Coverage for python / lsst / pipe / base / _instrument.py: 39%
132 statements
« prev ^ index » next coverage.py v7.13.1, created at 2026-01-31 01:07 -0800
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# (https://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 software is dual licensed under the GNU General Public License and also
10# under a 3-clause BSD license. Recipients may choose which of these licenses
11# to use; please see the files gpl-3.0.txt and/or bsd_license.txt,
12# respectively. If you choose the GPL option then the following text applies
13# (but note that there is still no warranty even if you opt for BSD instead):
14#
15# This program is free software: you can redistribute it and/or modify
16# it under the terms of the GNU General Public License as published by
17# the Free Software Foundation, either version 3 of the License, or
18# (at your option) any later version.
19#
20# This program is distributed in the hope that it will be useful,
21# but WITHOUT ANY WARRANTY; without even the implied warranty of
22# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
23# GNU General Public License for more details.
24#
25# You should have received a copy of the GNU General Public License
26# along with this program. If not, see <http://www.gnu.org/licenses/>.
28from __future__ import annotations
30__all__ = ("Instrument",)
32import contextlib
33import datetime
34from abc import ABCMeta, abstractmethod
35from collections.abc import Sequence
36from typing import TYPE_CHECKING, Any, Self, cast, final
38from lsst.daf.butler import DataCoordinate, DataId, DimensionPacker, DimensionRecord, Formatter, FormatterV2
39from lsst.daf.butler.registry import DataIdError
40from lsst.pex.config import Config, RegistryField
41from lsst.resources import ResourcePath, ResourcePathExpression
42from lsst.utils import doImportType
43from lsst.utils.introspection import get_full_type_name
45from ._observation_dimension_packer import observation_packer_registry
47if TYPE_CHECKING:
48 from lsst.daf.butler import Registry
51class Instrument(metaclass=ABCMeta):
52 """Base class for instrument-specific logic for the Gen3 Butler.
54 Parameters
55 ----------
56 collection_prefix : `str`, optional
57 Prefix for collection names to use instead of the instrument's own
58 name. This is primarily for use in simulated-data repositories, where
59 the instrument name may not be necessary and/or sufficient to
60 distinguish between collections.
62 Notes
63 -----
64 Concrete instrument subclasses must have the same construction signature as
65 the base class.
66 """
68 configPaths: Sequence[ResourcePathExpression] = ()
69 """Paths to config files to read for specific Tasks.
71 The paths in this list should contain files of the form `task.py`, for
72 each of the Tasks that requires special configuration.
73 """
75 policyName: str | None = None
76 """Instrument specific name to use when locating a policy or configuration
77 file in the file system."""
79 raw_definition: tuple[str, tuple[str, ...], str] | None = None
80 """Dataset type definition to use for "raw" datasets. This is a tuple
81 of the dataset type name, a tuple of dimension names, and the storage class
82 name. If `None` the ingest system will use its default definition."""
84 def __init__(self, collection_prefix: str | None = None):
85 if collection_prefix is None:
86 collection_prefix = self.getName()
87 self.collection_prefix = collection_prefix
89 @classmethod
90 @abstractmethod
91 def getName(cls) -> str:
92 """Return the short (dimension) name for this instrument.
94 This is not (in general) the same as the class name - it's what is used
95 as the value of the "instrument" field in data IDs, and is usually an
96 abbreviation of the full name.
97 """
98 raise NotImplementedError()
100 @abstractmethod
101 def register(self, registry: Registry, *, update: bool = False) -> None:
102 """Insert instrument, and other relevant records into `Registry`.
104 Parameters
105 ----------
106 registry : `lsst.daf.butler.Registry`
107 Registry client for the data repository to modify.
108 update : `bool`, optional
109 If `True` (`False` is default), update existing records if they
110 differ from the new ones.
112 Returns
113 -------
114 None
116 Raises
117 ------
118 lsst.daf.butler.registry.ConflictingDefinitionError
119 Raised if any existing record has the same key but a different
120 definition as one being registered.
122 Notes
123 -----
124 New records can always be added by calling this method multiple times,
125 as long as no existing records have changed (if existing records have
126 changed, ``update=True`` must be used). Old records can never be
127 removed by this method.
129 Implementations should guarantee that registration is atomic (the
130 registry should not be modified if any error occurs) and idempotent at
131 the level of individual dimension entries; new detectors and filters
132 should be added, but changes to any existing record should not be.
133 This can generally be achieved via a block like
134 """
135 raise NotImplementedError()
137 @classmethod
138 def fromName(cls, name: str, registry: Registry, collection_prefix: str | None = None) -> Self:
139 """Given an instrument name and a butler registry, retrieve a
140 corresponding instantiated instrument object.
142 Parameters
143 ----------
144 name : `str`
145 Name of the instrument (must match the return value of `getName`).
146 registry : `lsst.daf.butler.Registry`
147 Butler registry to query to find the information.
148 collection_prefix : `str`, optional
149 Prefix for collection names to use instead of the instrument's own
150 name. This is primarily for use in simulated-data repositories,
151 where the instrument name may not be necessary and/or sufficient to
152 distinguish between collections.
154 Returns
155 -------
156 instrument : `Instrument`
157 An instance of the relevant `Instrument`.
159 Notes
160 -----
161 The instrument must be registered in the corresponding butler.
163 Raises
164 ------
165 LookupError
166 Raised if the instrument is not known to the supplied registry.
167 ModuleNotFoundError
168 Raised if the class could not be imported. This could mean
169 that the relevant obs package has not been setup.
170 TypeError
171 Raised if the class name retrieved is not a string or the imported
172 symbol is not an `Instrument` subclass.
173 """
174 try:
175 records = list(registry.queryDimensionRecords("instrument", instrument=name))
176 except DataIdError:
177 records = None
178 if not records:
179 raise LookupError(f"No registered instrument with name '{name}'.")
180 cls_name = records[0].class_name
181 if not isinstance(cls_name, str):
182 raise TypeError(
183 f"Unexpected class name retrieved from {name} instrument dimension (got {cls_name})"
184 )
185 return cls._from_cls_name(cls_name, collection_prefix)
187 @classmethod
188 def from_string(
189 cls, name: str, registry: Registry | None = None, collection_prefix: str | None = None
190 ) -> Self:
191 """Return an instance from the short name or class name.
193 If the instrument name is not qualified (does not contain a '.') and a
194 butler registry is provided, this will attempt to load the instrument
195 using `Instrument.fromName()`. Otherwise the instrument will be
196 imported and instantiated.
198 Parameters
199 ----------
200 name : `str`
201 The name or fully-qualified class name of an instrument.
202 registry : `lsst.daf.butler.Registry`, optional
203 Butler registry to query to find information about the instrument,
204 by default `None`.
205 collection_prefix : `str`, optional
206 Prefix for collection names to use instead of the instrument's own
207 name. This is primarily for use in simulated-data repositories,
208 where the instrument name may not be necessary and/or sufficient
209 to distinguish between collections.
211 Returns
212 -------
213 instrument : `Instrument`
214 The instantiated instrument.
216 Raises
217 ------
218 RuntimeError
219 Raised if the instrument can not be imported, instantiated, or
220 obtained from the registry.
221 TypeError
222 Raised if the instrument is not a subclass of
223 `~lsst.pipe.base.Instrument`.
225 See Also
226 --------
227 Instrument.fromName : Constructing Instrument from a name.
228 """
229 if "." not in name and registry is not None:
230 try:
231 instr = cls.fromName(name, registry, collection_prefix=collection_prefix)
232 except Exception as err:
233 raise RuntimeError(
234 f"Could not get instrument from name: {name}. Failed with exception: {err}"
235 ) from err
236 else:
237 try:
238 instr_class = doImportType(name)
239 except Exception as err:
240 raise RuntimeError(
241 f"Could not import instrument: {name}. Failed with exception: {err}"
242 ) from err
243 instr = instr_class(collection_prefix=collection_prefix)
244 if not isinstance(instr, cls):
245 raise TypeError(f"{name} is not a {get_full_type_name(cls)} subclass.")
246 return instr
248 @classmethod
249 def from_data_id(cls, data_id: DataCoordinate, collection_prefix: str | None = None) -> Self:
250 """Instantiate an `Instrument` object from a fully-expanded data ID.
252 Parameters
253 ----------
254 data_id : `~lsst.daf.butler.DataCoordinate`
255 Expanded data ID that includes the instrument dimension.
256 collection_prefix : `str`, optional
257 Prefix for collection names to use instead of the instrument's own
258 name. This is primarily for use in simulated-data repositories,
259 where the instrument name may not be necessary and/or sufficient to
260 distinguish between collections.
262 Returns
263 -------
264 instrument : `Instrument`
265 An instance of the relevant `Instrument`.
267 Raises
268 ------
269 TypeError
270 Raised if the class name retrieved is not a string or the imported
271 symbol is not an `Instrument` subclass.
272 """
273 return cls._from_cls_name(
274 cast(DimensionRecord, data_id.records["instrument"]).class_name, collection_prefix
275 )
277 @classmethod
278 def _from_cls_name(cls, cls_name: str, collection_prefix: str | None = None) -> Self:
279 """Instantiate an `Instrument` object type name.
281 This just provides common error-handling for `fromName` and
282 `from_data_id`
284 Parameters
285 ----------
286 cls_name : `str`
287 Fully-qualified name of the type.
288 collection_prefix : `str`, optional
289 Prefix for collection names to use instead of the instrument's own
290 name. This is primarily for use in simulated-data repositories,
291 where the instrument name may not be necessary and/or sufficient to
292 distinguish between collections.
294 Returns
295 -------
296 instrument : `Instrument`
297 An instance of the relevant `Instrument`.
299 Raises
300 ------
301 TypeError
302 Raised if the class name retrieved is not a string or the imported
303 symbol is not an `Instrument` subclass.
304 """
305 instrument_cls: type = doImportType(cls_name)
306 if not issubclass(instrument_cls, cls):
307 raise TypeError(
308 f"{instrument_cls!r}, obtained from importing {cls_name}, is not a subclass "
309 f"of {get_full_type_name(cls)}."
310 )
311 return instrument_cls(collection_prefix=collection_prefix)
313 @staticmethod
314 def importAll(registry: Registry) -> dict[str, type[Instrument]]:
315 """Import all the instruments known to this registry.
317 This will ensure that all metadata translators have been registered.
319 Parameters
320 ----------
321 registry : `lsst.daf.butler.Registry`
322 Butler registry to query to find the information.
324 Returns
325 -------
326 imported : `dict` [`str`, `type` [`Instrument`]]
327 A mapping containing all the instrument classes that were loaded
328 successfully, keyed by their butler names.
330 Notes
331 -----
332 It is allowed for a particular instrument class to fail on import.
333 This might simply indicate that a particular obs package has
334 not been setup.
335 """
336 imported: dict[str, type[Instrument]] = {}
337 records = list(registry.queryDimensionRecords("instrument"))
338 for record in records:
339 cls = record.class_name
340 instrument_name: str = cast(str, record.name)
341 with contextlib.suppress(Exception):
342 instr = doImportType(cls)
343 assert issubclass(instr, Instrument)
344 imported[instrument_name] = instr
345 return imported
347 @abstractmethod
348 def getRawFormatter(self, dataId: DataId) -> type[Formatter | FormatterV2]:
349 """Return the Formatter class that should be used to read a particular
350 raw file.
352 Parameters
353 ----------
354 dataId : `DataId`
355 Dimension-based ID for the raw file or files being ingested.
357 Returns
358 -------
359 formatter : `type` \
360 [`lsst.daf.butler.Formatter` | `lsst.daf.butler.FormatterV2` ]
361 Class to be used that reads the file into the correct
362 Python object for the raw data.
363 """
364 raise NotImplementedError()
366 def applyConfigOverrides(self, name: str, config: Config) -> None:
367 """Apply instrument-specific overrides for a task config.
369 Parameters
370 ----------
371 name : `str`
372 Name of the object being configured; typically the _DefaultName
373 of a Task.
374 config : `lsst.pex.config.Config`
375 Config instance to which overrides should be applied.
376 """
377 for root in self.configPaths:
378 resource = ResourcePath(root, forceDirectory=True, forceAbsolute=True)
379 uri = resource.join(f"{name}.py", forceDirectory=False)
380 if uri.exists():
381 config.load(uri)
383 @staticmethod
384 def formatCollectionTimestamp(timestamp: str | datetime.datetime) -> str:
385 """Format a timestamp for use in a collection name.
387 Parameters
388 ----------
389 timestamp : `str` or `datetime.datetime`
390 Timestamp to format. May be a date or datetime string in extended
391 ISO format (assumed UTC), with or without a timezone specifier, a
392 datetime string in basic ISO format with a timezone specifier, a
393 naive `datetime.datetime` instance (assumed UTC) or a
394 timezone-aware `datetime.datetime` instance (converted to UTC).
395 This is intended to cover all forms that string ``CALIBDATE``
396 metadata values have taken in the past, as well as the format this
397 method itself writes out (to enable round-tripping).
399 Returns
400 -------
401 formatted : `str`
402 Standardized string form for the timestamp.
403 """
404 if isinstance(timestamp, str):
405 if "-" in timestamp:
406 # extended ISO format, with - and : delimiters
407 timestamp = datetime.datetime.fromisoformat(timestamp)
408 else:
409 # basic ISO format, with no delimiters (what this method
410 # returns)
411 timestamp = datetime.datetime.strptime(timestamp, "%Y%m%dT%H%M%S%z")
412 if not isinstance(timestamp, datetime.datetime):
413 raise TypeError(f"Unexpected date/time object: {timestamp!r}.")
414 if timestamp.tzinfo is not None:
415 timestamp = timestamp.astimezone(datetime.UTC)
416 return f"{timestamp:%Y%m%dT%H%M%S}Z"
418 @staticmethod
419 def makeCollectionTimestamp() -> str:
420 """Create a timestamp string for use in a collection name from the
421 current time.
423 Returns
424 -------
425 formatted : `str`
426 Standardized string form of the current time.
427 """
428 return Instrument.formatCollectionTimestamp(datetime.datetime.now(tz=datetime.UTC))
430 def makeDefaultRawIngestRunName(self) -> str:
431 """Make the default instrument-specific run collection string for raw
432 data ingest.
434 Returns
435 -------
436 coll : `str`
437 Run collection name to be used as the default for ingestion of
438 raws.
439 """
440 return self.makeCollectionName("raw", "all")
442 def makeUnboundedCalibrationRunName(self, *labels: str) -> str:
443 """Make a RUN collection name appropriate for inserting calibration
444 datasets whose validity ranges are unbounded.
446 Parameters
447 ----------
448 *labels : `str`
449 Extra strings to be included in the base name, using the default
450 delimiter for collection names. Usually this is the name of the
451 ticket on which the calibration collection is being created.
453 Returns
454 -------
455 name : `str`
456 Run collection name.
457 """
458 return self.makeCollectionName("calib", *labels, "unbounded")
460 def makeCuratedCalibrationRunName(self, calibDate: str, *labels: str) -> str:
461 """Make a RUN collection name appropriate for inserting curated
462 calibration datasets with the given ``CALIBDATE`` metadata value.
464 Parameters
465 ----------
466 calibDate : `str`
467 The ``CALIBDATE`` metadata value.
468 *labels : `str`
469 Strings to be included in the collection name (before
470 ``calibDate``, but after all other terms), using the default
471 delimiter for collection names. Usually this is the name of the
472 ticket on which the calibration collection is being created.
474 Returns
475 -------
476 name : `str`
477 Run collection name.
478 """
479 return self.makeCollectionName("calib", *labels, "curated", self.formatCollectionTimestamp(calibDate))
481 def makeCalibrationCollectionName(self, *labels: str) -> str:
482 """Make a CALIBRATION collection name appropriate for associating
483 calibration datasets with validity ranges.
485 Parameters
486 ----------
487 *labels : `str`
488 Strings to be appended to the base name, using the default
489 delimiter for collection names. Usually this is the name of the
490 ticket on which the calibration collection is being created.
492 Returns
493 -------
494 name : `str`
495 Calibration collection name.
496 """
497 return self.makeCollectionName("calib", *labels)
499 @staticmethod
500 def makeRefCatCollectionName(*labels: str) -> str:
501 """Return a global (not instrument-specific) name for a collection that
502 holds reference catalogs.
504 With no arguments, this returns the name of the collection that holds
505 all reference catalogs (usually a ``CHAINED`` collection, at least in
506 long-lived repos that may contain more than one reference catalog).
508 Parameters
509 ----------
510 *labels : `str`
511 Strings to be added to the global collection name, in order to
512 define a collection name for one or more reference catalogs being
513 ingested at the same time.
515 Returns
516 -------
517 name : `str`
518 Collection name.
520 Notes
521 -----
522 This is a ``staticmethod``, not a ``classmethod``, because it should
523 be the same for all instruments.
524 """
525 return "/".join(("refcats",) + labels)
527 def makeUmbrellaCollectionName(self) -> str:
528 """Return the name of the umbrella ``CHAINED`` collection for this
529 instrument that combines all standard recommended input collections.
531 This method should almost never be overridden by derived classes.
533 Returns
534 -------
535 name : `str`
536 Name for the umbrella collection.
537 """
538 return self.makeCollectionName("defaults")
540 def makeCollectionName(self, *labels: str) -> str:
541 """Get the instrument-specific collection string to use as derived
542 from the supplied labels.
544 Parameters
545 ----------
546 *labels : `str`
547 Strings to be combined with the instrument name to form a
548 collection name.
550 Returns
551 -------
552 name : `str`
553 Collection name to use that includes the instrument's recommended
554 prefix.
555 """
556 return "/".join((self.collection_prefix,) + labels)
558 @staticmethod
559 def make_dimension_packer_config_field(
560 doc: str = (
561 "How to pack visit+detector or exposure+detector data IDs into integers. "
562 "The default (None) is to delegate to the Instrument class for which "
563 "registered implementation to use (but still use the nested configuration "
564 "for that implementation)."
565 ),
566 ) -> RegistryField:
567 """Make an `lsst.pex.config.Field` that can be used to configure how
568 data IDs for this instrument are packed.
570 Parameters
571 ----------
572 doc : `str`, optional
573 Documentation for the config field.
575 Returns
576 -------
577 field : `lsst.pex.config.RegistryField`
578 A config field for which calling ``apply`` on the instance
579 attribute constructs an `lsst.daf.butler.DimensionPacker` that
580 defaults to the appropriate one for this instrument.
582 Notes
583 -----
584 This method is expected to be used whenever code requires a single
585 integer that represents the combination of a detector and either a
586 visit or exposure, but in most cases the `lsst.meas.base.IdGenerator`
587 class and its helper configs provide a simpler high-level interface
588 that should be used instead of calling this method directly.
590 This system is designed to work best when the configuration for the ID
591 packer is not overridden at all, allowing the appropriate instrument
592 class to determine the behavior for each data ID encountered. When the
593 configuration does need to be modified (most often when the scheme for
594 packing an instrument's data IDs is undergoing an upgrade), it is
595 important to ensure the overrides are only applied to data IDs with the
596 desired instrument value.
598 Unit tests of code that use a field produced by this method will often
599 want to explicitly set the packer to "observation" and manually set
600 its ``n_detectors`` and ``n_observations`` fields; this will make it
601 unnecessary for tests to provide expanded data IDs.
602 """
603 # The control flow here bounces around a bit when this RegistryField's
604 # apply() method is called, so it merits a thorough walkthrough
605 # somewhere, and that might as well be here:
606 #
607 # - If the config field's name is not `None`, that kind of packer is
608 # constructed and returned with the arguments to `apply`, in just the
609 # way it works with most RegistryFields or ConfigurableFields. But
610 # this is expected to be rare.
611 #
612 # - If the config fields' name is `None`, the `apply` method (which
613 # actually lives on the `pex.config.RegistryInstanceDict` class,
614 # since `RegistryField` is a descriptor), calls
615 # `_make_default_dimension_packer_dispatch` (which is final, and
616 # hence the base class implementation just below is the only one).
617 #
618 # - `_make_default_dimension_packer_dispatch` instantiates an
619 # `Instrument` instance of the type pointed at by the data ID (i.e.
620 # calling `Instrument.from_data_id`), then calls
621 # `_make_default_dimension_packer` on that.
622 #
623 # - The default implementation of `_make_default_dimension_packer` here
624 # in the base class picks the "observation" dimension packer, so if
625 # it's not overridden by a derived class everything proceeds as if
626 # the config field's name was set to that. Note that this sets which
627 # item in the registry is used, but it still pays attention to the
628 # configuration for that entry in the registry field.
629 #
630 # - A subclass implementation of `_make_default_dimension_packer` will
631 # take precedence over the base class, but it's expected that these
632 # will usually just delegate back to ``super()`` while changing the
633 # ``default`` argument to something other than "observation". Once
634 # again, this will control which packer entry in the registry is used
635 # but the result will still reflect the configuration for that packer
636 # in the registry field.
637 #
638 return observation_packer_registry.makeField(
639 doc, default=None, optional=True, on_none=Instrument._make_default_dimension_packer_dispatch
640 )
642 @staticmethod
643 @final
644 def make_default_dimension_packer(
645 data_id: DataCoordinate, is_exposure: bool | None = None
646 ) -> DimensionPacker:
647 """Return the default dimension packer for the given data ID.
649 Parameters
650 ----------
651 data_id : `lsst.daf.butler.DataCoordinate`
652 Data ID that identifies at least the ``instrument`` dimension. Must
653 have dimension records attached.
654 is_exposure : `bool`, optional
655 If `False`, construct a packer for visit+detector data IDs. If
656 `True`, construct a packer for exposure+detector data IDs. If
657 `None`, this is determined based on whether ``visit`` or
658 ``exposure`` is present in ``data_id``, with ``visit`` checked
659 first and hence used if both are present.
661 Returns
662 -------
663 packer : `lsst.daf.butler.DimensionPacker`
664 Object that packs {visit, detector} or {exposure, detector} data
665 IDs into integers.
667 Notes
668 -----
669 When using a dimension packer in task code, using
670 `make_dimension_packer_config_field` to make the packing algorithm
671 configurable is preferred over this method.
673 When obtaining a dimension packer to unpack IDs that were packed by
674 task code, it is similarly preferable to load the configuration for
675 that task and the existing packer configuration field there, to ensure
676 any config overrides are respected. That is sometimes quite difficult,
677 however, and since config overrides for dimension packers are expected
678 to be exceedingly rare, using this simpler method will almost always
679 work.
680 """
682 class _DummyConfig(Config):
683 packer = Instrument.make_dimension_packer_config_field()
685 config = _DummyConfig()
687 return config.packer.apply(data_id, is_exposure=is_exposure) # type: ignore
689 @staticmethod
690 @final
691 def _make_default_dimension_packer_dispatch(
692 config_dict: Any, data_id: DataCoordinate, is_exposure: bool | None = None
693 ) -> DimensionPacker:
694 """Dispatch method used to invoke `_make_dimension_packer`.
696 This method constructs the appropriate `Instrument` subclass from
697 config and then calls its `_make_default_dimension_packer`.
698 It is called when (as usual) the field returned by
699 `make_dimension_packer_config_field` is left to its default selection
700 of `None`.
702 All arguments and return values are the same as
703 `_make_default_dimension_packer.`
704 """
705 instrument = Instrument.from_data_id(data_id)
706 return instrument._make_default_dimension_packer(config_dict, data_id, is_exposure=is_exposure)
708 def _make_default_dimension_packer(
709 self,
710 config_dict: Any,
711 data_id: DataCoordinate,
712 is_exposure: bool | None = None,
713 default: str = "observation",
714 ) -> DimensionPacker:
715 """Construct return the default dimension packer for this instrument.
717 This method is a protected hook for subclasses to override the behavior
718 of `make_dimension_packer_config_field` when the packer is not selected
719 explicitly via configuration.
721 Parameters
722 ----------
723 config_dict
724 Mapping attribute of a `lsst.pex.config.Config` instance that
725 corresponds to a field created by `make_dimension_packer_config`
726 (the actual type of this object is a `lsst.pex.config`
727 implementation detail).
728 data_id : `lsst.daf.butler.DataCoordinate`
729 Data ID that identifies at least the ``instrument`` dimension. For
730 most configurations this must have dimension records attached.
731 is_exposure : `bool`, optional
732 If `False`, construct a packer for visit+detector data IDs. If
733 `True`, construct a packer for exposure+detector data IDs. If
734 `None`, this is determined based on whether ``visit`` or
735 ``exposure`` is present in ``data_id``, with ``visit`` checked
736 first and hence used if both are present.
737 default : `str`, optional
738 Registered name of the dimension packer to select when the
739 configured packer is `None` (as is usually the case). This is
740 intended primarily for derived classes delegating to `super` in
741 reimplementations of this method.
743 Returns
744 -------
745 packer : `lsst.daf.butler.DimensionPacker`
746 Object that packs {visit, detector} or {exposure, detector} data
747 IDs into integers.
748 """
749 return config_dict.apply_with(default, data_id, is_exposure=is_exposure)