definition"
372 )
373 if labelSpecifier.end is not None:
374 if labelSpecifier.end not in labels:
375 raise ValueError(
376 f"End of range subset, {labelSpecifier.end}, not found in pipeline definition"
377 )
379 labelSet = set()
380 for label in labels:
381 if labelSpecifier.begin is not None:
382 if label != labelSpecifier.begin:
383 continue
384 else:
385 labelSpecifier.begin = None
386 labelSet.add(label)
387 if labelSpecifier.end is not None and label == labelSpecifier.end:
388 break
389 return Pipeline.fromIR(self._pipelineIR.subset_from_labels(labelSet))
391 @staticmethod
392 def _parse_file_specifier(uri: ResourcePathExpression) -> tuple[ResourcePath, LabelSpecifier | None]:
393 """Split appart a uri and any possible label subsets"""
394 if isinstance(uri, str):
395 # This is to support legacy pipelines during transition
396 uri, num_replace = re.subn("[:](?!\\/\\/)", "#", uri)
397 if num_replace:
398 raise ValueError(
399 f"The pipeline file {uri} seems to use the legacy :"
400 " to separate labels, please use # instead."
401 )
402 if uri.count("#") > 1:
403 raise ValueError("Only one set of labels is allowed when specifying a pipeline to load")
404 # Everything else can be converted directly to ResourcePath.
405 uri = ResourcePath(uri)
406 label_subset = uri.fragment or None
408 specifier: LabelSpecifier | None
409 if label_subset is not None:
410 label_subset = urllib.parse.unquote(label_subset)
411 args: dict[str, set[str] | str | None]
412 # labels supplied as a list
413 if "," in label_subset:
414 if ".." in label_subset:
415 raise ValueError(
416 "Can only specify a list of labels or a rangewhen loading a Pipline not both"
417 )
418 args = {"labels": set(label_subset.split(","))}
419 # labels supplied as a range
420 elif ".." in label_subset:
421 # Try to de-structure the labelSubset, this will fail if more
422 # than one range is specified
423 begin, end, *rest = label_subset.split("..")
424 if rest:
425 raise ValueError("Only one range can be specified when loading a pipeline")
426 args = {"begin": begin if begin else None, "end": end if end else None}
427 # Assume anything else is a single label
428 else:
429 args = {"labels": {label_subset}}
431 # MyPy doesn't like how cavalier kwarg construction is with types.
432 specifier = LabelSpecifier(**args) # type: ignore
433 else:
434 specifier = None
436 return uri, specifier
438 @classmethod
439 def fromString(cls, pipeline_string: str) -> Pipeline:
440 """Create a pipeline from string formatted as a pipeline document.
442 Parameters
443 ----------
444 pipeline_string : `str`
445 A string that is formatted according like a pipeline document
447 Returns
448 -------
449 pipeline: `Pipeline`
450 """
451 pipeline = cls.fromIR(pipelineIR.PipelineIR.from_string(pipeline_string))
452 return pipeline
454 @classmethod
455 def fromIR(cls, deserialized_pipeline: pipelineIR.PipelineIR) -> Pipeline:
456 """Create a pipeline from an already created `PipelineIR` object.
458 Parameters
459 ----------
460 deserialized_pipeline: `PipelineIR`
461 An already created pipeline intermediate representation object
463 Returns
464 -------
465 pipeline: `Pipeline`
466 """
467 pipeline = cls.__new__(cls)
468 pipeline._pipelineIR = deserialized_pipeline
469 return pipeline
471 @classmethod
472 def fromPipeline(cls, pipeline: Pipeline) -> Pipeline:
473 """Create a new pipeline by copying an already existing `Pipeline`.
475 Parameters
476 ----------
477 pipeline: `Pipeline`
478 An already created pipeline intermediate representation object
480 Returns
481 -------
482 pipeline: `Pipeline`
483 """
484 return cls.fromIR(copy.deepcopy(pipeline._pipelineIR))
486 def __str__(self) -> str:
487 return str(self._pipelineIR)
489 def mergePipeline(self, pipeline: Pipeline) -> None:
490 """Merge another in-memory `Pipeline` object into this one.
492 This merges another pipeline into this object, as if it were declared
493 in the import block of the yaml definition of this pipeline. This
494 modifies this pipeline in place.
496 Parameters
497 ----------
498 pipeline : `Pipeline`
499 The `Pipeline` object that is to be merged into this object.
500 """
501 self._pipelineIR.merge_pipelines((pipeline._pipelineIR,))
503 def addLabelToSubset(self, subset: str, label: str) -> None:
504 """Add a task label from the specified subset.
506 Parameters
507 ----------
508 subset : `str`
509 The labeled subset to modify
510 label : `str`
511 The task label to add to the specified subset.
513 Raises
514 ------
515 ValueError
516 Raised if the specified subset does not exist within the pipeline.
517 Raised if the specified label does not exist within the pipeline.
518 """
519 if label not in self._pipelineIR.tasks:
520 raise ValueError(f"Label {label} does not appear within the pipeline")
521 if subset not in self._pipelineIR.labeled_subsets:
522 raise ValueError(f"Subset {subset} does not appear within the pipeline")
523 self._pipelineIR.labeled_subsets[subset].subset.add(label)
525 def removeLabelFromSubset(self, subset: str, label: str) -> None:
526 """Remove a task label from the specified subset.
528 Parameters
529 ----------
530 subset : `str`
531 The labeled subset to modify
532 label : `str`
533 The task label to remove from the specified subset.
535 Raises
536 ------
537 ValueError
538 Raised if the specified subset does not exist in the pipeline.
539 Raised if the specified label does not exist within the specified
540 subset.
541 """
542 if subset not in self._pipelineIR.labeled_subsets:
543 raise ValueError(f"Subset {subset} does not appear within the pipeline")
544 if label not in self._pipelineIR.labeled_subsets[subset].subset:
545 raise ValueError(f"Label {label} does not appear within the pipeline")
546 self._pipelineIR.labeled_subsets[subset].subset.remove(label)
548 def findSubsetsWithLabel(self, label: str) -> set[str]:
549 """Find any subsets which may contain the specified label.
551 This function returns the name of subsets which return the specified
552 label. May return an empty set if there are no subsets, or no subsets
553 containing the specified label.
555 Parameters
556 ----------
557 label : `str`
558 The task label to use in membership check
560 Returns
561 -------
562 subsets : `set` of `str`
563 Returns a set (possibly empty) of subsets names which contain the
564 specified label.
566 Raises
567 ------
568 ValueError
569 Raised if the specified label does not exist within this pipeline.
570 """
571 results = set()
572 if label not in self._pipelineIR.tasks:
573 raise ValueError(f"Label {label} does not appear within the pipeline")
574 for subset in self._pipelineIR.labeled_subsets.values():
575 if label in subset.subset:
576 results.add(subset.label)
577 return results
579 def addInstrument(self, instrument: Instrument | str) -> None:
580 """Add an instrument to the pipeline, or replace an instrument that is
581 already defined.
583 Parameters
584 ----------
585 instrument : `~lsst.daf.butler.instrument.Instrument` or `str`
586 Either a derived class object of a `lsst.daf.butler.instrument` or
587 a string corresponding to a fully qualified
588 `lsst.daf.butler.instrument` name.
589 """
590 if isinstance(instrument, str):
591 pass
592 else:
593 # TODO: assume that this is a subclass of Instrument, no type
594 # checking
595 instrument = get_full_type_name(instrument)
596 self._pipelineIR.instrument = instrument
598 def getInstrument(self) -> str | None:
599 """Get the instrument from the pipeline.
601 Returns
602 -------
603 instrument : `str`, or None
604 The fully qualified name of a `lsst.obs.base.Instrument` subclass,
605 name, or None if the pipeline does not have an instrument.
606 """
607 return self._pipelineIR.instrument
609 def get_data_id(self, universe: DimensionUniverse) -> DataCoordinate:
610 """Return a data ID with all dimension constraints embedded in the
611 pipeline.
613 Parameters
614 ----------
615 universe : `lsst.daf.butler.DimensionUniverse`
616 Object that defines all dimensions.
618 Returns
619 -------
620 data_id : `lsst.daf.butler.DataCoordinate`
621 Data ID with all dimension constraints embedded in the
622 pipeline.
623 """
624 instrument_class_name = self._pipelineIR.instrument
625 if instrument_class_name is not None:
626 instrument_class = doImportType(instrument_class_name)
627 if instrument_class is not None:
628 return DataCoordinate.standardize(instrument=instrument_class.getName(), universe=universe)
629 return DataCoordinate.makeEmpty(universe)
631 def addTask(self, task: type[PipelineTask] | str, label: str) -> None:
632 """Add a new task to the pipeline, or replace a task that is already
633 associated with the supplied label.
635 Parameters
636 ----------
637 task: `PipelineTask` or `str`
638 Either a derived class object of a `PipelineTask` or a string
639 corresponding to a fully qualified `PipelineTask` name.
640 label: `str`
641 A label that is used to identify the `PipelineTask` being added
642 """
643 if isinstance(task, str):
644 taskName = task
645 elif issubclass(task, PipelineTask):
646 taskName = get_full_type_name(task)
647 else:
648 raise ValueError(
649 "task must be either a child class of PipelineTask or a string containing"
650 " a fully qualified name to one"
651 )
652 if not label:
653 # in some cases (with command line-generated pipeline) tasks can
654 # be defined without label which is not acceptable, use task
655 # _DefaultName in that case
656 if isinstance(task, str):
657 task_class = doImportType(task)
658 label = task_class._DefaultName
659 self._pipelineIR.tasks[label] = pipelineIR.TaskIR(label, taskName)
661 def removeTask(self, label: str) -> None:
662 """Remove a task from the pipeline.
664 Parameters
665 ----------
666 label : `str`
667 The label used to identify the task that is to be removed
669 Raises
670 ------
671 KeyError
672 If no task with that label exists in the pipeline
674 """
675 self._pipelineIR.tasks.pop(label)
677 def addConfigOverride(self, label: str, key: str, value: object) -> None:
678 """Apply single config override.
680 Parameters
681 ----------
682 label : `str`
683 Label of the task.
684 key: `str`
685 Fully-qualified field name.
686 value : object
687 Value to be given to a field.
688 """
689 self._addConfigImpl(label, pipelineIR.ConfigIR(rest={key: value}))
691 def addConfigFile(self, label: str, filename: str) -> None:
692 """Add overrides from a specified file.
694 Parameters
695 ----------
696 label : `str`
697 The label used to identify the task associated with config to
698 modify
699 filename : `str`
700 Path to the override file.
701 """
702 self._addConfigImpl(label, pipelineIR.ConfigIR(file=[filename]))
704 def addConfigPython(self, label: str, pythonString: str) -> None:
705 """Add Overrides by running a snippet of python code against a config.
707 Parameters
708 ----------
709 label : `str`
710 The label used to identity the task associated with config to
711 modify.
712 pythonString: `str`
713 A string which is valid python code to be executed. This is done
714 with config as the only local accessible value.
715 """
716 self._addConfigImpl(label, pipelineIR.ConfigIR(python=pythonString))
718 def _addConfigImpl(self, label: str, newConfig: pipelineIR.ConfigIR) -> None:
719 if label == "parameters":
720 self._pipelineIR.parameters.mapping.update(newConfig.rest)
721 if newConfig.file:
722 raise ValueError("Setting parameters section with config file is not supported")
723 if newConfig.python:
724 raise ValueError("Setting parameters section using python block in unsupported")
725 return
726 if label not in self._pipelineIR.tasks:
727 raise LookupError(f"There are no tasks labeled '{label}' in the pipeline")
728 self._pipelineIR.tasks[label].add_or_update_config(newConfig)
730 def write_to_uri(self, uri: ResourcePathExpression) -> None:
731 """Write the pipeline to a file or directory.
733 Parameters
734 ----------
735 uri : convertible to `~lsst.resources.ResourcePath`
736 URI to write to; may have any scheme with
737 `~lsst.resources.ResourcePath` write support or no scheme for a
738 local file/directory. Should have a ``.yaml`` extension.
739 """
740 self._pipelineIR.write_to_uri(uri)
742 def toExpandedPipeline(self) -> Generator[TaskDef, None, None]:
743 r"""Return a generator of `TaskDef`\s which can be used to create
744 quantum graphs.
746 Returns
747 -------
748 generator : generator of `TaskDef`
749 The generator returned will be the sorted iterator of tasks which
750 are to be used in constructing a quantum graph.
752 Raises
753 ------
754 NotImplementedError
755 If a dataId is supplied in a config block. This is in place for
756 future use
757 """
758 taskDefs = []
759 for label in self._pipelineIR.tasks:
760 taskDefs.append(self._buildTaskDef(label))
762 # lets evaluate the contracts
763 if self._pipelineIR.contracts is not None:
764 label_to_config = {x.label: x.config for x in taskDefs}
765 for contract in self._pipelineIR.contracts:
766 # execute this in its own line so it can raise a good error
767 # message if there was problems with the eval
768 success = eval(contract.contract, None, label_to_config)
769 if not success:
770 extra_info = f": {contract.msg}" if contract.msg is not None else ""
771 raise pipelineIR.ContractError(
772 f"Contract(s) '{contract.contract}' were not satisfied{extra_info}"
773 )
775 taskDefs = sorted(taskDefs, key=lambda x: x.label)
776 yield from pipeTools.orderPipeline(taskDefs)
778 def _buildTaskDef(self, label: str) -> TaskDef:
779 if (taskIR := self._pipelineIR.tasks.get(label)) is None:
780 raise NameError(f"Label {label} does not appear in this pipeline")
781 taskClass: type[PipelineTask] = doImportType(taskIR.klass)
782 taskName = get_full_type_name(taskClass)
783 config = taskClass.ConfigClass()
784 instrument: PipeBaseInstrument | None = None
785 if (instrumentName := self._pipelineIR.instrument) is not None:
786 instrument_cls: type = doImportType(instrumentName)
787 instrument = instrument_cls()
788 config.applyConfigOverrides(
789 instrument,
790 getattr(taskClass, "_DefaultName", ""),
791 taskIR.config,
792 self._pipelineIR.parameters,
793 label,
794 )
795 return TaskDef(taskName=taskName, config=config, taskClass=taskClass, label=label)
797 def __iter__(self) -> Generator[TaskDef, None, None]:
798 return self.toExpandedPipeline()
800 def __getitem__(self, item: str) -> TaskDef:
801 return self._buildTaskDef(item)
803 def __len__(self) -> int:
804 return len(self._pipelineIR.tasks)
806 def __eq__(self, other: object) -> bool:
807 if not isinstance(other, Pipeline):
808 return False
809 elif self._pipelineIR == other._pipelineIR:
810 # Shortcut: if the IR is the same, the expanded pipeline must be
811 # the same as well. But the converse is not true.
812 return True
813 else:
814 self_expanded = {td.label: (td.taskClass,) for td in self}
815 other_expanded = {td.label: (td.taskClass,) for td in other}
816 if self_expanded != other_expanded:
817 return False
818 # After DM-27847, we should compare configuration here, or better,
819 # delegated to TaskDef.__eq__ after making that compare configurations.
820 raise NotImplementedError(
821 "Pipelines cannot be compared because config instances cannot be compared; see DM-27847."
822 )
825@dataclass(frozen=True)
826class TaskDatasetTypes:
827 """An immutable struct that extracts and classifies the dataset types used
828 by a `PipelineTask`
829 """
831 initInputs: NamedValueSet[DatasetType]
832 """Dataset types that are needed as inputs in order to construct this Task.
834 Task-level `initInputs` may be classified as either
835 `~PipelineDatasetTypes.initInputs` or
836 `~PipelineDatasetTypes.initIntermediates` at the Pipeline level.
837 """
839 initOutputs: NamedValueSet[DatasetType]
840 """Dataset types that may be written after constructing this Task.
842 Task-level `initOutputs` may be classified as either
843 `~PipelineDatasetTypes.initOutputs` or
844 `~PipelineDatasetTypes.initIntermediates` at the Pipeline level.
845 """
847 inputs: NamedValueSet[DatasetType]
848 """Dataset types that are regular inputs to this Task.
850 If an input dataset needed for a Quantum cannot be found in the input
851 collection(s) or produced by another Task in the Pipeline, that Quantum
852 (and all dependent Quanta) will not be produced.
854 Task-level `inputs` may be classified as either
855 `~PipelineDatasetTypes.inputs` or `~PipelineDatasetTypes.intermediates`
856 at the Pipeline level.
857 """
859 queryConstraints: NamedValueSet[DatasetType]
860 """Regular inputs that should not be used as constraints on the initial
861 QuantumGraph generation data ID query, according to their tasks
862 (`NamedValueSet`).
863 """
865 prerequisites: NamedValueSet[DatasetType]
866 """Dataset types that are prerequisite inputs to this Task.
868 Prerequisite inputs must exist in the input collection(s) before the
869 pipeline is run, but do not constrain the graph - if a prerequisite is
870 missing for a Quantum, `PrerequisiteMissingError` is raised.
872 Prerequisite inputs are not resolved until the second stage of
873 QuantumGraph generation.
874 """
876 outputs: NamedValueSet[DatasetType]
877 """Dataset types that are produced by this Task.
879 Task-level `outputs` may be classified as either
880 `~PipelineDatasetTypes.outputs` or `~PipelineDatasetTypes.intermediates`
881 at the Pipeline level.
882 """
884 @classmethod
885 def fromTaskDef(
886 cls,
887 taskDef: TaskDef,
888 *,
889 registry: Registry,
890 include_configs: bool = True,
891 storage_class_mapping: Mapping[str, str] | None = None,
892 ) -> TaskDatasetTypes:
893 """Extract and classify the dataset types from a single `PipelineTask`.
895 Parameters
896 ----------
897 taskDef: `TaskDef`
898 An instance of a `TaskDef` class for a particular `PipelineTask`.
899 registry: `Registry`
900 Registry used to construct normalized
901 `~lsst.daf.butler.DatasetType` objects and retrieve those that are
902 incomplete.
903 include_configs : `bool`, optional
904 If `True` (default) include config dataset types as
905 ``initOutputs``.
906 storage_class_mapping : `~collections.abc.Mapping` of `str` to \
907 `StorageClass`, optional
908 If a taskdef contains a component dataset type that is unknown
909 to the registry, its parent `~lsst.daf.butler.StorageClass` will
910 be looked up in this mapping if it is supplied. If the mapping does
911 not contain the composite dataset type, or the mapping is not
912 supplied an exception will be raised.
914 Returns
915 -------
916 types: `TaskDatasetTypes`
917 The dataset types used by this task.
919 Raises
920 ------
921 ValueError
922 Raised if dataset type connection definition differs from
923 registry definition.
924 LookupError
925 Raised if component parent StorageClass could not be determined
926 and storage_class_mapping does not contain the composite type, or
927 is set to None.
928 """
930 def makeDatasetTypesSet(
931 connectionType: str,
932 is_input: bool,
933 freeze: bool = True,
934 ) -> NamedValueSet[DatasetType]:
935 """Construct a set of true `~lsst.daf.butler.DatasetType` objects.
937 Parameters
938 ----------
939 connectionType : `str`
940 Name of the connection type to produce a set for, corresponds
941 to an attribute of type `list` on the connection class instance
942 is_input : `bool`
943 These are input dataset types, else they are output dataset
944 types.
945 freeze : `bool`, optional
946 If `True`, call `NamedValueSet.freeze` on the object returned.
948 Returns
949 -------
950 datasetTypes : `NamedValueSet`
951 A set of all datasetTypes which correspond to the input
952 connection type specified in the connection class of this
953 `PipelineTask`
955 Raises
956 ------
957 ValueError
958 Raised if dataset type connection definition differs from
959 registry definition.
960 LookupError
961 Raised if component parent StorageClass could not be determined
962 and storage_class_mapping does not contain the composite type,
963 or is set to None.
965 Notes
966 -----
967 This function is a closure over the variables ``registry`` and
968 ``taskDef``, and ``storage_class_mapping``.
969 """
970 datasetTypes = NamedValueSet[DatasetType]()
971 for c in iterConnections(taskDef.connections, connectionType):
972 dimensions = set(getattr(c, "dimensions", set()))
973 if "skypix" in dimensions:
974 try:
975 datasetType = registry.getDatasetType(c.name)
976 except LookupError as err:
977 raise LookupError(
978 f"DatasetType '{c.name}' referenced by "
979 f"{type(taskDef.connections).__name__} uses 'skypix' as a dimension "
980 "placeholder, but does not already exist in the registry. "
981 "Note that reference catalog names are now used as the dataset "
982 "type name instead of 'ref_cat'."
983 ) from err
984 rest1 = set(registry.dimensions.extract(dimensions - set(["skypix"])).names)
985 rest2 = set(
986 dim.name for dim in datasetType.dimensions if not isinstance(dim, SkyPixDimension)
987 )
988 if rest1 != rest2:
989 raise ValueError(
990 f"Non-skypix dimensions for dataset type {c.name} declared in "
991 f"connections ({rest1}) are inconsistent with those in "
992 f"registry's version of this dataset ({rest2})."
993 )
994 else:
995 # Component dataset types are not explicitly in the
996 # registry. This complicates consistency checks with
997 # registry and requires we work out the composite storage
998 # class.
999 registryDatasetType = None
1000 try:
1001 registryDatasetType = registry.getDatasetType(c.name)
1002 except KeyError:
1003 compositeName, componentName = DatasetType.splitDatasetTypeName(c.name)
1004 if componentName:
1005 if storage_class_mapping is None or compositeName not in storage_class_mapping:
1006 raise LookupError(
1007 "Component parent class cannot be determined, and "
1008 "composite name was not in storage class mapping, or no "
1009 "storage_class_mapping was supplied"
1010 )
1011 else:
1012 parentStorageClass = storage_class_mapping[compositeName]
1013 else:
1014 parentStorageClass = None
1015 datasetType = c.makeDatasetType(
1016 registry.dimensions, parentStorageClass=parentStorageClass
1017 )
1018 registryDatasetType = datasetType
1019 else:
1020 datasetType = c.makeDatasetType(
1021 registry.dimensions, parentStorageClass=registryDatasetType.parentStorageClass
1022 )
1024 if registryDatasetType and datasetType != registryDatasetType:
1025 # The dataset types differ but first check to see if
1026 # they are compatible before raising.
1027 if is_input:
1028 # This DatasetType must be compatible on get.
1029 is_compatible = datasetType.is_compatible_with(registryDatasetType)
1030 else:
1031 # Has to be able to be converted to expect type
1032 # on put.
1033 is_compatible = registryDatasetType.is_compatible_with(datasetType)
1034 if is_compatible:
1035 # For inputs we want the pipeline to use the
1036 # pipeline definition, for outputs it should use
1037 # the registry definition.
1038 if not is_input:
1039 datasetType = registryDatasetType
1040 _LOG.debug(
1041 "Dataset types differ (task %s != registry %s) but are compatible"
1042 " for %s in %s.",
1043 datasetType,
1044 registryDatasetType,
1045 "input" if is_input else "output",
1046 taskDef.label,
1047 )
1048 else:
1049 try:
1050 # Explicitly check for storage class just to
1051 # make more specific message.
1052 _ = datasetType.storageClass
1053 except KeyError:
1054 raise ValueError(
1055 "Storage class does not exist for supplied dataset type "
1056 f"{datasetType} for {taskDef.label}."
1057 ) from None
1058 raise ValueError(
1059 f"Supplied dataset type ({datasetType}) inconsistent with "
1060 f"registry definition ({registryDatasetType}) "
1061 f"for {taskDef.label}."
1062 )
1063 datasetTypes.add(datasetType)
1064 if freeze:
1065 datasetTypes.freeze()
1066 return datasetTypes
1068 # optionally add initOutput dataset for config
1069 initOutputs = makeDatasetTypesSet("initOutputs", is_input=False, freeze=False)
1070 if include_configs:
1071 initOutputs.add(
1072 DatasetType(
1073 taskDef.configDatasetName,
1074 registry.dimensions.empty,
1075 storageClass="Config",
1076 )
1077 )
1078 initOutputs.freeze()
1080 # optionally add output dataset for metadata
1081 outputs = makeDatasetTypesSet("outputs", is_input=False, freeze=False)
1083 # Metadata is supposed to be of the TaskMetadata type, its dimensions
1084 # correspond to a task quantum.
1085 dimensions = registry.dimensions.extract(taskDef.connections.dimensions)
1087 # Allow the storage class definition to be read from the existing
1088 # dataset type definition if present.
1089 try:
1090 current = registry.getDatasetType(taskDef.metadataDatasetName)
1091 except KeyError:
1092 # No previous definition so use the default.
1093 storageClass = "TaskMetadata" if _TASK_METADATA_TYPE is TaskMetadata else "PropertySet"
1094 else:
1095 storageClass = current.storageClass.name
1096 outputs.update({DatasetType(taskDef.metadataDatasetName, dimensions, storageClass)})
1098 if taskDef.logOutputDatasetName is not None:
1099 # Log output dimensions correspond to a task quantum.
1100 dimensions = registry.dimensions.extract(taskDef.connections.dimensions)
1101 outputs.update({DatasetType(taskDef.logOutputDatasetName, dimensions, "ButlerLogRecords")})
1103 outputs.freeze()
1105 inputs = makeDatasetTypesSet("inputs", is_input=True)
1106 queryConstraints = NamedValueSet(
1107 inputs[c.name]
1108 for c in cast(Iterable[Input], iterConnections(taskDef.connections, "inputs"))
1109 if not c.deferGraphConstraint
1110 )
1112 return cls(
1113 initInputs=makeDatasetTypesSet("initInputs", is_input=True),
1114 initOutputs=initOutputs,
1115 inputs=inputs,
1116 queryConstraints=queryConstraints,
1117 prerequisites=makeDatasetTypesSet("prerequisiteInputs", is_input=True),
1118 outputs=outputs,
1119 )
1122@dataclass(frozen=True)
1123class PipelineDatasetTypes:
1124 """An immutable struct that classifies the dataset types used in a
1125 `Pipeline`.
1126 """
1128 packagesDatasetName: ClassVar[str] = "packages"
1129 """Name of a dataset type used to save package versions.
1130 """
1132 initInputs: NamedValueSet[DatasetType]
1133 """Dataset types that are needed as inputs in order to construct the Tasks
1134 in this Pipeline.
1136 This does not include dataset types that are produced when constructing
1137 other Tasks in the Pipeline (these are classified as `initIntermediates`).
1138 """
1140 initOutputs: NamedValueSet[DatasetType]
1141 """Dataset types that may be written after constructing the Tasks in this
1142 Pipeline.
1144 This does not include dataset types that are also used as inputs when
1145 constructing other Tasks in the Pipeline (these are classified as
1146 `initIntermediates`).
1147 """
1149 initIntermediates: NamedValueSet[DatasetType]
1150 """Dataset types that are both used when constructing one or more Tasks
1151 in the Pipeline and produced as a side-effect of constructing another
1152 Task in the Pipeline.
1153 """
1155 inputs: NamedValueSet[DatasetType]
1156 """Dataset types that are regular inputs for the full pipeline.
1158 If an input dataset needed for a Quantum cannot be found in the input
1159 collection(s), that Quantum (and all dependent Quanta) will not be
1160 produced.
1161 """
1163 queryConstraints: NamedValueSet[DatasetType]
1164 """Regular inputs that should be used as constraints on the initial
1165 QuantumGraph generation data ID query, according to their tasks
1166 (`NamedValueSet`).
1167 """
1169 prerequisites: NamedValueSet[DatasetType]
1170 """Dataset types that are prerequisite inputs for the full Pipeline.
1172 Prerequisite inputs must exist in the input collection(s) before the
1173 pipeline is run, but do not constrain the graph - if a prerequisite is
1174 missing for a Quantum, `PrerequisiteMissingError` is raised.
1176 Prerequisite inputs are not resolved until the second stage of
1177 QuantumGraph generation.
1178 """
1180 intermediates: NamedValueSet[DatasetType]
1181 """Dataset types that are output by one Task in the Pipeline and consumed
1182 as inputs by one or more other Tasks in the Pipeline.
1183 """
1185 outputs: NamedValueSet[DatasetType]
1186 """Dataset types that are output by a Task in the Pipeline and not consumed
1187 by any other Task in the Pipeline.
1188 """
1190 byTask: Mapping[str, TaskDatasetTypes]
1191 """Per-Task dataset types, keyed by label in the `Pipeline`.
1193 This is guaranteed to be zip-iterable with the `Pipeline` itself (assuming
1194 neither has been modified since the dataset types were extracted, of
1195 course).
1196 """
1198 @classmethod
1199 def fromPipeline(
1200 cls,
1201 pipeline: Pipeline | Iterable[TaskDef],
1202 *,
1203 registry: Registry,
1204 include_configs: bool = True,
1205 include_packages: bool = True,
1206 ) -> PipelineDatasetTypes:
1207 """Extract and classify the dataset types from all tasks in a
1208 `Pipeline`.
1210 Parameters
1211 ----------
1212 pipeline: `Pipeline` or `~collections.abc.Iterable` [ `TaskDef` ]
1213 A collection of tasks that can be run together.
1214 registry: `Registry`
1215 Registry used to construct normalized
1216 `~lsst.daf.butler.DatasetType` objects and retrieve those that are
1217 incomplete.
1218 include_configs : `bool`, optional
1219 If `True` (default) include config dataset types as
1220 ``initOutputs``.
1221 include_packages : `bool`, optional
1222 If `True` (default) include the dataset type for software package
1223 versions in ``initOutputs``.
1225 Returns
1226 -------
1227 types: `PipelineDatasetTypes`
1228 The dataset types used by this `Pipeline`.
1230 Raises
1231 ------
1232 ValueError
1233 Raised if Tasks are inconsistent about which datasets are marked
1234 prerequisite. This indicates that the Tasks cannot be run as part
1235 of the same `Pipeline`.
1236 """
1237 allInputs = NamedValueSet[DatasetType]()
1238 allOutputs = NamedValueSet[DatasetType]()
1239 allInitInputs = NamedValueSet[DatasetType]()
1240 allInitOutputs = NamedValueSet[DatasetType]()
1241 prerequisites = NamedValueSet[DatasetType]()
1242 queryConstraints = NamedValueSet[DatasetType]()
1243 byTask = dict()
1244 if include_packages:
1245 allInitOutputs.add(
1246 DatasetType(
1247 cls.packagesDatasetName,
1248 registry.dimensions.empty,
1249 storageClass="Packages",
1250 )
1251 )
1252 # create a list of TaskDefs in case the input is a generator
1253 pipeline = list(pipeline)
1255 # collect all the output dataset types
1256 typeStorageclassMap: dict[str, str] = {}
1257 for taskDef in pipeline:
1258 for outConnection in iterConnections(taskDef.connections, "outputs"):
1259 typeStorageclassMap[outConnection.name] = outConnection.storageClass
1261 for taskDef in pipeline:
1262 thisTask = TaskDatasetTypes.fromTaskDef(
1263 taskDef,
1264 registry=registry,
1265 include_configs=include_configs,
1266 storage_class_mapping=typeStorageclassMap,
1267 )
1268 allInitInputs.update(thisTask.initInputs)
1269 allInitOutputs.update(thisTask.initOutputs)
1270 allInputs.update(thisTask.inputs)
1271 # Inputs are query constraints if any task considers them a query
1272 # constraint.
1273 queryConstraints.update(thisTask.queryConstraints)
1274 prerequisites.update(thisTask.prerequisites)
1275 allOutputs.update(thisTask.outputs)
1276 byTask[taskDef.label] = thisTask
1277 if not prerequisites.isdisjoint(allInputs):
1278 raise ValueError(
1279 "{} marked as both prerequisites and regular inputs".format(
1280 {dt.name for dt in allInputs & prerequisites}
1281 )
1282 )
1283 if not prerequisites.isdisjoint(allOutputs):
1284 raise ValueError(
1285 "{} marked as both prerequisites and outputs".format(
1286 {dt.name for dt in allOutputs & prerequisites}
1287 )
1288 )
1289 # Make sure that components which are marked as inputs get treated as
1290 # intermediates if there is an output which produces the composite
1291 # containing the component
1292 intermediateComponents = NamedValueSet[DatasetType]()
1293 intermediateComposites = NamedValueSet[DatasetType]()
1294 outputNameMapping = {dsType.name: dsType for dsType in allOutputs}
1295 for dsType in allInputs:
1296 # get the name of a possible component
1297 name, component = dsType.nameAndComponent()
1298 # if there is a component name, that means this is a component
1299 # DatasetType, if there is an output which produces the parent of
1300 # this component, treat this input as an intermediate
1301 if component is not None:
1302 # This needs to be in this if block, because someone might have
1303 # a composite that is a pure input from existing data
1304 if name in outputNameMapping:
1305 intermediateComponents.add(dsType)
1306 intermediateComposites.add(outputNameMapping[name])
1308 def checkConsistency(a: NamedValueSet, b: NamedValueSet) -> None:
1309 common = a.names & b.names
1310 for name in common:
1311 # Any compatibility is allowed. This function does not know
1312 # if a dataset type is to be used for input or output.
1313 if not (a[name].is_compatible_with(b[name]) or b[name].is_compatible_with(a[name])):
1314 raise ValueError(f"Conflicting definitions for dataset type: {a[name]} != {b[name]}.")
1316 checkConsistency(allInitInputs, allInitOutputs)
1317 checkConsistency(allInputs, allOutputs)
1318 checkConsistency(allInputs, intermediateComposites)
1319 checkConsistency(allOutputs, intermediateComposites)
1321 def frozen(s: Set[DatasetType]) -> NamedValueSet[DatasetType]:
1322 assert isinstance(s, NamedValueSet)
1323 s.freeze()
1324 return s
1326 inputs = frozen(allInputs - allOutputs - intermediateComponents)
1328 return cls(
1329 initInputs=frozen(allInitInputs - allInitOutputs),
1330 initIntermediates=frozen(allInitInputs & allInitOutputs),
1331 initOutputs=frozen(allInitOutputs - allInitInputs),
1332 inputs=inputs,
1333 queryConstraints=frozen(queryConstraints & inputs),
1334 # If there are storage class differences in inputs and outputs
1335 # the intermediates have to choose priority. Here choose that
1336 # inputs to tasks much match the requested storage class by
1337 # applying the inputs over the top of the outputs.
1338 intermediates=frozen(allOutputs & allInputs | intermediateComponents),
1339 outputs=frozen(allOutputs - allInputs - intermediateComposites),
1340 prerequisites=frozen(prerequisites),
1341 byTask=MappingProxyType(byTask), # MappingProxyType -> frozen view of dict for immutability
1342 )
1344 @classmethod
1345 def initOutputNames(
1346 cls,
1347 pipeline: Pipeline | Iterable[TaskDef],
1348 *,
1349 include_configs: bool = True,
1350 include_packages: bool = True,
1351 ) -> Iterator[str]:
1352 """Return the names of dataset types ot task initOutputs, Configs,
1353 and package versions for a pipeline.
1355 Parameters
1356 ----------
1357 pipeline: `Pipeline` or `~collections.abc.Iterable` [ `TaskDef` ]
1358 A `Pipeline` instance or collection of `TaskDef` instances.
1359 include_configs : `bool`, optional
1360 If `True` (default) include config dataset types.
1361 include_packages : `bool`, optional
1362 If `True` (default) include the dataset type for package versions.
1364 Yields
1365 ------
1366 datasetTypeName : `str`
1367 Name of the dataset type.
1368 """
1369 if include_packages:
1370 # Package versions dataset type
1371 yield cls.packagesDatasetName
1373 if isinstance(pipeline, Pipeline):
1374 pipeline = pipeline.toExpandedPipeline()
1376 for taskDef in pipeline:
1377 # all task InitOutputs
1378 for name in taskDef.connections.initOutputs:
1379 attribute = getattr(taskDef.connections, name)
1380 yield attribute.name
1382 # config dataset name
1383 if include_configs:
1384 yield taskDef.configDatasetName