Coverage for python/lsst/daf/butler/direct_butler.py: 11%

689 statements  

« prev     ^ index     » next       coverage.py v7.3.2, created at 2023-10-27 09:44 +0000

1# This file is part of daf_butler. 

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 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/>. 

27 

28"""Butler top level classes. 

29""" 

30from __future__ import annotations 

31 

32__all__ = ( 

33 "DirectButler", 

34 "ButlerValidationError", 

35) 

36 

37import collections.abc 

38import contextlib 

39import io 

40import logging 

41import numbers 

42import os 

43import warnings 

44from collections import Counter, defaultdict 

45from collections.abc import Iterable, Iterator, MutableMapping, Sequence 

46from typing import TYPE_CHECKING, Any, ClassVar, TextIO 

47 

48from deprecated.sphinx import deprecated 

49from lsst.resources import ResourcePath, ResourcePathExpression 

50from lsst.utils.introspection import get_class_of 

51from lsst.utils.logging import VERBOSE, getLogger 

52from sqlalchemy.exc import IntegrityError 

53 

54from ._butler import Butler 

55from ._butler_config import ButlerConfig 

56from ._config import Config 

57from ._dataset_existence import DatasetExistence 

58from ._dataset_ref import DatasetIdGenEnum, DatasetRef 

59from ._dataset_type import DatasetType 

60from ._deferredDatasetHandle import DeferredDatasetHandle 

61from ._exceptions import ValidationError 

62from ._file_dataset import FileDataset 

63from ._limited_butler import LimitedButler 

64from ._registry_shim import RegistryShim 

65from ._storage_class import StorageClass, StorageClassFactory 

66from ._timespan import Timespan 

67from .datastore import DatasetRefURIs, Datastore, NullDatastore 

68from .dimensions import ( 

69 DataCoordinate, 

70 DataId, 

71 DataIdValue, 

72 Dimension, 

73 DimensionElement, 

74 DimensionRecord, 

75 DimensionUniverse, 

76) 

77from .progress import Progress 

78from .registry import ( 

79 CollectionType, 

80 ConflictingDefinitionError, 

81 DataIdError, 

82 MissingDatasetTypeError, 

83 NoDefaultCollectionError, 

84 Registry, 

85 RegistryDefaults, 

86 _RegistryFactory, 

87) 

88from .registry.sql_registry import SqlRegistry 

89from .transfers import RepoExportContext 

90from .utils import transactional 

91 

92if TYPE_CHECKING: 

93 from lsst.resources import ResourceHandleProtocol 

94 

95 from .transfers import RepoImportBackend 

96 

97_LOG = getLogger(__name__) 

98 

99 

100class ButlerValidationError(ValidationError): 

101 """There is a problem with the Butler configuration.""" 

102 

103 pass 

104 

105 

106class DirectButler(Butler): 

107 """Main entry point for the data access system. 

108 

109 Parameters 

110 ---------- 

111 config : `ButlerConfig`, `Config` or `str`, optional. 

112 Configuration. Anything acceptable to the 

113 `ButlerConfig` constructor. If a directory path 

114 is given the configuration will be read from a ``butler.yaml`` file in 

115 that location. If `None` is given default values will be used. 

116 butler : `DirectButler`, optional. 

117 If provided, construct a new Butler that uses the same registry and 

118 datastore as the given one, but with the given collection and run. 

119 Incompatible with the ``config``, ``searchPaths``, and ``writeable`` 

120 arguments. 

121 collections : `str` or `~collections.abc.Iterable` [ `str` ], optional 

122 An expression specifying the collections to be searched (in order) when 

123 reading datasets. 

124 This may be a `str` collection name or an iterable thereof. 

125 See :ref:`daf_butler_collection_expressions` for more information. 

126 These collections are not registered automatically and must be 

127 manually registered before they are used by any method, but they may be 

128 manually registered after the `Butler` is initialized. 

129 run : `str`, optional 

130 Name of the `~CollectionType.RUN` collection new datasets should be 

131 inserted into. If ``collections`` is `None` and ``run`` is not `None`, 

132 ``collections`` will be set to ``[run]``. If not `None`, this 

133 collection will automatically be registered. If this is not set (and 

134 ``writeable`` is not set either), a read-only butler will be created. 

135 searchPaths : `list` of `str`, optional 

136 Directory paths to search when calculating the full Butler 

137 configuration. Not used if the supplied config is already a 

138 `ButlerConfig`. 

139 writeable : `bool`, optional 

140 Explicitly sets whether the butler supports write operations. If not 

141 provided, a read-write butler is created if any of ``run``, ``tags``, 

142 or ``chains`` is non-empty. 

143 inferDefaults : `bool`, optional 

144 If `True` (default) infer default data ID values from the values 

145 present in the datasets in ``collections``: if all collections have the 

146 same value (or no value) for a governor dimension, that value will be 

147 the default for that dimension. Nonexistent collections are ignored. 

148 If a default value is provided explicitly for a governor dimension via 

149 ``**kwargs``, no default will be inferred for that dimension. 

150 without_datastore : `bool`, optional 

151 If `True` do not attach a datastore to this butler. Any attempts 

152 to use a datastore will fail. 

153 **kwargs : `str` 

154 Default data ID key-value pairs. These may only identify "governor" 

155 dimensions like ``instrument`` and ``skymap``. 

156 """ 

157 

158 def __init__( 

159 self, 

160 config: Config | ResourcePathExpression | None = None, 

161 *, 

162 butler: DirectButler | None = None, 

163 collections: Any = None, 

164 run: str | None = None, 

165 searchPaths: Sequence[ResourcePathExpression] | None = None, 

166 writeable: bool | None = None, 

167 inferDefaults: bool = True, 

168 without_datastore: bool = False, 

169 **kwargs: str, 

170 ): 

171 defaults = RegistryDefaults(collections=collections, run=run, infer=inferDefaults, **kwargs) 

172 # Load registry, datastore, etc. from config or existing butler. 

173 if butler is not None: 

174 if config is not None or searchPaths is not None or writeable is not None: 

175 raise TypeError( 

176 "Cannot pass 'config', 'searchPaths', or 'writeable' arguments with 'butler' argument." 

177 ) 

178 self._registry = butler._registry.copy(defaults) 

179 self._datastore = butler._datastore 

180 self.storageClasses = butler.storageClasses 

181 self._config: ButlerConfig = butler._config 

182 else: 

183 self._config = ButlerConfig(config, searchPaths=searchPaths, without_datastore=without_datastore) 

184 try: 

185 butlerRoot = self._config.get("root", self._config.configDir) 

186 if writeable is None: 

187 writeable = run is not None 

188 self._registry = _RegistryFactory(self._config).from_config( 

189 butlerRoot=butlerRoot, writeable=writeable, defaults=defaults 

190 ) 

191 if without_datastore: 

192 self._datastore = NullDatastore(None, None) 

193 else: 

194 self._datastore = Datastore.fromConfig( 

195 self._config, self._registry.getDatastoreBridgeManager(), butlerRoot=butlerRoot 

196 ) 

197 # TODO: Once datastore drops dependency on registry we can 

198 # construct datastore first and pass opaque tables to registry 

199 # constructor. 

200 self._registry.make_datastore_tables(self._datastore.get_opaque_table_definitions()) 

201 self.storageClasses = StorageClassFactory() 

202 self.storageClasses.addFromConfig(self._config) 

203 except Exception: 

204 # Failures here usually mean that configuration is incomplete, 

205 # just issue an error message which includes config file URI. 

206 _LOG.error(f"Failed to instantiate Butler from config {self._config.configFile}.") 

207 raise 

208 

209 # For execution butler the datastore needs a special 

210 # dependency-inversion trick. This is not used by regular butler, 

211 # but we do not have a way to distinguish regular butler from execution 

212 # butler. 

213 self._datastore.set_retrieve_dataset_type_method(self._retrieve_dataset_type) 

214 

215 if "run" in self._config or "collection" in self._config: 

216 raise ValueError("Passing a run or collection via configuration is no longer supported.") 

217 

218 self._registry_shim = RegistryShim(self) 

219 

220 GENERATION: ClassVar[int] = 3 

221 """This is a Generation 3 Butler. 

222 

223 This attribute may be removed in the future, once the Generation 2 Butler 

224 interface has been fully retired; it should only be used in transitional 

225 code. 

226 """ 

227 

228 def _retrieve_dataset_type(self, name: str) -> DatasetType | None: 

229 """Return DatasetType defined in registry given dataset type name.""" 

230 try: 

231 return self._registry.getDatasetType(name) 

232 except MissingDatasetTypeError: 

233 return None 

234 

235 @classmethod 

236 def _unpickle( 

237 cls, 

238 config: ButlerConfig, 

239 collections: tuple[str, ...] | None, 

240 run: str | None, 

241 defaultDataId: dict[str, str], 

242 writeable: bool, 

243 ) -> DirectButler: 

244 """Callable used to unpickle a Butler. 

245 

246 We prefer not to use ``Butler.__init__`` directly so we can force some 

247 of its many arguments to be keyword-only (note that ``__reduce__`` 

248 can only invoke callables with positional arguments). 

249 

250 Parameters 

251 ---------- 

252 config : `ButlerConfig` 

253 Butler configuration, already coerced into a true `ButlerConfig` 

254 instance (and hence after any search paths for overrides have been 

255 utilized). 

256 collections : `tuple` [ `str` ] 

257 Names of the default collections to read from. 

258 run : `str`, optional 

259 Name of the default `~CollectionType.RUN` collection to write to. 

260 defaultDataId : `dict` [ `str`, `str` ] 

261 Default data ID values. 

262 writeable : `bool` 

263 Whether the Butler should support write operations. 

264 

265 Returns 

266 ------- 

267 butler : `Butler` 

268 A new `Butler` instance. 

269 """ 

270 # MyPy doesn't recognize that the kwargs below are totally valid; it 

271 # seems to think '**defaultDataId* is a _positional_ argument! 

272 return cls( 

273 config=config, 

274 collections=collections, 

275 run=run, 

276 writeable=writeable, 

277 **defaultDataId, # type: ignore 

278 ) 

279 

280 def __reduce__(self) -> tuple: 

281 """Support pickling.""" 

282 return ( 

283 DirectButler._unpickle, 

284 ( 

285 self._config, 

286 self.collections, 

287 self.run, 

288 self._registry.defaults.dataId.byName(), 

289 self._registry.isWriteable(), 

290 ), 

291 ) 

292 

293 def __str__(self) -> str: 

294 return "Butler(collections={}, run={}, datastore='{}', registry='{}')".format( 

295 self.collections, self.run, self._datastore, self._registry 

296 ) 

297 

298 def isWriteable(self) -> bool: 

299 # Docstring inherited. 

300 return self._registry.isWriteable() 

301 

302 @contextlib.contextmanager 

303 def transaction(self) -> Iterator[None]: 

304 """Context manager supporting `Butler` transactions. 

305 

306 Transactions can be nested. 

307 """ 

308 with self._registry.transaction(), self._datastore.transaction(): 

309 yield 

310 

311 def _standardizeArgs( 

312 self, 

313 datasetRefOrType: DatasetRef | DatasetType | str, 

314 dataId: DataId | None = None, 

315 for_put: bool = True, 

316 **kwargs: Any, 

317 ) -> tuple[DatasetType, DataId | None]: 

318 """Standardize the arguments passed to several Butler APIs. 

319 

320 Parameters 

321 ---------- 

322 datasetRefOrType : `DatasetRef`, `DatasetType`, or `str` 

323 When `DatasetRef` the `dataId` should be `None`. 

324 Otherwise the `DatasetType` or name thereof. 

325 dataId : `dict` or `DataCoordinate` 

326 A `dict` of `Dimension` link name, value pairs that label the 

327 `DatasetRef` within a Collection. When `None`, a `DatasetRef` 

328 should be provided as the second argument. 

329 for_put : `bool`, optional 

330 If `True` this call is invoked as part of a `Butler.put()`. 

331 Otherwise it is assumed to be part of a `Butler.get()`. This 

332 parameter is only relevant if there is dataset type 

333 inconsistency. 

334 **kwargs 

335 Additional keyword arguments used to augment or construct a 

336 `DataCoordinate`. See `DataCoordinate.standardize` 

337 parameters. 

338 

339 Returns 

340 ------- 

341 datasetType : `DatasetType` 

342 A `DatasetType` instance extracted from ``datasetRefOrType``. 

343 dataId : `dict` or `DataId`, optional 

344 Argument that can be used (along with ``kwargs``) to construct a 

345 `DataId`. 

346 

347 Notes 

348 ----- 

349 Butler APIs that conceptually need a DatasetRef also allow passing a 

350 `DatasetType` (or the name of one) and a `DataId` (or a dict and 

351 keyword arguments that can be used to construct one) separately. This 

352 method accepts those arguments and always returns a true `DatasetType` 

353 and a `DataId` or `dict`. 

354 

355 Standardization of `dict` vs `DataId` is best handled by passing the 

356 returned ``dataId`` (and ``kwargs``) to `Registry` APIs, which are 

357 generally similarly flexible. 

358 """ 

359 externalDatasetType: DatasetType | None = None 

360 internalDatasetType: DatasetType | None = None 

361 if isinstance(datasetRefOrType, DatasetRef): 

362 if dataId is not None or kwargs: 

363 raise ValueError("DatasetRef given, cannot use dataId as well") 

364 externalDatasetType = datasetRefOrType.datasetType 

365 dataId = datasetRefOrType.dataId 

366 else: 

367 # Don't check whether DataId is provided, because Registry APIs 

368 # can usually construct a better error message when it wasn't. 

369 if isinstance(datasetRefOrType, DatasetType): 

370 externalDatasetType = datasetRefOrType 

371 else: 

372 internalDatasetType = self._registry.getDatasetType(datasetRefOrType) 

373 

374 # Check that they are self-consistent 

375 if externalDatasetType is not None: 

376 internalDatasetType = self._registry.getDatasetType(externalDatasetType.name) 

377 if externalDatasetType != internalDatasetType: 

378 # We can allow differences if they are compatible, depending 

379 # on whether this is a get or a put. A get requires that 

380 # the python type associated with the datastore can be 

381 # converted to the user type. A put requires that the user 

382 # supplied python type can be converted to the internal 

383 # type expected by registry. 

384 relevantDatasetType = internalDatasetType 

385 if for_put: 

386 is_compatible = internalDatasetType.is_compatible_with(externalDatasetType) 

387 else: 

388 is_compatible = externalDatasetType.is_compatible_with(internalDatasetType) 

389 relevantDatasetType = externalDatasetType 

390 if not is_compatible: 

391 raise ValueError( 

392 f"Supplied dataset type ({externalDatasetType}) inconsistent with " 

393 f"registry definition ({internalDatasetType})" 

394 ) 

395 # Override the internal definition. 

396 internalDatasetType = relevantDatasetType 

397 

398 assert internalDatasetType is not None 

399 return internalDatasetType, dataId 

400 

401 def _rewrite_data_id( 

402 self, dataId: DataId | None, datasetType: DatasetType, **kwargs: Any 

403 ) -> tuple[DataId | None, dict[str, Any]]: 

404 """Rewrite a data ID taking into account dimension records. 

405 

406 Take a Data ID and keyword args and rewrite it if necessary to 

407 allow the user to specify dimension records rather than dimension 

408 primary values. 

409 

410 This allows a user to include a dataId dict with keys of 

411 ``exposure.day_obs`` and ``exposure.seq_num`` instead of giving 

412 the integer exposure ID. It also allows a string to be given 

413 for a dimension value rather than the integer ID if that is more 

414 convenient. For example, rather than having to specifying the 

415 detector with ``detector.full_name``, a string given for ``detector`` 

416 will be interpreted as the full name and converted to the integer 

417 value. 

418 

419 Keyword arguments can also use strings for dimensions like detector 

420 and exposure but python does not allow them to include ``.`` and 

421 so the ``exposure.day_obs`` syntax can not be used in a keyword 

422 argument. 

423 

424 Parameters 

425 ---------- 

426 dataId : `dict` or `DataCoordinate` 

427 A `dict` of `Dimension` link name, value pairs that will label the 

428 `DatasetRef` within a Collection. 

429 datasetType : `DatasetType` 

430 The dataset type associated with this dataId. Required to 

431 determine the relevant dimensions. 

432 **kwargs 

433 Additional keyword arguments used to augment or construct a 

434 `DataId`. See `DataId` parameters. 

435 

436 Returns 

437 ------- 

438 dataId : `dict` or `DataCoordinate` 

439 The, possibly rewritten, dataId. If given a `DataCoordinate` and 

440 no keyword arguments, the original dataId will be returned 

441 unchanged. 

442 **kwargs : `dict` 

443 Any unused keyword arguments (would normally be empty dict). 

444 """ 

445 # Do nothing if we have a standalone DataCoordinate. 

446 if isinstance(dataId, DataCoordinate) and not kwargs: 

447 return dataId, kwargs 

448 

449 # Process dimension records that are using record information 

450 # rather than ids 

451 newDataId: dict[str, DataIdValue] = {} 

452 byRecord: dict[str, dict[str, Any]] = defaultdict(dict) 

453 

454 # if all the dataId comes from keyword parameters we do not need 

455 # to do anything here because they can't be of the form 

456 # exposure.obs_id because a "." is not allowed in a keyword parameter. 

457 if dataId: 

458 for k, v in dataId.items(): 

459 # If we have a Dimension we do not need to do anything 

460 # because it cannot be a compound key. 

461 if isinstance(k, str) and "." in k: 

462 # Someone is using a more human-readable dataId 

463 dimensionName, record = k.split(".", 1) 

464 byRecord[dimensionName][record] = v 

465 elif isinstance(k, Dimension): 

466 newDataId[k.name] = v 

467 else: 

468 newDataId[k] = v 

469 

470 # Go through the updated dataId and check the type in case someone is 

471 # using an alternate key. We have already filtered out the compound 

472 # keys dimensions.record format. 

473 not_dimensions = {} 

474 

475 # Will need to look in the dataId and the keyword arguments 

476 # and will remove them if they need to be fixed or are unrecognized. 

477 for dataIdDict in (newDataId, kwargs): 

478 # Use a list so we can adjust the dict safely in the loop 

479 for dimensionName in list(dataIdDict): 

480 value = dataIdDict[dimensionName] 

481 try: 

482 dimension = self.dimensions.getStaticDimensions()[dimensionName] 

483 except KeyError: 

484 # This is not a real dimension 

485 not_dimensions[dimensionName] = value 

486 del dataIdDict[dimensionName] 

487 continue 

488 

489 # Convert an integral type to an explicit int to simplify 

490 # comparisons here 

491 if isinstance(value, numbers.Integral): 

492 value = int(value) 

493 

494 if not isinstance(value, dimension.primaryKey.getPythonType()): 

495 for alternate in dimension.alternateKeys: 

496 if isinstance(value, alternate.getPythonType()): 

497 byRecord[dimensionName][alternate.name] = value 

498 del dataIdDict[dimensionName] 

499 _LOG.debug( 

500 "Converting dimension %s to %s.%s=%s", 

501 dimensionName, 

502 dimensionName, 

503 alternate.name, 

504 value, 

505 ) 

506 break 

507 else: 

508 _LOG.warning( 

509 "Type mismatch found for value '%r' provided for dimension %s. " 

510 "Could not find matching alternative (primary key has type %s) " 

511 "so attempting to use as-is.", 

512 value, 

513 dimensionName, 

514 dimension.primaryKey.getPythonType(), 

515 ) 

516 

517 # By this point kwargs and newDataId should only include valid 

518 # dimensions. Merge kwargs in to the new dataId and log if there 

519 # are dimensions in both (rather than calling update). 

520 for k, v in kwargs.items(): 

521 if k in newDataId and newDataId[k] != v: 

522 _LOG.debug( 

523 "Keyword arg %s overriding explicit value in dataId of %s with %s", k, newDataId[k], v 

524 ) 

525 newDataId[k] = v 

526 # No need to retain any values in kwargs now. 

527 kwargs = {} 

528 

529 # If we have some unrecognized dimensions we have to try to connect 

530 # them to records in other dimensions. This is made more complicated 

531 # by some dimensions having records with clashing names. A mitigation 

532 # is that we can tell by this point which dimensions are missing 

533 # for the DatasetType but this does not work for calibrations 

534 # where additional dimensions can be used to constrain the temporal 

535 # axis. 

536 if not_dimensions: 

537 # Search for all dimensions even if we have been given a value 

538 # explicitly. In some cases records are given as well as the 

539 # actually dimension and this should not be an error if they 

540 # match. 

541 mandatoryDimensions = datasetType.dimensions.names # - provided 

542 

543 candidateDimensions: set[str] = set() 

544 candidateDimensions.update(mandatoryDimensions) 

545 

546 # For calibrations we may well be needing temporal dimensions 

547 # so rather than always including all dimensions in the scan 

548 # restrict things a little. It is still possible for there 

549 # to be confusion over day_obs in visit vs exposure for example. 

550 # If we are not searching calibration collections things may 

551 # fail but they are going to fail anyway because of the 

552 # ambiguousness of the dataId... 

553 if datasetType.isCalibration(): 

554 for dim in self.dimensions.getStaticDimensions(): 

555 if dim.temporal: 

556 candidateDimensions.add(str(dim)) 

557 

558 # Look up table for the first association with a dimension 

559 guessedAssociation: dict[str, dict[str, Any]] = defaultdict(dict) 

560 

561 # Keep track of whether an item is associated with multiple 

562 # dimensions. 

563 counter: Counter[str] = Counter() 

564 assigned: dict[str, set[str]] = defaultdict(set) 

565 

566 # Go through the missing dimensions and associate the 

567 # given names with records within those dimensions 

568 matched_dims = set() 

569 for dimensionName in candidateDimensions: 

570 dimension = self.dimensions.getStaticDimensions()[dimensionName] 

571 fields = dimension.metadata.names | dimension.uniqueKeys.names 

572 for field in not_dimensions: 

573 if field in fields: 

574 guessedAssociation[dimensionName][field] = not_dimensions[field] 

575 counter[dimensionName] += 1 

576 assigned[field].add(dimensionName) 

577 matched_dims.add(field) 

578 

579 # Calculate the fields that matched nothing. 

580 never_found = set(not_dimensions) - matched_dims 

581 

582 if never_found: 

583 raise ValueError(f"Unrecognized keyword args given: {never_found}") 

584 

585 # There is a chance we have allocated a single dataId item 

586 # to multiple dimensions. Need to decide which should be retained. 

587 # For now assume that the most popular alternative wins. 

588 # This means that day_obs with seq_num will result in 

589 # exposure.day_obs and not visit.day_obs 

590 # Also prefer an explicitly missing dimension over an inferred 

591 # temporal dimension. 

592 for fieldName, assignedDimensions in assigned.items(): 

593 if len(assignedDimensions) > 1: 

594 # Pick the most popular (preferring mandatory dimensions) 

595 requiredButMissing = assignedDimensions.intersection(mandatoryDimensions) 

596 if requiredButMissing: 

597 candidateDimensions = requiredButMissing 

598 else: 

599 candidateDimensions = assignedDimensions 

600 

601 # If this is a choice between visit and exposure and 

602 # neither was a required part of the dataset type, 

603 # (hence in this branch) always prefer exposure over 

604 # visit since exposures are always defined and visits 

605 # are defined from exposures. 

606 if candidateDimensions == {"exposure", "visit"}: 

607 candidateDimensions = {"exposure"} 

608 

609 # Select the relevant items and get a new restricted 

610 # counter. 

611 theseCounts = {k: v for k, v in counter.items() if k in candidateDimensions} 

612 duplicatesCounter: Counter[str] = Counter() 

613 duplicatesCounter.update(theseCounts) 

614 

615 # Choose the most common. If they are equally common 

616 # we will pick the one that was found first. 

617 # Returns a list of tuples 

618 selected = duplicatesCounter.most_common(1)[0][0] 

619 

620 _LOG.debug( 

621 "Ambiguous dataId entry '%s' associated with multiple dimensions: %s." 

622 " Removed ambiguity by choosing dimension %s.", 

623 fieldName, 

624 ", ".join(assignedDimensions), 

625 selected, 

626 ) 

627 

628 for candidateDimension in assignedDimensions: 

629 if candidateDimension != selected: 

630 del guessedAssociation[candidateDimension][fieldName] 

631 

632 # Update the record look up dict with the new associations 

633 for dimensionName, values in guessedAssociation.items(): 

634 if values: # A dict might now be empty 

635 _LOG.debug( 

636 "Assigned non-dimension dataId keys to dimension %s: %s", dimensionName, values 

637 ) 

638 byRecord[dimensionName].update(values) 

639 

640 if byRecord: 

641 # Some record specifiers were found so we need to convert 

642 # them to the Id form 

643 for dimensionName, values in byRecord.items(): 

644 if dimensionName in newDataId: 

645 _LOG.debug( 

646 "DataId specified explicit %s dimension value of %s in addition to" 

647 " general record specifiers for it of %s. Ignoring record information.", 

648 dimensionName, 

649 newDataId[dimensionName], 

650 str(values), 

651 ) 

652 # Get the actual record and compare with these values. 

653 try: 

654 recs = list(self._registry.queryDimensionRecords(dimensionName, dataId=newDataId)) 

655 except DataIdError: 

656 raise ValueError( 

657 f"Could not find dimension '{dimensionName}'" 

658 f" with dataId {newDataId} as part of comparing with" 

659 f" record values {byRecord[dimensionName]}" 

660 ) from None 

661 if len(recs) == 1: 

662 errmsg: list[str] = [] 

663 for k, v in values.items(): 

664 if (recval := getattr(recs[0], k)) != v: 

665 errmsg.append(f"{k}({recval} != {v})") 

666 if errmsg: 

667 raise ValueError( 

668 f"Dimension {dimensionName} in dataId has explicit value" 

669 " inconsistent with records: " + ", ".join(errmsg) 

670 ) 

671 else: 

672 # Multiple matches for an explicit dimension 

673 # should never happen but let downstream complain. 

674 pass 

675 continue 

676 

677 # Build up a WHERE expression 

678 bind = dict(values.items()) 

679 where = " AND ".join(f"{dimensionName}.{k} = {k}" for k in bind) 

680 

681 # Hopefully we get a single record that matches 

682 records = set( 

683 self._registry.queryDimensionRecords( 

684 dimensionName, dataId=newDataId, where=where, bind=bind, **kwargs 

685 ) 

686 ) 

687 

688 if len(records) != 1: 

689 if len(records) > 1: 

690 # visit can have an ambiguous answer without involving 

691 # visit_system. The default visit_system is defined 

692 # by the instrument. 

693 if ( 

694 dimensionName == "visit" 

695 and "visit_system_membership" in self.dimensions 

696 and "visit_system" in self.dimensions["instrument"].metadata 

697 ): 

698 instrument_records = list( 

699 self._registry.queryDimensionRecords( 

700 "instrument", 

701 dataId=newDataId, 

702 **kwargs, 

703 ) 

704 ) 

705 if len(instrument_records) == 1: 

706 visit_system = instrument_records[0].visit_system 

707 if visit_system is None: 

708 # Set to a value that will never match. 

709 visit_system = -1 

710 

711 # Look up each visit in the 

712 # visit_system_membership records. 

713 for rec in records: 

714 membership = list( 

715 self._registry.queryDimensionRecords( 

716 # Use bind to allow zero results. 

717 # This is a fully-specified query. 

718 "visit_system_membership", 

719 where="instrument = inst AND visit_system = system AND visit = v", 

720 bind=dict( 

721 inst=instrument_records[0].name, system=visit_system, v=rec.id 

722 ), 

723 ) 

724 ) 

725 if membership: 

726 # This record is the right answer. 

727 records = {rec} 

728 break 

729 

730 # The ambiguity may have been resolved so check again. 

731 if len(records) > 1: 

732 _LOG.debug( 

733 "Received %d records from constraints of %s", len(records), str(values) 

734 ) 

735 for r in records: 

736 _LOG.debug("- %s", str(r)) 

737 raise ValueError( 

738 f"DataId specification for dimension {dimensionName} is not" 

739 f" uniquely constrained to a single dataset by {values}." 

740 f" Got {len(records)} results." 

741 ) 

742 else: 

743 raise ValueError( 

744 f"DataId specification for dimension {dimensionName} matched no" 

745 f" records when constrained by {values}" 

746 ) 

747 

748 # Get the primary key from the real dimension object 

749 dimension = self.dimensions.getStaticDimensions()[dimensionName] 

750 if not isinstance(dimension, Dimension): 

751 raise RuntimeError( 

752 f"{dimension.name} is not a true dimension, and cannot be used in data IDs." 

753 ) 

754 newDataId[dimensionName] = getattr(records.pop(), dimension.primaryKey.name) 

755 

756 return newDataId, kwargs 

757 

758 def _findDatasetRef( 

759 self, 

760 datasetRefOrType: DatasetRef | DatasetType | str, 

761 dataId: DataId | None = None, 

762 *, 

763 collections: Any = None, 

764 predict: bool = False, 

765 run: str | None = None, 

766 datastore_records: bool = False, 

767 **kwargs: Any, 

768 ) -> DatasetRef: 

769 """Shared logic for methods that start with a search for a dataset in 

770 the registry. 

771 

772 Parameters 

773 ---------- 

774 datasetRefOrType : `DatasetRef`, `DatasetType`, or `str` 

775 When `DatasetRef` the `dataId` should be `None`. 

776 Otherwise the `DatasetType` or name thereof. 

777 dataId : `dict` or `DataCoordinate`, optional 

778 A `dict` of `Dimension` link name, value pairs that label the 

779 `DatasetRef` within a Collection. When `None`, a `DatasetRef` 

780 should be provided as the first argument. 

781 collections : Any, optional 

782 Collections to be searched, overriding ``self.collections``. 

783 Can be any of the types supported by the ``collections`` argument 

784 to butler construction. 

785 predict : `bool`, optional 

786 If `True`, return a newly created `DatasetRef` with a unique 

787 dataset ID if finding a reference in the `Registry` fails. 

788 Defaults to `False`. 

789 run : `str`, optional 

790 Run collection name to use for creating `DatasetRef` for predicted 

791 datasets. Only used if ``predict`` is `True`. 

792 datastore_records : `bool`, optional 

793 If `True` add datastore records to returned `DatasetRef`. 

794 **kwargs 

795 Additional keyword arguments used to augment or construct a 

796 `DataId`. See `DataId` parameters. 

797 

798 Returns 

799 ------- 

800 ref : `DatasetRef` 

801 A reference to the dataset identified by the given arguments. 

802 This can be the same dataset reference as given if it was 

803 resolved. 

804 

805 Raises 

806 ------ 

807 LookupError 

808 Raised if no matching dataset exists in the `Registry` (and 

809 ``predict`` is `False`). 

810 ValueError 

811 Raised if a resolved `DatasetRef` was passed as an input, but it 

812 differs from the one found in the registry. 

813 TypeError 

814 Raised if no collections were provided. 

815 """ 

816 datasetType, dataId = self._standardizeArgs(datasetRefOrType, dataId, for_put=False, **kwargs) 

817 if isinstance(datasetRefOrType, DatasetRef): 

818 if collections is not None: 

819 warnings.warn("Collections should not be specified with DatasetRef", stacklevel=3) 

820 # May need to retrieve datastore records if requested. 

821 if datastore_records and datasetRefOrType._datastore_records is None: 

822 datasetRefOrType = self._registry.get_datastore_records(datasetRefOrType) 

823 return datasetRefOrType 

824 timespan: Timespan | None = None 

825 

826 dataId, kwargs = self._rewrite_data_id(dataId, datasetType, **kwargs) 

827 

828 if datasetType.isCalibration(): 

829 # Because this is a calibration dataset, first try to make a 

830 # standardize the data ID without restricting the dimensions to 

831 # those of the dataset type requested, because there may be extra 

832 # dimensions that provide temporal information for a validity-range 

833 # lookup. 

834 dataId = DataCoordinate.standardize( 

835 dataId, universe=self.dimensions, defaults=self._registry.defaults.dataId, **kwargs 

836 ) 

837 if dataId.graph.temporal: 

838 dataId = self._registry.expandDataId(dataId) 

839 timespan = dataId.timespan 

840 else: 

841 # Standardize the data ID to just the dimensions of the dataset 

842 # type instead of letting registry.findDataset do it, so we get the 

843 # result even if no dataset is found. 

844 dataId = DataCoordinate.standardize( 

845 dataId, graph=datasetType.dimensions, defaults=self._registry.defaults.dataId, **kwargs 

846 ) 

847 # Always lookup the DatasetRef, even if one is given, to ensure it is 

848 # present in the current collection. 

849 ref = self._registry.findDataset( 

850 datasetType, 

851 dataId, 

852 collections=collections, 

853 timespan=timespan, 

854 datastore_records=datastore_records, 

855 ) 

856 if ref is None: 

857 if predict: 

858 if run is None: 

859 run = self.run 

860 if run is None: 

861 raise TypeError("Cannot predict dataset ID/location with run=None.") 

862 return DatasetRef(datasetType, dataId, run=run) 

863 else: 

864 if collections is None: 

865 collections = self._registry.defaults.collections 

866 raise LookupError( 

867 f"Dataset {datasetType.name} with data ID {dataId} " 

868 f"could not be found in collections {collections}." 

869 ) 

870 if datasetType != ref.datasetType: 

871 # If they differ it is because the user explicitly specified 

872 # a compatible dataset type to this call rather than using the 

873 # registry definition. The DatasetRef must therefore be recreated 

874 # using the user definition such that the expected type is 

875 # returned. 

876 ref = DatasetRef( 

877 datasetType, ref.dataId, run=ref.run, id=ref.id, datastore_records=ref._datastore_records 

878 ) 

879 

880 return ref 

881 

882 # TODO: remove on DM-40067. 

883 @transactional 

884 @deprecated( 

885 reason="Butler.put() now behaves like Butler.putDirect() when given a DatasetRef." 

886 " Please use Butler.put(). Be aware that you may need to adjust your usage if you" 

887 " were relying on the run parameter to determine the run." 

888 " Will be removed after v26.0.", 

889 version="v26.0", 

890 category=FutureWarning, 

891 ) 

892 def putDirect(self, obj: Any, ref: DatasetRef, /) -> DatasetRef: 

893 # Docstring inherited. 

894 return self.put(obj, ref) 

895 

896 @transactional 

897 def put( 

898 self, 

899 obj: Any, 

900 datasetRefOrType: DatasetRef | DatasetType | str, 

901 /, 

902 dataId: DataId | None = None, 

903 *, 

904 run: str | None = None, 

905 **kwargs: Any, 

906 ) -> DatasetRef: 

907 """Store and register a dataset. 

908 

909 Parameters 

910 ---------- 

911 obj : `object` 

912 The dataset. 

913 datasetRefOrType : `DatasetRef`, `DatasetType`, or `str` 

914 When `DatasetRef` is provided, ``dataId`` should be `None`. 

915 Otherwise the `DatasetType` or name thereof. If a fully resolved 

916 `DatasetRef` is given the run and ID are used directly. 

917 dataId : `dict` or `DataCoordinate` 

918 A `dict` of `Dimension` link name, value pairs that label the 

919 `DatasetRef` within a Collection. When `None`, a `DatasetRef` 

920 should be provided as the second argument. 

921 run : `str`, optional 

922 The name of the run the dataset should be added to, overriding 

923 ``self.run``. Not used if a resolved `DatasetRef` is provided. 

924 **kwargs 

925 Additional keyword arguments used to augment or construct a 

926 `DataCoordinate`. See `DataCoordinate.standardize` 

927 parameters. Not used if a resolve `DatasetRef` is provided. 

928 

929 Returns 

930 ------- 

931 ref : `DatasetRef` 

932 A reference to the stored dataset, updated with the correct id if 

933 given. 

934 

935 Raises 

936 ------ 

937 TypeError 

938 Raised if the butler is read-only or if no run has been provided. 

939 """ 

940 if isinstance(datasetRefOrType, DatasetRef): 

941 # This is a direct put of predefined DatasetRef. 

942 _LOG.debug("Butler put direct: %s", datasetRefOrType) 

943 if run is not None: 

944 warnings.warn("Run collection is not used for DatasetRef", stacklevel=3) 

945 # If registry already has a dataset with the same dataset ID, 

946 # dataset type and DataId, then _importDatasets will do nothing and 

947 # just return an original ref. We have to raise in this case, there 

948 # is a datastore check below for that. 

949 self._registry._importDatasets([datasetRefOrType], expand=True) 

950 # Before trying to write to the datastore check that it does not 

951 # know this dataset. This is prone to races, of course. 

952 if self._datastore.knows(datasetRefOrType): 

953 raise ConflictingDefinitionError(f"Datastore already contains dataset: {datasetRefOrType}") 

954 # Try to write dataset to the datastore, if it fails due to a race 

955 # with another write, the content of stored data may be 

956 # unpredictable. 

957 try: 

958 self._datastore.put(obj, datasetRefOrType) 

959 except IntegrityError as e: 

960 raise ConflictingDefinitionError(f"Datastore already contains dataset: {e}") from e 

961 return datasetRefOrType 

962 

963 _LOG.debug("Butler put: %s, dataId=%s, run=%s", datasetRefOrType, dataId, run) 

964 if not self.isWriteable(): 

965 raise TypeError("Butler is read-only.") 

966 datasetType, dataId = self._standardizeArgs(datasetRefOrType, dataId, **kwargs) 

967 

968 # Handle dimension records in dataId 

969 dataId, kwargs = self._rewrite_data_id(dataId, datasetType, **kwargs) 

970 

971 # Add Registry Dataset entry. 

972 dataId = self._registry.expandDataId(dataId, graph=datasetType.dimensions, **kwargs) 

973 (ref,) = self._registry.insertDatasets(datasetType, run=run, dataIds=[dataId]) 

974 self._datastore.put(obj, ref) 

975 

976 return ref 

977 

978 # TODO: remove on DM-40067. 

979 @deprecated( 

980 reason="Butler.get() now behaves like Butler.getDirect() when given a DatasetRef." 

981 " Please use Butler.get(). Will be removed after v26.0.", 

982 version="v26.0", 

983 category=FutureWarning, 

984 ) 

985 def getDirect( 

986 self, 

987 ref: DatasetRef, 

988 *, 

989 parameters: dict[str, Any] | None = None, 

990 storageClass: StorageClass | str | None = None, 

991 ) -> Any: 

992 """Retrieve a stored dataset. 

993 

994 Parameters 

995 ---------- 

996 ref : `DatasetRef` 

997 Resolved reference to an already stored dataset. 

998 parameters : `dict` 

999 Additional StorageClass-defined options to control reading, 

1000 typically used to efficiently read only a subset of the dataset. 

1001 storageClass : `StorageClass` or `str`, optional 

1002 The storage class to be used to override the Python type 

1003 returned by this method. By default the returned type matches 

1004 the dataset type definition for this dataset. Specifying a 

1005 read `StorageClass` can force a different type to be returned. 

1006 This type must be compatible with the original type. 

1007 

1008 Returns 

1009 ------- 

1010 obj : `object` 

1011 The dataset. 

1012 """ 

1013 return self._datastore.get(ref, parameters=parameters, storageClass=storageClass) 

1014 

1015 # TODO: remove on DM-40067. 

1016 @deprecated( 

1017 reason="Butler.getDeferred() now behaves like getDirectDeferred() when given a DatasetRef. " 

1018 "Please use Butler.getDeferred(). Will be removed after v26.0.", 

1019 version="v26.0", 

1020 category=FutureWarning, 

1021 ) 

1022 def getDirectDeferred( 

1023 self, 

1024 ref: DatasetRef, 

1025 *, 

1026 parameters: dict[str, Any] | None = None, 

1027 storageClass: str | StorageClass | None = None, 

1028 ) -> DeferredDatasetHandle: 

1029 """Create a `DeferredDatasetHandle` which can later retrieve a dataset, 

1030 from a resolved `DatasetRef`. 

1031 

1032 Parameters 

1033 ---------- 

1034 ref : `DatasetRef` 

1035 Resolved reference to an already stored dataset. 

1036 parameters : `dict` 

1037 Additional StorageClass-defined options to control reading, 

1038 typically used to efficiently read only a subset of the dataset. 

1039 storageClass : `StorageClass` or `str`, optional 

1040 The storage class to be used to override the Python type 

1041 returned by this method. By default the returned type matches 

1042 the dataset type definition for this dataset. Specifying a 

1043 read `StorageClass` can force a different type to be returned. 

1044 This type must be compatible with the original type. 

1045 

1046 Returns 

1047 ------- 

1048 obj : `DeferredDatasetHandle` 

1049 A handle which can be used to retrieve a dataset at a later time. 

1050 

1051 Raises 

1052 ------ 

1053 LookupError 

1054 Raised if no matching dataset exists in the `Registry`. 

1055 """ 

1056 # Check that dataset is known to the datastore. 

1057 if not self._datastore.knows(ref): 

1058 raise LookupError(f"Dataset reference {ref} is not known to datastore.") 

1059 return DeferredDatasetHandle(butler=self, ref=ref, parameters=parameters, storageClass=storageClass) 

1060 

1061 def getDeferred( 

1062 self, 

1063 datasetRefOrType: DatasetRef | DatasetType | str, 

1064 /, 

1065 dataId: DataId | None = None, 

1066 *, 

1067 parameters: dict | None = None, 

1068 collections: Any = None, 

1069 storageClass: str | StorageClass | None = None, 

1070 **kwargs: Any, 

1071 ) -> DeferredDatasetHandle: 

1072 """Create a `DeferredDatasetHandle` which can later retrieve a dataset, 

1073 after an immediate registry lookup. 

1074 

1075 Parameters 

1076 ---------- 

1077 datasetRefOrType : `DatasetRef`, `DatasetType`, or `str` 

1078 When `DatasetRef` the `dataId` should be `None`. 

1079 Otherwise the `DatasetType` or name thereof. 

1080 dataId : `dict` or `DataCoordinate`, optional 

1081 A `dict` of `Dimension` link name, value pairs that label the 

1082 `DatasetRef` within a Collection. When `None`, a `DatasetRef` 

1083 should be provided as the first argument. 

1084 parameters : `dict` 

1085 Additional StorageClass-defined options to control reading, 

1086 typically used to efficiently read only a subset of the dataset. 

1087 collections : Any, optional 

1088 Collections to be searched, overriding ``self.collections``. 

1089 Can be any of the types supported by the ``collections`` argument 

1090 to butler construction. 

1091 storageClass : `StorageClass` or `str`, optional 

1092 The storage class to be used to override the Python type 

1093 returned by this method. By default the returned type matches 

1094 the dataset type definition for this dataset. Specifying a 

1095 read `StorageClass` can force a different type to be returned. 

1096 This type must be compatible with the original type. 

1097 **kwargs 

1098 Additional keyword arguments used to augment or construct a 

1099 `DataId`. See `DataId` parameters. 

1100 

1101 Returns 

1102 ------- 

1103 obj : `DeferredDatasetHandle` 

1104 A handle which can be used to retrieve a dataset at a later time. 

1105 

1106 Raises 

1107 ------ 

1108 LookupError 

1109 Raised if no matching dataset exists in the `Registry` or 

1110 datastore. 

1111 ValueError 

1112 Raised if a resolved `DatasetRef` was passed as an input, but it 

1113 differs from the one found in the registry. 

1114 TypeError 

1115 Raised if no collections were provided. 

1116 """ 

1117 if isinstance(datasetRefOrType, DatasetRef): 

1118 # Do the quick check first and if that fails, check for artifact 

1119 # existence. This is necessary for datastores that are configured 

1120 # in trust mode where there won't be a record but there will be 

1121 # a file. 

1122 if self._datastore.knows(datasetRefOrType) or self._datastore.exists(datasetRefOrType): 

1123 ref = datasetRefOrType 

1124 else: 

1125 raise LookupError(f"Dataset reference {datasetRefOrType} does not exist.") 

1126 else: 

1127 ref = self._findDatasetRef(datasetRefOrType, dataId, collections=collections, **kwargs) 

1128 return DeferredDatasetHandle(butler=self, ref=ref, parameters=parameters, storageClass=storageClass) 

1129 

1130 def get( 

1131 self, 

1132 datasetRefOrType: DatasetRef | DatasetType | str, 

1133 /, 

1134 dataId: DataId | None = None, 

1135 *, 

1136 parameters: dict[str, Any] | None = None, 

1137 collections: Any = None, 

1138 storageClass: StorageClass | str | None = None, 

1139 **kwargs: Any, 

1140 ) -> Any: 

1141 """Retrieve a stored dataset. 

1142 

1143 Parameters 

1144 ---------- 

1145 datasetRefOrType : `DatasetRef`, `DatasetType`, or `str` 

1146 When `DatasetRef` the `dataId` should be `None`. 

1147 Otherwise the `DatasetType` or name thereof. 

1148 If a resolved `DatasetRef`, the associated dataset 

1149 is returned directly without additional querying. 

1150 dataId : `dict` or `DataCoordinate` 

1151 A `dict` of `Dimension` link name, value pairs that label the 

1152 `DatasetRef` within a Collection. When `None`, a `DatasetRef` 

1153 should be provided as the first argument. 

1154 parameters : `dict` 

1155 Additional StorageClass-defined options to control reading, 

1156 typically used to efficiently read only a subset of the dataset. 

1157 collections : Any, optional 

1158 Collections to be searched, overriding ``self.collections``. 

1159 Can be any of the types supported by the ``collections`` argument 

1160 to butler construction. 

1161 storageClass : `StorageClass` or `str`, optional 

1162 The storage class to be used to override the Python type 

1163 returned by this method. By default the returned type matches 

1164 the dataset type definition for this dataset. Specifying a 

1165 read `StorageClass` can force a different type to be returned. 

1166 This type must be compatible with the original type. 

1167 **kwargs 

1168 Additional keyword arguments used to augment or construct a 

1169 `DataCoordinate`. See `DataCoordinate.standardize` 

1170 parameters. 

1171 

1172 Returns 

1173 ------- 

1174 obj : `object` 

1175 The dataset. 

1176 

1177 Raises 

1178 ------ 

1179 LookupError 

1180 Raised if no matching dataset exists in the `Registry`. 

1181 TypeError 

1182 Raised if no collections were provided. 

1183 

1184 Notes 

1185 ----- 

1186 When looking up datasets in a `~CollectionType.CALIBRATION` collection, 

1187 this method requires that the given data ID include temporal dimensions 

1188 beyond the dimensions of the dataset type itself, in order to find the 

1189 dataset with the appropriate validity range. For example, a "bias" 

1190 dataset with native dimensions ``{instrument, detector}`` could be 

1191 fetched with a ``{instrument, detector, exposure}`` data ID, because 

1192 ``exposure`` is a temporal dimension. 

1193 """ 

1194 _LOG.debug("Butler get: %s, dataId=%s, parameters=%s", datasetRefOrType, dataId, parameters) 

1195 ref = self._findDatasetRef( 

1196 datasetRefOrType, dataId, collections=collections, datastore_records=True, **kwargs 

1197 ) 

1198 return self._datastore.get(ref, parameters=parameters, storageClass=storageClass) 

1199 

1200 def getURIs( 

1201 self, 

1202 datasetRefOrType: DatasetRef | DatasetType | str, 

1203 /, 

1204 dataId: DataId | None = None, 

1205 *, 

1206 predict: bool = False, 

1207 collections: Any = None, 

1208 run: str | None = None, 

1209 **kwargs: Any, 

1210 ) -> DatasetRefURIs: 

1211 """Return the URIs associated with the dataset. 

1212 

1213 Parameters 

1214 ---------- 

1215 datasetRefOrType : `DatasetRef`, `DatasetType`, or `str` 

1216 When `DatasetRef` the `dataId` should be `None`. 

1217 Otherwise the `DatasetType` or name thereof. 

1218 dataId : `dict` or `DataCoordinate` 

1219 A `dict` of `Dimension` link name, value pairs that label the 

1220 `DatasetRef` within a Collection. When `None`, a `DatasetRef` 

1221 should be provided as the first argument. 

1222 predict : `bool` 

1223 If `True`, allow URIs to be returned of datasets that have not 

1224 been written. 

1225 collections : Any, optional 

1226 Collections to be searched, overriding ``self.collections``. 

1227 Can be any of the types supported by the ``collections`` argument 

1228 to butler construction. 

1229 run : `str`, optional 

1230 Run to use for predictions, overriding ``self.run``. 

1231 **kwargs 

1232 Additional keyword arguments used to augment or construct a 

1233 `DataCoordinate`. See `DataCoordinate.standardize` 

1234 parameters. 

1235 

1236 Returns 

1237 ------- 

1238 uris : `DatasetRefURIs` 

1239 The URI to the primary artifact associated with this dataset (if 

1240 the dataset was disassembled within the datastore this may be 

1241 `None`), and the URIs to any components associated with the dataset 

1242 artifact. (can be empty if there are no components). 

1243 """ 

1244 ref = self._findDatasetRef( 

1245 datasetRefOrType, dataId, predict=predict, run=run, collections=collections, **kwargs 

1246 ) 

1247 return self._datastore.getURIs(ref, predict) 

1248 

1249 def getURI( 

1250 self, 

1251 datasetRefOrType: DatasetRef | DatasetType | str, 

1252 /, 

1253 dataId: DataId | None = None, 

1254 *, 

1255 predict: bool = False, 

1256 collections: Any = None, 

1257 run: str | None = None, 

1258 **kwargs: Any, 

1259 ) -> ResourcePath: 

1260 """Return the URI to the Dataset. 

1261 

1262 Parameters 

1263 ---------- 

1264 datasetRefOrType : `DatasetRef`, `DatasetType`, or `str` 

1265 When `DatasetRef` the `dataId` should be `None`. 

1266 Otherwise the `DatasetType` or name thereof. 

1267 dataId : `dict` or `DataCoordinate` 

1268 A `dict` of `Dimension` link name, value pairs that label the 

1269 `DatasetRef` within a Collection. When `None`, a `DatasetRef` 

1270 should be provided as the first argument. 

1271 predict : `bool` 

1272 If `True`, allow URIs to be returned of datasets that have not 

1273 been written. 

1274 collections : Any, optional 

1275 Collections to be searched, overriding ``self.collections``. 

1276 Can be any of the types supported by the ``collections`` argument 

1277 to butler construction. 

1278 run : `str`, optional 

1279 Run to use for predictions, overriding ``self.run``. 

1280 **kwargs 

1281 Additional keyword arguments used to augment or construct a 

1282 `DataCoordinate`. See `DataCoordinate.standardize` 

1283 parameters. 

1284 

1285 Returns 

1286 ------- 

1287 uri : `lsst.resources.ResourcePath` 

1288 URI pointing to the Dataset within the datastore. If the 

1289 Dataset does not exist in the datastore, and if ``predict`` is 

1290 `True`, the URI will be a prediction and will include a URI 

1291 fragment "#predicted". 

1292 If the datastore does not have entities that relate well 

1293 to the concept of a URI the returned URI string will be 

1294 descriptive. The returned URI is not guaranteed to be obtainable. 

1295 

1296 Raises 

1297 ------ 

1298 LookupError 

1299 A URI has been requested for a dataset that does not exist and 

1300 guessing is not allowed. 

1301 ValueError 

1302 Raised if a resolved `DatasetRef` was passed as an input, but it 

1303 differs from the one found in the registry. 

1304 TypeError 

1305 Raised if no collections were provided. 

1306 RuntimeError 

1307 Raised if a URI is requested for a dataset that consists of 

1308 multiple artifacts. 

1309 """ 

1310 primary, components = self.getURIs( 

1311 datasetRefOrType, dataId=dataId, predict=predict, collections=collections, run=run, **kwargs 

1312 ) 

1313 

1314 if primary is None or components: 

1315 raise RuntimeError( 

1316 f"Dataset ({datasetRefOrType}) includes distinct URIs for components. " 

1317 "Use Butler.getURIs() instead." 

1318 ) 

1319 return primary 

1320 

1321 def retrieveArtifacts( 

1322 self, 

1323 refs: Iterable[DatasetRef], 

1324 destination: ResourcePathExpression, 

1325 transfer: str = "auto", 

1326 preserve_path: bool = True, 

1327 overwrite: bool = False, 

1328 ) -> list[ResourcePath]: 

1329 # Docstring inherited. 

1330 return self._datastore.retrieveArtifacts( 

1331 refs, 

1332 ResourcePath(destination), 

1333 transfer=transfer, 

1334 preserve_path=preserve_path, 

1335 overwrite=overwrite, 

1336 ) 

1337 

1338 def exists( 

1339 self, 

1340 dataset_ref_or_type: DatasetRef | DatasetType | str, 

1341 /, 

1342 data_id: DataId | None = None, 

1343 *, 

1344 full_check: bool = True, 

1345 collections: Any = None, 

1346 **kwargs: Any, 

1347 ) -> DatasetExistence: 

1348 # Docstring inherited. 

1349 existence = DatasetExistence.UNRECOGNIZED 

1350 

1351 if isinstance(dataset_ref_or_type, DatasetRef): 

1352 if collections is not None: 

1353 warnings.warn("Collections should not be specified with DatasetRef", stacklevel=2) 

1354 if data_id is not None: 

1355 warnings.warn("A DataID should not be specified with DatasetRef", stacklevel=2) 

1356 ref = dataset_ref_or_type 

1357 registry_ref = self._registry.getDataset(dataset_ref_or_type.id) 

1358 if registry_ref is not None: 

1359 existence |= DatasetExistence.RECORDED 

1360 

1361 if dataset_ref_or_type != registry_ref: 

1362 # This could mean that storage classes differ, so we should 

1363 # check for that but use the registry ref for the rest of 

1364 # the method. 

1365 if registry_ref.is_compatible_with(dataset_ref_or_type): 

1366 # Use the registry version from now on. 

1367 ref = registry_ref 

1368 else: 

1369 raise ValueError( 

1370 f"The ref given to exists() ({ref}) has the same dataset ID as one " 

1371 f"in registry but has different incompatible values ({registry_ref})." 

1372 ) 

1373 else: 

1374 try: 

1375 ref = self._findDatasetRef(dataset_ref_or_type, data_id, collections=collections, **kwargs) 

1376 except (LookupError, TypeError, NoDefaultCollectionError): 

1377 return existence 

1378 existence |= DatasetExistence.RECORDED 

1379 

1380 if self._datastore.knows(ref): 

1381 existence |= DatasetExistence.DATASTORE 

1382 

1383 if full_check: 

1384 if self._datastore.exists(ref): 

1385 existence |= DatasetExistence._ARTIFACT 

1386 elif existence.value != DatasetExistence.UNRECOGNIZED.value: 

1387 # Do not add this flag if we have no other idea about a dataset. 

1388 existence |= DatasetExistence(DatasetExistence._ASSUMED) 

1389 

1390 return existence 

1391 

1392 def _exists_many( 

1393 self, 

1394 refs: Iterable[DatasetRef], 

1395 /, 

1396 *, 

1397 full_check: bool = True, 

1398 ) -> dict[DatasetRef, DatasetExistence]: 

1399 # Docstring inherited. 

1400 existence = {ref: DatasetExistence.UNRECOGNIZED for ref in refs} 

1401 

1402 # Registry does not have a bulk API to check for a ref. 

1403 for ref in refs: 

1404 registry_ref = self._registry.getDataset(ref.id) 

1405 if registry_ref is not None: 

1406 # It is possible, albeit unlikely, that the given ref does 

1407 # not match the one in registry even though the UUID matches. 

1408 # When checking a single ref we raise, but it's impolite to 

1409 # do that when potentially hundreds of refs are being checked. 

1410 # We could change the API to only accept UUIDs and that would 

1411 # remove the ability to even check and remove the worry 

1412 # about differing storage classes. Given the ongoing discussion 

1413 # on refs vs UUIDs and whether to raise or have a new 

1414 # private flag, treat this as a private API for now. 

1415 existence[ref] |= DatasetExistence.RECORDED 

1416 

1417 # Ask datastore if it knows about these refs. 

1418 knows = self._datastore.knows_these(refs) 

1419 for ref, known in knows.items(): 

1420 if known: 

1421 existence[ref] |= DatasetExistence.DATASTORE 

1422 

1423 if full_check: 

1424 mexists = self._datastore.mexists(refs) 

1425 for ref, exists in mexists.items(): 

1426 if exists: 

1427 existence[ref] |= DatasetExistence._ARTIFACT 

1428 else: 

1429 # Do not set this flag if nothing is known about the dataset. 

1430 for ref in existence: 

1431 if existence[ref] != DatasetExistence.UNRECOGNIZED: 

1432 existence[ref] |= DatasetExistence._ASSUMED 

1433 

1434 return existence 

1435 

1436 # TODO: remove on DM-40079. 

1437 @deprecated( 

1438 reason="Butler.datasetExists() has been replaced by Butler.exists(). Will be removed after v26.0.", 

1439 version="v26.0", 

1440 category=FutureWarning, 

1441 ) 

1442 def datasetExists( 

1443 self, 

1444 datasetRefOrType: DatasetRef | DatasetType | str, 

1445 dataId: DataId | None = None, 

1446 *, 

1447 collections: Any = None, 

1448 **kwargs: Any, 

1449 ) -> bool: 

1450 """Return True if the Dataset is actually present in the Datastore. 

1451 

1452 Parameters 

1453 ---------- 

1454 datasetRefOrType : `DatasetRef`, `DatasetType`, or `str` 

1455 When `DatasetRef` the `dataId` should be `None`. 

1456 Otherwise the `DatasetType` or name thereof. 

1457 dataId : `dict` or `DataCoordinate` 

1458 A `dict` of `Dimension` link name, value pairs that label the 

1459 `DatasetRef` within a Collection. When `None`, a `DatasetRef` 

1460 should be provided as the first argument. 

1461 collections : Any, optional 

1462 Collections to be searched, overriding ``self.collections``. 

1463 Can be any of the types supported by the ``collections`` argument 

1464 to butler construction. 

1465 **kwargs 

1466 Additional keyword arguments used to augment or construct a 

1467 `DataCoordinate`. See `DataCoordinate.standardize` 

1468 parameters. 

1469 

1470 Raises 

1471 ------ 

1472 LookupError 

1473 Raised if the dataset is not even present in the Registry. 

1474 ValueError 

1475 Raised if a resolved `DatasetRef` was passed as an input, but it 

1476 differs from the one found in the registry. 

1477 NoDefaultCollectionError 

1478 Raised if no collections were provided. 

1479 """ 

1480 # A resolved ref may be given that is not known to this butler. 

1481 if isinstance(datasetRefOrType, DatasetRef): 

1482 ref = self._registry.getDataset(datasetRefOrType.id) 

1483 if ref is None: 

1484 raise LookupError( 

1485 f"Resolved DatasetRef with id {datasetRefOrType.id} is not known to registry." 

1486 ) 

1487 else: 

1488 ref = self._findDatasetRef(datasetRefOrType, dataId, collections=collections, **kwargs) 

1489 return self._datastore.exists(ref) 

1490 

1491 def removeRuns(self, names: Iterable[str], unstore: bool = True) -> None: 

1492 # Docstring inherited. 

1493 if not self.isWriteable(): 

1494 raise TypeError("Butler is read-only.") 

1495 names = list(names) 

1496 refs: list[DatasetRef] = [] 

1497 for name in names: 

1498 collectionType = self._registry.getCollectionType(name) 

1499 if collectionType is not CollectionType.RUN: 

1500 raise TypeError(f"The collection type of '{name}' is {collectionType.name}, not RUN.") 

1501 refs.extend(self._registry.queryDatasets(..., collections=name, findFirst=True)) 

1502 with self._datastore.transaction(), self._registry.transaction(): 

1503 if unstore: 

1504 self._datastore.trash(refs) 

1505 else: 

1506 self._datastore.forget(refs) 

1507 for name in names: 

1508 self._registry.removeCollection(name) 

1509 if unstore: 

1510 # Point of no return for removing artifacts 

1511 self._datastore.emptyTrash() 

1512 

1513 def pruneDatasets( 

1514 self, 

1515 refs: Iterable[DatasetRef], 

1516 *, 

1517 disassociate: bool = True, 

1518 unstore: bool = False, 

1519 tags: Iterable[str] = (), 

1520 purge: bool = False, 

1521 ) -> None: 

1522 # docstring inherited from LimitedButler 

1523 

1524 if not self.isWriteable(): 

1525 raise TypeError("Butler is read-only.") 

1526 if purge: 

1527 if not disassociate: 

1528 raise TypeError("Cannot pass purge=True without disassociate=True.") 

1529 if not unstore: 

1530 raise TypeError("Cannot pass purge=True without unstore=True.") 

1531 elif disassociate: 

1532 tags = tuple(tags) 

1533 if not tags: 

1534 raise TypeError("No tags provided but disassociate=True.") 

1535 for tag in tags: 

1536 collectionType = self._registry.getCollectionType(tag) 

1537 if collectionType is not CollectionType.TAGGED: 

1538 raise TypeError( 

1539 f"Cannot disassociate from collection '{tag}' " 

1540 f"of non-TAGGED type {collectionType.name}." 

1541 ) 

1542 # Transform possibly-single-pass iterable into something we can iterate 

1543 # over multiple times. 

1544 refs = list(refs) 

1545 # Pruning a component of a DatasetRef makes no sense since registry 

1546 # doesn't know about components and datastore might not store 

1547 # components in a separate file 

1548 for ref in refs: 

1549 if ref.datasetType.component(): 

1550 raise ValueError(f"Can not prune a component of a dataset (ref={ref})") 

1551 # We don't need an unreliable Datastore transaction for this, because 

1552 # we've been extra careful to ensure that Datastore.trash only involves 

1553 # mutating the Registry (it can _look_ at Datastore-specific things, 

1554 # but shouldn't change them), and hence all operations here are 

1555 # Registry operations. 

1556 with self._datastore.transaction(), self._registry.transaction(): 

1557 if unstore: 

1558 self._datastore.trash(refs) 

1559 if purge: 

1560 self._registry.removeDatasets(refs) 

1561 elif disassociate: 

1562 assert tags, "Guaranteed by earlier logic in this function." 

1563 for tag in tags: 

1564 self._registry.disassociate(tag, refs) 

1565 # We've exited the Registry transaction, and apparently committed. 

1566 # (if there was an exception, everything rolled back, and it's as if 

1567 # nothing happened - and we never get here). 

1568 # Datastore artifacts are not yet gone, but they're clearly marked 

1569 # as trash, so if we fail to delete now because of (e.g.) filesystem 

1570 # problems we can try again later, and if manual administrative 

1571 # intervention is required, it's pretty clear what that should entail: 

1572 # deleting everything on disk and in private Datastore tables that is 

1573 # in the dataset_location_trash table. 

1574 if unstore: 

1575 # Point of no return for removing artifacts 

1576 self._datastore.emptyTrash() 

1577 

1578 @transactional 

1579 def ingest( 

1580 self, 

1581 *datasets: FileDataset, 

1582 transfer: str | None = "auto", 

1583 run: str | None = None, 

1584 idGenerationMode: DatasetIdGenEnum | None = None, 

1585 record_validation_info: bool = True, 

1586 ) -> None: 

1587 # Docstring inherited. 

1588 if not self.isWriteable(): 

1589 raise TypeError("Butler is read-only.") 

1590 

1591 _LOG.verbose("Ingesting %d file dataset%s.", len(datasets), "" if len(datasets) == 1 else "s") 

1592 if not datasets: 

1593 return 

1594 

1595 if idGenerationMode is not None: 

1596 warnings.warn( 

1597 "The idGenerationMode parameter is no longer used and is ignored. " 

1598 " Will be removed after v26.0", 

1599 FutureWarning, 

1600 stacklevel=2, 

1601 ) 

1602 

1603 progress = Progress("lsst.daf.butler.Butler.ingest", level=logging.DEBUG) 

1604 

1605 # We need to reorganize all the inputs so that they are grouped 

1606 # by dataset type and run. Multiple refs in a single FileDataset 

1607 # are required to share the run and dataset type. 

1608 GroupedData = MutableMapping[tuple[DatasetType, str], list[FileDataset]] 

1609 groupedData: GroupedData = defaultdict(list) 

1610 

1611 # Track DataIDs that are being ingested so we can spot issues early 

1612 # with duplication. Retain previous FileDataset so we can report it. 

1613 groupedDataIds: MutableMapping[ 

1614 tuple[DatasetType, str], dict[DataCoordinate, FileDataset] 

1615 ] = defaultdict(dict) 

1616 

1617 used_run = False 

1618 

1619 # And the nested loop that populates it: 

1620 for dataset in progress.wrap(datasets, desc="Grouping by dataset type"): 

1621 # Somewhere to store pre-existing refs if we have an 

1622 # execution butler. 

1623 existingRefs: list[DatasetRef] = [] 

1624 

1625 for ref in dataset.refs: 

1626 assert ref.run is not None # For mypy 

1627 group_key = (ref.datasetType, ref.run) 

1628 

1629 if ref.dataId in groupedDataIds[group_key]: 

1630 raise ConflictingDefinitionError( 

1631 f"Ingest conflict. Dataset {dataset.path} has same" 

1632 " DataId as other ingest dataset" 

1633 f" {groupedDataIds[group_key][ref.dataId].path} " 

1634 f" ({ref.dataId})" 

1635 ) 

1636 

1637 groupedDataIds[group_key][ref.dataId] = dataset 

1638 

1639 if existingRefs: 

1640 if len(dataset.refs) != len(existingRefs): 

1641 # Keeping track of partially pre-existing datasets is hard 

1642 # and should generally never happen. For now don't allow 

1643 # it. 

1644 raise ConflictingDefinitionError( 

1645 f"For dataset {dataset.path} some dataIds already exist" 

1646 " in registry but others do not. This is not supported." 

1647 ) 

1648 

1649 # Store expanded form in the original FileDataset. 

1650 dataset.refs = existingRefs 

1651 else: 

1652 groupedData[group_key].append(dataset) 

1653 

1654 if not used_run and run is not None: 

1655 warnings.warn( 

1656 "All DatasetRefs to be ingested had resolved dataset IDs. The value given to the " 

1657 f"'run' parameter ({run!r}) was not used and the parameter will be removed in the future.", 

1658 category=FutureWarning, 

1659 stacklevel=3, # Take into account the @transactional decorator. 

1660 ) 

1661 

1662 # Now we can bulk-insert into Registry for each DatasetType. 

1663 for (datasetType, this_run), grouped_datasets in progress.iter_item_chunks( 

1664 groupedData.items(), desc="Bulk-inserting datasets by type" 

1665 ): 

1666 refs_to_import = [] 

1667 for dataset in grouped_datasets: 

1668 refs_to_import.extend(dataset.refs) 

1669 

1670 n_refs = len(refs_to_import) 

1671 _LOG.verbose( 

1672 "Importing %d ref%s of dataset type %r into run %r", 

1673 n_refs, 

1674 "" if n_refs == 1 else "s", 

1675 datasetType.name, 

1676 this_run, 

1677 ) 

1678 

1679 # Import the refs and expand the DataCoordinates since we can't 

1680 # guarantee that they are expanded and Datastore will need 

1681 # the records. 

1682 imported_refs = self._registry._importDatasets(refs_to_import, expand=True) 

1683 assert set(imported_refs) == set(refs_to_import) 

1684 

1685 # Replace all the refs in the FileDataset with expanded versions. 

1686 # Pull them off in the order we put them on the list. 

1687 for dataset in grouped_datasets: 

1688 n_dataset_refs = len(dataset.refs) 

1689 dataset.refs = imported_refs[:n_dataset_refs] 

1690 del imported_refs[:n_dataset_refs] 

1691 

1692 # Bulk-insert everything into Datastore. 

1693 # We do not know if any of the registry entries already existed 

1694 # (_importDatasets only complains if they exist but differ) so 

1695 # we have to catch IntegrityError explicitly. 

1696 try: 

1697 self._datastore.ingest( 

1698 *datasets, transfer=transfer, record_validation_info=record_validation_info 

1699 ) 

1700 except IntegrityError as e: 

1701 raise ConflictingDefinitionError(f"Datastore already contains one or more datasets: {e}") from e 

1702 

1703 @contextlib.contextmanager 

1704 def export( 

1705 self, 

1706 *, 

1707 directory: str | None = None, 

1708 filename: str | None = None, 

1709 format: str | None = None, 

1710 transfer: str | None = None, 

1711 ) -> Iterator[RepoExportContext]: 

1712 # Docstring inherited. 

1713 if directory is None and transfer is not None: 

1714 raise TypeError("Cannot transfer without providing a directory.") 

1715 if transfer == "move": 

1716 raise TypeError("Transfer may not be 'move': export is read-only") 

1717 if format is None: 

1718 if filename is None: 

1719 raise TypeError("At least one of 'filename' or 'format' must be provided.") 

1720 else: 

1721 _, format = os.path.splitext(filename) 

1722 if not format: 

1723 raise ValueError("Please specify a file extension to determine export format.") 

1724 format = format[1:] # Strip leading "."" 

1725 elif filename is None: 

1726 filename = f"export.{format}" 

1727 if directory is not None: 

1728 filename = os.path.join(directory, filename) 

1729 formats = self._config["repo_transfer_formats"] 

1730 if format not in formats: 

1731 raise ValueError(f"Unknown export format {format!r}, allowed: {','.join(formats.keys())}") 

1732 BackendClass = get_class_of(formats[format, "export"]) 

1733 with open(filename, "w") as stream: 

1734 backend = BackendClass(stream, universe=self.dimensions) 

1735 try: 

1736 helper = RepoExportContext( 

1737 self._registry, self._datastore, backend=backend, directory=directory, transfer=transfer 

1738 ) 

1739 yield helper 

1740 except BaseException: 

1741 raise 

1742 else: 

1743 helper._finish() 

1744 

1745 def import_( 

1746 self, 

1747 *, 

1748 directory: ResourcePathExpression | None = None, 

1749 filename: ResourcePathExpression | TextIO | None = None, 

1750 format: str | None = None, 

1751 transfer: str | None = None, 

1752 skip_dimensions: set | None = None, 

1753 ) -> None: 

1754 # Docstring inherited. 

1755 if not self.isWriteable(): 

1756 raise TypeError("Butler is read-only.") 

1757 if format is None: 

1758 if filename is None: 

1759 raise TypeError("At least one of 'filename' or 'format' must be provided.") 

1760 else: 

1761 _, format = os.path.splitext(filename) # type: ignore 

1762 elif filename is None: 

1763 filename = ResourcePath(f"export.{format}", forceAbsolute=False) 

1764 if directory is not None: 

1765 directory = ResourcePath(directory, forceDirectory=True) 

1766 # mypy doesn't think this will work but it does in python >= 3.10. 

1767 if isinstance(filename, ResourcePathExpression): # type: ignore 

1768 filename = ResourcePath(filename, forceAbsolute=False) # type: ignore 

1769 if not filename.isabs() and directory is not None: 

1770 potential = directory.join(filename) 

1771 exists_in_cwd = filename.exists() 

1772 exists_in_dir = potential.exists() 

1773 if exists_in_cwd and exists_in_dir: 

1774 _LOG.warning( 

1775 "A relative path for filename was specified (%s) which exists relative to cwd. " 

1776 "Additionally, the file exists relative to the given search directory (%s). " 

1777 "Using the export file in the given directory.", 

1778 filename, 

1779 potential, 

1780 ) 

1781 # Given they specified an explicit directory and that 

1782 # directory has the export file in it, assume that that 

1783 # is what was meant despite the file in cwd. 

1784 filename = potential 

1785 elif exists_in_dir: 

1786 filename = potential 

1787 elif not exists_in_cwd and not exists_in_dir: 

1788 # Raise early. 

1789 raise FileNotFoundError( 

1790 f"Export file could not be found in {filename.abspath()} or {potential.abspath()}." 

1791 ) 

1792 BackendClass: type[RepoImportBackend] = get_class_of( 

1793 self._config["repo_transfer_formats"][format]["import"] 

1794 ) 

1795 

1796 def doImport(importStream: TextIO | ResourceHandleProtocol) -> None: 

1797 backend = BackendClass(importStream, self._registry) # type: ignore[call-arg] 

1798 backend.register() 

1799 with self.transaction(): 

1800 backend.load( 

1801 self._datastore, 

1802 directory=directory, 

1803 transfer=transfer, 

1804 skip_dimensions=skip_dimensions, 

1805 ) 

1806 

1807 if isinstance(filename, ResourcePath): 

1808 # We can not use open() here at the moment because of 

1809 # DM-38589 since yaml does stream.read(8192) in a loop. 

1810 stream = io.StringIO(filename.read().decode()) 

1811 doImport(stream) 

1812 else: 

1813 doImport(filename) # type: ignore 

1814 

1815 def transfer_from( 

1816 self, 

1817 source_butler: LimitedButler, 

1818 source_refs: Iterable[DatasetRef], 

1819 transfer: str = "auto", 

1820 skip_missing: bool = True, 

1821 register_dataset_types: bool = False, 

1822 transfer_dimensions: bool = False, 

1823 ) -> collections.abc.Collection[DatasetRef]: 

1824 # Docstring inherited. 

1825 if not self.isWriteable(): 

1826 raise TypeError("Butler is read-only.") 

1827 progress = Progress("lsst.daf.butler.Butler.transfer_from", level=VERBOSE) 

1828 

1829 # Will iterate through the refs multiple times so need to convert 

1830 # to a list if this isn't a collection. 

1831 if not isinstance(source_refs, collections.abc.Collection): 

1832 source_refs = list(source_refs) 

1833 

1834 original_count = len(source_refs) 

1835 _LOG.info("Transferring %d datasets into %s", original_count, str(self)) 

1836 

1837 # In some situations the datastore artifact may be missing 

1838 # and we do not want that registry entry to be imported. 

1839 # Asking datastore is not sufficient, the records may have been 

1840 # purged, we have to ask for the (predicted) URI and check 

1841 # existence explicitly. Execution butler is set up exactly like 

1842 # this with no datastore records. 

1843 artifact_existence: dict[ResourcePath, bool] = {} 

1844 if skip_missing: 

1845 dataset_existence = source_butler._datastore.mexists( 

1846 source_refs, artifact_existence=artifact_existence 

1847 ) 

1848 source_refs = [ref for ref, exists in dataset_existence.items() if exists] 

1849 filtered_count = len(source_refs) 

1850 n_missing = original_count - filtered_count 

1851 _LOG.verbose( 

1852 "%d dataset%s removed because the artifact does not exist. Now have %d.", 

1853 n_missing, 

1854 "" if n_missing == 1 else "s", 

1855 filtered_count, 

1856 ) 

1857 

1858 # Importing requires that we group the refs by dataset type and run 

1859 # before doing the import. 

1860 source_dataset_types = set() 

1861 grouped_refs = defaultdict(list) 

1862 for ref in source_refs: 

1863 grouped_refs[ref.datasetType, ref.run].append(ref) 

1864 source_dataset_types.add(ref.datasetType) 

1865 

1866 # Check to see if the dataset type in the source butler has 

1867 # the same definition in the target butler and register missing 

1868 # ones if requested. Registration must happen outside a transaction. 

1869 newly_registered_dataset_types = set() 

1870 for datasetType in source_dataset_types: 

1871 if register_dataset_types: 

1872 # Let this raise immediately if inconsistent. Continuing 

1873 # on to find additional inconsistent dataset types 

1874 # might result in additional unwanted dataset types being 

1875 # registered. 

1876 if self._registry.registerDatasetType(datasetType): 

1877 newly_registered_dataset_types.add(datasetType) 

1878 else: 

1879 # If the dataset type is missing, let it fail immediately. 

1880 target_dataset_type = self._registry.getDatasetType(datasetType.name) 

1881 if target_dataset_type != datasetType: 

1882 raise ConflictingDefinitionError( 

1883 "Source butler dataset type differs from definition" 

1884 f" in target butler: {datasetType} !=" 

1885 f" {target_dataset_type}" 

1886 ) 

1887 if newly_registered_dataset_types: 

1888 # We may have registered some even if there were inconsistencies 

1889 # but should let people know (or else remove them again). 

1890 _LOG.verbose( 

1891 "Registered the following dataset types in the target Butler: %s", 

1892 ", ".join(d.name for d in newly_registered_dataset_types), 

1893 ) 

1894 else: 

1895 _LOG.verbose("All required dataset types are known to the target Butler") 

1896 

1897 dimension_records: dict[DimensionElement, dict[DataCoordinate, DimensionRecord]] = defaultdict(dict) 

1898 if transfer_dimensions: 

1899 # Collect all the dimension records for these refs. 

1900 # All dimensions are to be copied but the list of valid dimensions 

1901 # come from this butler's universe. 

1902 elements = frozenset( 

1903 element 

1904 for element in self.dimensions.getStaticElements() 

1905 if element.hasTable() and element.viewOf is None 

1906 ) 

1907 dataIds = {ref.dataId for ref in source_refs} 

1908 # This logic comes from saveDataIds. 

1909 for dataId in dataIds: 

1910 # Need an expanded record, if not expanded that we need a full 

1911 # butler with registry (allow mocks with registry too). 

1912 if not dataId.hasRecords(): 

1913 if registry := getattr(source_butler, "registry", None): 

1914 dataId = registry.expandDataId(dataId) 

1915 else: 

1916 raise TypeError("Input butler needs to be a full butler to expand DataId.") 

1917 # If this butler doesn't know about a dimension in the source 

1918 # butler things will break later. 

1919 for record in dataId.records.values(): 

1920 if record is not None and record.definition in elements: 

1921 dimension_records[record.definition].setdefault(record.dataId, record) 

1922 

1923 handled_collections: set[str] = set() 

1924 

1925 # Do all the importing in a single transaction. 

1926 with self.transaction(): 

1927 if dimension_records: 

1928 _LOG.verbose("Ensuring that dimension records exist for transferred datasets.") 

1929 for element, r in dimension_records.items(): 

1930 records = [r[dataId] for dataId in r] 

1931 # Assume that if the record is already present that we can 

1932 # use it without having to check that the record metadata 

1933 # is consistent. 

1934 self._registry.insertDimensionData(element, *records, skip_existing=True) 

1935 

1936 n_imported = 0 

1937 for (datasetType, run), refs_to_import in progress.iter_item_chunks( 

1938 grouped_refs.items(), desc="Importing to registry by run and dataset type" 

1939 ): 

1940 if run not in handled_collections: 

1941 # May need to create output collection. If source butler 

1942 # has a registry, ask for documentation string. 

1943 run_doc = None 

1944 if registry := getattr(source_butler, "registry", None): 

1945 run_doc = registry.getCollectionDocumentation(run) 

1946 registered = self._registry.registerRun(run, doc=run_doc) 

1947 handled_collections.add(run) 

1948 if registered: 

1949 _LOG.verbose("Creating output run %s", run) 

1950 

1951 n_refs = len(refs_to_import) 

1952 _LOG.verbose( 

1953 "Importing %d ref%s of dataset type %s into run %s", 

1954 n_refs, 

1955 "" if n_refs == 1 else "s", 

1956 datasetType.name, 

1957 run, 

1958 ) 

1959 

1960 # Assume we are using UUIDs and the source refs will match 

1961 # those imported. 

1962 imported_refs = self._registry._importDatasets(refs_to_import) 

1963 assert set(imported_refs) == set(refs_to_import) 

1964 n_imported += len(imported_refs) 

1965 

1966 assert len(source_refs) == n_imported 

1967 _LOG.verbose("Imported %d datasets into destination butler", n_imported) 

1968 

1969 # Ask the datastore to transfer. The datastore has to check that 

1970 # the source datastore is compatible with the target datastore. 

1971 accepted, rejected = self._datastore.transfer_from( 

1972 source_butler._datastore, 

1973 source_refs, 

1974 transfer=transfer, 

1975 artifact_existence=artifact_existence, 

1976 ) 

1977 if rejected: 

1978 # For now, accept the registry entries but not the files. 

1979 _LOG.warning( 

1980 "%d datasets were rejected and %d accepted for dataset type %s in run %r.", 

1981 len(rejected), 

1982 len(accepted), 

1983 datasetType, 

1984 run, 

1985 ) 

1986 

1987 return source_refs 

1988 

1989 def validateConfiguration( 

1990 self, 

1991 logFailures: bool = False, 

1992 datasetTypeNames: Iterable[str] | None = None, 

1993 ignore: Iterable[str] | None = None, 

1994 ) -> None: 

1995 # Docstring inherited. 

1996 if datasetTypeNames: 

1997 datasetTypes = [self._registry.getDatasetType(name) for name in datasetTypeNames] 

1998 else: 

1999 datasetTypes = list(self._registry.queryDatasetTypes()) 

2000 

2001 # filter out anything from the ignore list 

2002 if ignore: 

2003 ignore = set(ignore) 

2004 datasetTypes = [ 

2005 e for e in datasetTypes if e.name not in ignore and e.nameAndComponent()[0] not in ignore 

2006 ] 

2007 else: 

2008 ignore = set() 

2009 

2010 # For each datasetType that has an instrument dimension, create 

2011 # a DatasetRef for each defined instrument 

2012 datasetRefs = [] 

2013 

2014 # Find all the registered instruments (if "instrument" is in the 

2015 # universe). 

2016 if "instrument" in self.dimensions: 

2017 instruments = {record.name for record in self._registry.queryDimensionRecords("instrument")} 

2018 

2019 for datasetType in datasetTypes: 

2020 if "instrument" in datasetType.dimensions: 

2021 # In order to create a conforming dataset ref, create 

2022 # fake DataCoordinate values for the non-instrument 

2023 # dimensions. The type of the value does not matter here. 

2024 dataId = {dim.name: 1 for dim in datasetType.dimensions if dim.name != "instrument"} 

2025 

2026 for instrument in instruments: 

2027 datasetRef = DatasetRef( 

2028 datasetType, 

2029 DataCoordinate.standardize( 

2030 dataId, instrument=instrument, graph=datasetType.dimensions 

2031 ), 

2032 run="validate", 

2033 ) 

2034 datasetRefs.append(datasetRef) 

2035 

2036 entities: list[DatasetType | DatasetRef] = [] 

2037 entities.extend(datasetTypes) 

2038 entities.extend(datasetRefs) 

2039 

2040 datastoreErrorStr = None 

2041 try: 

2042 self._datastore.validateConfiguration(entities, logFailures=logFailures) 

2043 except ValidationError as e: 

2044 datastoreErrorStr = str(e) 

2045 

2046 # Also check that the LookupKeys used by the datastores match 

2047 # registry and storage class definitions 

2048 keys = self._datastore.getLookupKeys() 

2049 

2050 failedNames = set() 

2051 failedDataId = set() 

2052 for key in keys: 

2053 if key.name is not None: 

2054 if key.name in ignore: 

2055 continue 

2056 

2057 # skip if specific datasetType names were requested and this 

2058 # name does not match 

2059 if datasetTypeNames and key.name not in datasetTypeNames: 

2060 continue 

2061 

2062 # See if it is a StorageClass or a DatasetType 

2063 if key.name in self.storageClasses: 

2064 pass 

2065 else: 

2066 try: 

2067 self._registry.getDatasetType(key.name) 

2068 except KeyError: 

2069 if logFailures: 

2070 _LOG.critical( 

2071 "Key '%s' does not correspond to a DatasetType or StorageClass", key 

2072 ) 

2073 failedNames.add(key) 

2074 else: 

2075 # Dimensions are checked for consistency when the Butler 

2076 # is created and rendezvoused with a universe. 

2077 pass 

2078 

2079 # Check that the instrument is a valid instrument 

2080 # Currently only support instrument so check for that 

2081 if key.dataId: 

2082 dataIdKeys = set(key.dataId) 

2083 if {"instrument"} != dataIdKeys: 

2084 if logFailures: 

2085 _LOG.critical("Key '%s' has unsupported DataId override", key) 

2086 failedDataId.add(key) 

2087 elif key.dataId["instrument"] not in instruments: 

2088 if logFailures: 

2089 _LOG.critical("Key '%s' has unknown instrument", key) 

2090 failedDataId.add(key) 

2091 

2092 messages = [] 

2093 

2094 if datastoreErrorStr: 

2095 messages.append(datastoreErrorStr) 

2096 

2097 for failed, msg in ( 

2098 (failedNames, "Keys without corresponding DatasetType or StorageClass entry: "), 

2099 (failedDataId, "Keys with bad DataId entries: "), 

2100 ): 

2101 if failed: 

2102 msg += ", ".join(str(k) for k in failed) 

2103 messages.append(msg) 

2104 

2105 if messages: 

2106 raise ValidationError(";\n".join(messages)) 

2107 

2108 @property 

2109 def collections(self) -> Sequence[str]: 

2110 """The collections to search by default, in order 

2111 (`~collections.abc.Sequence` [ `str` ]). 

2112 

2113 This is an alias for ``self.registry.defaults.collections``. It cannot 

2114 be set directly in isolation, but all defaults may be changed together 

2115 by assigning a new `RegistryDefaults` instance to 

2116 ``self.registry.defaults``. 

2117 """ 

2118 return self._registry.defaults.collections 

2119 

2120 @property 

2121 def run(self) -> str | None: 

2122 """Name of the run this butler writes outputs to by default (`str` or 

2123 `None`). 

2124 

2125 This is an alias for ``self.registry.defaults.run``. It cannot be set 

2126 directly in isolation, but all defaults may be changed together by 

2127 assigning a new `RegistryDefaults` instance to 

2128 ``self.registry.defaults``. 

2129 """ 

2130 return self._registry.defaults.run 

2131 

2132 @property 

2133 def registry(self) -> Registry: 

2134 """The object that manages dataset metadata and relationships 

2135 (`Registry`). 

2136 

2137 Many operations that don't involve reading or writing butler datasets 

2138 are accessible only via `Registry` methods. Eventually these methods 

2139 will be replaced by equivalent `Butler` methods. 

2140 """ 

2141 return self._registry_shim 

2142 

2143 @property 

2144 def dimensions(self) -> DimensionUniverse: 

2145 # Docstring inherited. 

2146 return self._registry.dimensions 

2147 

2148 _registry: SqlRegistry 

2149 """The object that manages dataset metadata and relationships 

2150 (`SqlRegistry`). 

2151 

2152 Most operations that don't involve reading or writing butler datasets are 

2153 accessible only via `SqlRegistry` methods. 

2154 """ 

2155 

2156 datastore: Datastore 

2157 """The object that manages actual dataset storage (`Datastore`). 

2158 

2159 Direct user access to the datastore should rarely be necessary; the primary 

2160 exception is the case where a `Datastore` implementation provides extra 

2161 functionality beyond what the base class defines. 

2162 """ 

2163 

2164 storageClasses: StorageClassFactory 

2165 """An object that maps known storage class names to objects that fully 

2166 describe them (`StorageClassFactory`). 

2167 """