Coverage for python/lsst/pipe/base/tests/simpleQGraph.py: 29%
155 statements
« prev ^ index » next coverage.py v6.5.0, created at 2023-04-12 09:13 +0000
« prev ^ index » next coverage.py v6.5.0, created at 2023-04-12 09:13 +0000
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/>.
22"""Bunch of common classes and methods for use in unit tests.
23"""
24from __future__ import annotations
26__all__ = ["AddTaskConfig", "AddTask", "AddTaskFactoryMock"]
28import itertools
29import logging
30from collections.abc import Iterable, Mapping
31from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Type, Union, cast
33import lsst.daf.butler.tests as butlerTests
34import lsst.pex.config as pexConfig
35import numpy
36from lsst.daf.butler import Butler, Config, DataId, DatasetRef, DatasetType, Formatter, LimitedButler
37from lsst.daf.butler.core.logging import ButlerLogRecords
38from lsst.resources import ResourcePath
39from lsst.utils import doImportType
40from lsst.utils.introspection import get_full_type_name
42from .. import connectionTypes as cT
43from .._instrument import Instrument
44from ..config import PipelineTaskConfig
45from ..connections import PipelineTaskConnections
46from ..graph import QuantumGraph
47from ..graphBuilder import DatasetQueryConstraintVariant as DSQVariant
48from ..graphBuilder import GraphBuilder
49from ..pipeline import Pipeline, TaskDatasetTypes, TaskDef
50from ..pipelineTask import PipelineTask
51from ..struct import Struct
52from ..task import _TASK_FULL_METADATA_TYPE
53from ..taskFactory import TaskFactory
55if TYPE_CHECKING: 55 ↛ 56line 55 didn't jump to line 56, because the condition on line 55 was never true
56 from lsst.daf.butler import Registry
58_LOG = logging.getLogger(__name__)
61class SimpleInstrument(Instrument):
62 def __init__(self, *args: Any, **kwargs: Any):
63 pass
65 @staticmethod
66 def getName() -> str:
67 return "INSTRU"
69 def getRawFormatter(self, dataId: DataId) -> Type[Formatter]:
70 return Formatter
72 def register(self, registry: Registry, *, update: bool = False) -> None:
73 pass
76class AddTaskConnections(
77 PipelineTaskConnections,
78 dimensions=("instrument", "detector"),
79 defaultTemplates={"in_tmpl": "_in", "out_tmpl": "_out"},
80):
81 """Connections for AddTask, has one input and two outputs,
82 plus one init output.
83 """
85 input = cT.Input(
86 name="add_dataset{in_tmpl}",
87 dimensions=["instrument", "detector"],
88 storageClass="NumpyArray",
89 doc="Input dataset type for this task",
90 )
91 output = cT.Output(
92 name="add_dataset{out_tmpl}",
93 dimensions=["instrument", "detector"],
94 storageClass="NumpyArray",
95 doc="Output dataset type for this task",
96 )
97 output2 = cT.Output(
98 name="add2_dataset{out_tmpl}",
99 dimensions=["instrument", "detector"],
100 storageClass="NumpyArray",
101 doc="Output dataset type for this task",
102 )
103 initout = cT.InitOutput(
104 name="add_init_output{out_tmpl}",
105 storageClass="NumpyArray",
106 doc="Init Output dataset type for this task",
107 )
110class AddTaskConfig(PipelineTaskConfig, pipelineConnections=AddTaskConnections):
111 """Config for AddTask."""
113 addend = pexConfig.Field[int](doc="amount to add", default=3)
116class AddTask(PipelineTask):
117 """Trivial PipelineTask for testing, has some extras useful for specific
118 unit tests.
119 """
121 ConfigClass = AddTaskConfig
122 _DefaultName = "add_task"
124 initout = numpy.array([999])
125 """InitOutputs for this task"""
127 taskFactory: Optional[AddTaskFactoryMock] = None
128 """Factory that makes instances"""
130 def run(self, input: int) -> Struct: # type: ignore
131 if self.taskFactory:
132 # do some bookkeeping
133 if self.taskFactory.stopAt == self.taskFactory.countExec:
134 raise RuntimeError("pretend something bad happened")
135 self.taskFactory.countExec += 1
137 self.config = cast(AddTaskConfig, self.config)
138 self.metadata.add("add", self.config.addend)
139 output = input + self.config.addend
140 output2 = output + self.config.addend
141 _LOG.info("input = %s, output = %s, output2 = %s", input, output, output2)
142 return Struct(output=output, output2=output2)
145class AddTaskFactoryMock(TaskFactory):
146 """Special task factory that instantiates AddTask.
148 It also defines some bookkeeping variables used by AddTask to report
149 progress to unit tests.
150 """
152 def __init__(self, stopAt: int = -1):
153 self.countExec = 0 # incremented by AddTask
154 self.stopAt = stopAt # AddTask raises exception at this call to run()
156 def makeTask(
157 self, taskDef: TaskDef, butler: LimitedButler, initInputRefs: Iterable[DatasetRef] | None
158 ) -> PipelineTask:
159 taskClass = taskDef.taskClass
160 assert taskClass is not None
161 task = taskClass(config=taskDef.config, initInputs=None, name=taskDef.label)
162 task.taskFactory = self # type: ignore
163 return task
166def registerDatasetTypes(registry: Registry, pipeline: Union[Pipeline, Iterable[TaskDef]]) -> None:
167 """Register all dataset types used by tasks in a registry.
169 Copied and modified from `PreExecInit.initializeDatasetTypes`.
171 Parameters
172 ----------
173 registry : `~lsst.daf.butler.Registry`
174 Registry instance.
175 pipeline : `typing.Iterable` of `TaskDef`
176 Iterable of TaskDef instances, likely the output of the method
177 toExpandedPipeline on a `~lsst.pipe.base.Pipeline` object
178 """
179 for taskDef in pipeline:
180 configDatasetType = DatasetType(
181 taskDef.configDatasetName, {}, storageClass="Config", universe=registry.dimensions
182 )
183 storageClass = "Packages"
184 packagesDatasetType = DatasetType(
185 "packages", {}, storageClass=storageClass, universe=registry.dimensions
186 )
187 datasetTypes = TaskDatasetTypes.fromTaskDef(taskDef, registry=registry)
188 for datasetType in itertools.chain(
189 datasetTypes.initInputs,
190 datasetTypes.initOutputs,
191 datasetTypes.inputs,
192 datasetTypes.outputs,
193 datasetTypes.prerequisites,
194 [configDatasetType, packagesDatasetType],
195 ):
196 _LOG.info("Registering %s with registry", datasetType)
197 # this is a no-op if it already exists and is consistent,
198 # and it raises if it is inconsistent. But components must be
199 # skipped
200 if not datasetType.isComponent():
201 registry.registerDatasetType(datasetType)
204def makeSimplePipeline(nQuanta: int, instrument: Optional[str] = None) -> Pipeline:
205 """Make a simple Pipeline for tests.
207 This is called by ``makeSimpleQGraph`` if no pipeline is passed to that
208 function. It can also be used to customize the pipeline used by
209 ``makeSimpleQGraph`` function by calling this first and passing the result
210 to it.
212 Parameters
213 ----------
214 nQuanta : `int`
215 The number of quanta to add to the pipeline.
216 instrument : `str` or `None`, optional
217 The importable name of an instrument to be added to the pipeline or
218 if no instrument should be added then an empty string or `None`, by
219 default None
221 Returns
222 -------
223 pipeline : `~lsst.pipe.base.Pipeline`
224 The created pipeline object.
225 """
226 pipeline = Pipeline("test pipeline")
227 # make a bunch of tasks that execute in well defined order (via data
228 # dependencies)
229 for lvl in range(nQuanta):
230 pipeline.addTask(AddTask, f"task{lvl}")
231 pipeline.addConfigOverride(f"task{lvl}", "connections.in_tmpl", lvl)
232 pipeline.addConfigOverride(f"task{lvl}", "connections.out_tmpl", lvl + 1)
233 if instrument:
234 pipeline.addInstrument(instrument)
235 return pipeline
238def makeSimpleButler(root: str, run: str = "test", inMemory: bool = True) -> Butler:
239 """Create new data butler instance.
241 Parameters
242 ----------
243 root : `str`
244 Path or URI to the root location of the new repository.
245 run : `str`, optional
246 Run collection name.
247 inMemory : `bool`, optional
248 If true make in-memory repository.
250 Returns
251 -------
252 butler : `~lsst.daf.butler.Butler`
253 Data butler instance.
254 """
255 root_path = ResourcePath(root, forceDirectory=True)
256 if not root_path.isLocal:
257 raise ValueError(f"Only works with local root not {root_path}")
258 config = Config()
259 if not inMemory:
260 config["registry", "db"] = f"sqlite:///{root_path.ospath}/gen3.sqlite"
261 config["datastore", "cls"] = "lsst.daf.butler.datastores.fileDatastore.FileDatastore"
262 repo = butlerTests.makeTestRepo(str(root_path), {}, config=config)
263 butler = Butler(butler=repo, run=run)
264 return butler
267def populateButler(
268 pipeline: Pipeline, butler: Butler, datasetTypes: Dict[Optional[str], List[str]] | None = None
269) -> None:
270 """Populate data butler with data needed for test.
272 Initializes data butler with a bunch of items:
273 - registers dataset types which are defined by pipeline
274 - create dimensions data for (instrument, detector)
275 - adds datasets based on ``datasetTypes`` dictionary, if dictionary is
276 missing then a single dataset with type "add_dataset0" is added
278 All datasets added to butler have ``dataId={instrument=instrument,
279 detector=0}`` where ``instrument`` is extracted from pipeline, "INSTR" is
280 used if pipeline is missing instrument definition. Type of the dataset is
281 guessed from dataset type name (assumes that pipeline is made of `AddTask`
282 tasks).
284 Parameters
285 ----------
286 pipeline : `~lsst.pipe.base.Pipeline`
287 Pipeline instance.
288 butler : `~lsst.daf.butler.Butler`
289 Data butler instance.
290 datasetTypes : `dict` [ `str`, `list` ], optional
291 Dictionary whose keys are collection names and values are lists of
292 dataset type names. By default a single dataset of type "add_dataset0"
293 is added to a ``butler.run`` collection.
294 """
296 # Add dataset types to registry
297 taskDefs = list(pipeline.toExpandedPipeline())
298 registerDatasetTypes(butler.registry, taskDefs)
300 instrument = pipeline.getInstrument()
301 if instrument is not None:
302 instrument_class = doImportType(instrument)
303 instrumentName = instrument_class.getName()
304 instrumentClass = get_full_type_name(instrument_class)
305 else:
306 instrumentName = "INSTR"
307 instrumentClass = None
309 # Add all needed dimensions to registry
310 butler.registry.insertDimensionData("instrument", dict(name=instrumentName, class_name=instrumentClass))
311 butler.registry.insertDimensionData("detector", dict(instrument=instrumentName, id=0, full_name="det0"))
313 taskDefMap = dict((taskDef.label, taskDef) for taskDef in taskDefs)
314 # Add inputs to butler
315 if not datasetTypes:
316 datasetTypes = {None: ["add_dataset0"]}
317 for run, dsTypes in datasetTypes.items():
318 if run is not None:
319 butler.registry.registerRun(run)
320 for dsType in dsTypes:
321 if dsType == "packages":
322 # Version is intentionally inconsistent.
323 # Dict is convertible to Packages if Packages is installed.
324 data: Any = {"python": "9.9.99"}
325 butler.put(data, dsType, run=run)
326 else:
327 if dsType.endswith("_config"):
328 # find a config from matching task name or make a new one
329 taskLabel, _, _ = dsType.rpartition("_")
330 taskDef = taskDefMap.get(taskLabel)
331 if taskDef is not None:
332 data = taskDef.config
333 else:
334 data = AddTaskConfig()
335 elif dsType.endswith("_metadata"):
336 data = _TASK_FULL_METADATA_TYPE()
337 elif dsType.endswith("_log"):
338 data = ButlerLogRecords.from_records([])
339 else:
340 data = numpy.array([0.0, 1.0, 2.0, 5.0])
341 butler.put(data, dsType, run=run, instrument=instrumentName, detector=0)
344def makeSimpleQGraph(
345 nQuanta: int = 5,
346 pipeline: Optional[Pipeline] = None,
347 butler: Optional[Butler] = None,
348 root: Optional[str] = None,
349 callPopulateButler: bool = True,
350 run: str = "test",
351 instrument: Optional[str] = None,
352 skipExistingIn: Any = None,
353 inMemory: bool = True,
354 userQuery: str = "",
355 datasetTypes: Optional[Dict[Optional[str], List[str]]] = None,
356 datasetQueryConstraint: DSQVariant = DSQVariant.ALL,
357 makeDatastoreRecords: bool = False,
358 resolveRefs: bool = False,
359 bind: Optional[Mapping[str, Any]] = None,
360) -> Tuple[Butler, QuantumGraph]:
361 """Make simple QuantumGraph for tests.
363 Makes simple one-task pipeline with AddTask, sets up in-memory registry
364 and butler, fills them with minimal data, and generates QuantumGraph with
365 all of that.
367 Parameters
368 ----------
369 nQuanta : `int`
370 Number of quanta in a graph, only used if ``pipeline`` is None.
371 pipeline : `~lsst.pipe.base.Pipeline`
372 If `None` then a pipeline is made with `AddTask` and default
373 `AddTaskConfig`.
374 butler : `~lsst.daf.butler.Butler`, optional
375 Data butler instance, if None then new data butler is created by
376 calling `makeSimpleButler`.
377 callPopulateButler : `bool`, optional
378 If True insert datasets into the butler prior to building a graph.
379 If False butler argument must not be None, and must be pre-populated.
380 Defaults to True.
381 root : `str`
382 Path or URI to the root location of the new repository. Only used if
383 ``butler`` is None.
384 run : `str`, optional
385 Name of the RUN collection to add to butler, only used if ``butler``
386 is None.
387 instrument : `str` or `None`, optional
388 The importable name of an instrument to be added to the pipeline or
389 if no instrument should be added then an empty string or `None`, by
390 default `None`. Only used if ``pipeline`` is `None`.
391 skipExistingIn
392 Expressions representing the collections to search for existing
393 output datasets that should be skipped. See
394 :ref:`daf_butler_ordered_collection_searches`.
395 inMemory : `bool`, optional
396 If true make in-memory repository, only used if ``butler`` is `None`.
397 userQuery : `str`, optional
398 The user query to pass to ``makeGraph``, by default an empty string.
399 datasetTypes : `dict` [ `str`, `list` ], optional
400 Dictionary whose keys are collection names and values are lists of
401 dataset type names. By default a single dataset of type "add_dataset0"
402 is added to a ``butler.run`` collection.
403 datasetQueryQConstraint : `DatasetQueryConstraintVariant`
404 The query constraint variant that should be used to constrain the
405 query based on dataset existence, defaults to
406 `DatasetQueryConstraintVariant.ALL`.
407 makeDatastoreRecords : `bool`, optional
408 If `True` then add datstore records to generated quanta.
409 resolveRefs : `bool`, optional
410 If `True` then resolve all input references and generate random dataset
411 IDs for all output and intermediate datasets.
412 bind : `Mapping`, optional
413 Mapping containing literal values that should be injected into the
414 ``userQuery`` expression, keyed by the identifiers they replace.
416 Returns
417 -------
418 butler : `~lsst.daf.butler.Butler`
419 Butler instance
420 qgraph : `~lsst.pipe.base.QuantumGraph`
421 Quantum graph instance
422 """
424 if pipeline is None:
425 pipeline = makeSimplePipeline(nQuanta=nQuanta, instrument=instrument)
427 if butler is None:
428 if root is None:
429 raise ValueError("Must provide `root` when `butler` is None")
430 if callPopulateButler is False:
431 raise ValueError("populateButler can only be False when butler is supplied as an argument")
432 butler = makeSimpleButler(root, run=run, inMemory=inMemory)
434 if callPopulateButler:
435 populateButler(pipeline, butler, datasetTypes=datasetTypes)
437 # Make the graph
438 _LOG.debug("Instantiating GraphBuilder, skipExistingIn=%s", skipExistingIn)
439 builder = GraphBuilder(
440 registry=butler.registry,
441 skipExistingIn=skipExistingIn,
442 datastore=butler.datastore if makeDatastoreRecords else None,
443 )
444 _LOG.debug(
445 "Calling GraphBuilder.makeGraph, collections=%r, run=%r, userQuery=%r bind=%s",
446 butler.collections,
447 run or butler.run,
448 userQuery,
449 bind,
450 )
451 qgraph = builder.makeGraph(
452 pipeline,
453 collections=butler.collections,
454 run=run or butler.run,
455 userQuery=userQuery,
456 datasetQueryConstraint=datasetQueryConstraint,
457 resolveRefs=resolveRefs,
458 bind=bind,
459 )
461 return butler, qgraph