Coverage for python/lsst/daf/butler/datastores/fileDatastore.py: 85%

921 statements  

« prev     ^ index     » next       coverage.py v6.5.0, created at 2022-12-08 14:18 -0800

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 program is free software: you can redistribute it and/or modify 

10# it under the terms of the GNU General Public License as published by 

11# the Free Software Foundation, either version 3 of the License, or 

12# (at your option) any later version. 

13# 

14# This program is distributed in the hope that it will be useful, 

15# but WITHOUT ANY WARRANTY; without even the implied warranty of 

16# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 

17# GNU General Public License for more details. 

18# 

19# You should have received a copy of the GNU General Public License 

20# along with this program. If not, see <http://www.gnu.org/licenses/>. 

21from __future__ import annotations 

22 

23"""Generic file-based datastore code.""" 

24 

25__all__ = ("FileDatastore",) 

26 

27import hashlib 

28import logging 

29from collections import defaultdict 

30from dataclasses import dataclass 

31from typing import ( 

32 TYPE_CHECKING, 

33 Any, 

34 ClassVar, 

35 Dict, 

36 Iterable, 

37 List, 

38 Mapping, 

39 Optional, 

40 Sequence, 

41 Set, 

42 Tuple, 

43 Type, 

44 Union, 

45) 

46 

47from lsst.daf.butler import ( 

48 CompositesMap, 

49 Config, 

50 DatasetId, 

51 DatasetRef, 

52 DatasetRefURIs, 

53 DatasetType, 

54 DatasetTypeNotSupportedError, 

55 Datastore, 

56 DatastoreCacheManager, 

57 DatastoreConfig, 

58 DatastoreDisabledCacheManager, 

59 DatastoreRecordData, 

60 DatastoreValidationError, 

61 FileDataset, 

62 FileDescriptor, 

63 FileTemplates, 

64 FileTemplateValidationError, 

65 Formatter, 

66 FormatterFactory, 

67 Location, 

68 LocationFactory, 

69 Progress, 

70 StorageClass, 

71 StoredDatastoreItemInfo, 

72 StoredFileInfo, 

73 ddl, 

74) 

75from lsst.daf.butler.core.repoRelocation import replaceRoot 

76from lsst.daf.butler.core.utils import transactional 

77from lsst.daf.butler.registry.interfaces import DatastoreRegistryBridge, ReadOnlyDatabaseError 

78from lsst.resources import ResourcePath, ResourcePathExpression 

79from lsst.utils.introspection import get_class_of, get_instance_of 

80from lsst.utils.iteration import chunk_iterable 

81 

82# For VERBOSE logging usage. 

83from lsst.utils.logging import VERBOSE, getLogger 

84from lsst.utils.timer import time_this 

85from sqlalchemy import BigInteger, String 

86 

87from ..registry.interfaces import FakeDatasetRef 

88from .genericDatastore import GenericBaseDatastore 

89 

90if TYPE_CHECKING: 90 ↛ 91line 90 didn't jump to line 91, because the condition on line 90 was never true

91 from lsst.daf.butler import AbstractDatastoreCacheManager, LookupKey 

92 from lsst.daf.butler.registry.interfaces import DatasetIdRef, DatastoreRegistryBridgeManager 

93 

94log = getLogger(__name__) 

95 

96 

97class _IngestPrepData(Datastore.IngestPrepData): 

98 """Helper class for FileDatastore ingest implementation. 

99 

100 Parameters 

101 ---------- 

102 datasets : `list` of `FileDataset` 

103 Files to be ingested by this datastore. 

104 """ 

105 

106 def __init__(self, datasets: List[FileDataset]): 

107 super().__init__(ref for dataset in datasets for ref in dataset.refs) 

108 self.datasets = datasets 

109 

110 

111@dataclass(frozen=True) 

112class DatastoreFileGetInformation: 

113 """Collection of useful parameters needed to retrieve a file from 

114 a Datastore. 

115 """ 

116 

117 location: Location 

118 """The location from which to read the dataset.""" 

119 

120 formatter: Formatter 

121 """The `Formatter` to use to deserialize the dataset.""" 

122 

123 info: StoredFileInfo 

124 """Stored information about this file and its formatter.""" 

125 

126 assemblerParams: Mapping[str, Any] 

127 """Parameters to use for post-processing the retrieved dataset.""" 

128 

129 formatterParams: Mapping[str, Any] 

130 """Parameters that were understood by the associated formatter.""" 

131 

132 component: Optional[str] 

133 """The component to be retrieved (can be `None`).""" 

134 

135 readStorageClass: StorageClass 

136 """The `StorageClass` of the dataset being read.""" 

137 

138 

139class FileDatastore(GenericBaseDatastore): 

140 """Generic Datastore for file-based implementations. 

141 

142 Should always be sub-classed since key abstract methods are missing. 

143 

144 Parameters 

145 ---------- 

146 config : `DatastoreConfig` or `str` 

147 Configuration as either a `Config` object or URI to file. 

148 bridgeManager : `DatastoreRegistryBridgeManager` 

149 Object that manages the interface between `Registry` and datastores. 

150 butlerRoot : `str`, optional 

151 New datastore root to use to override the configuration value. 

152 

153 Raises 

154 ------ 

155 ValueError 

156 If root location does not exist and ``create`` is `False` in the 

157 configuration. 

158 """ 

159 

160 defaultConfigFile: ClassVar[Optional[str]] = None 

161 """Path to configuration defaults. Accessed within the ``config`` resource 

162 or relative to a search path. Can be None if no defaults specified. 

163 """ 

164 

165 root: ResourcePath 

166 """Root directory URI of this `Datastore`.""" 

167 

168 locationFactory: LocationFactory 

169 """Factory for creating locations relative to the datastore root.""" 

170 

171 formatterFactory: FormatterFactory 

172 """Factory for creating instances of formatters.""" 

173 

174 templates: FileTemplates 

175 """File templates that can be used by this `Datastore`.""" 

176 

177 composites: CompositesMap 

178 """Determines whether a dataset should be disassembled on put.""" 

179 

180 defaultConfigFile = "datastores/fileDatastore.yaml" 

181 """Path to configuration defaults. Accessed within the ``config`` resource 

182 or relative to a search path. Can be None if no defaults specified. 

183 """ 

184 

185 @classmethod 

186 def setConfigRoot(cls, root: str, config: Config, full: Config, overwrite: bool = True) -> None: 

187 """Set any filesystem-dependent config options for this Datastore to 

188 be appropriate for a new empty repository with the given root. 

189 

190 Parameters 

191 ---------- 

192 root : `str` 

193 URI to the root of the data repository. 

194 config : `Config` 

195 A `Config` to update. Only the subset understood by 

196 this component will be updated. Will not expand 

197 defaults. 

198 full : `Config` 

199 A complete config with all defaults expanded that can be 

200 converted to a `DatastoreConfig`. Read-only and will not be 

201 modified by this method. 

202 Repository-specific options that should not be obtained 

203 from defaults when Butler instances are constructed 

204 should be copied from ``full`` to ``config``. 

205 overwrite : `bool`, optional 

206 If `False`, do not modify a value in ``config`` if the value 

207 already exists. Default is always to overwrite with the provided 

208 ``root``. 

209 

210 Notes 

211 ----- 

212 If a keyword is explicitly defined in the supplied ``config`` it 

213 will not be overridden by this method if ``overwrite`` is `False`. 

214 This allows explicit values set in external configs to be retained. 

215 """ 

216 Config.updateParameters( 

217 DatastoreConfig, 

218 config, 

219 full, 

220 toUpdate={"root": root}, 

221 toCopy=("cls", ("records", "table")), 

222 overwrite=overwrite, 

223 ) 

224 

225 @classmethod 

226 def makeTableSpec(cls, datasetIdColumnType: type) -> ddl.TableSpec: 

227 return ddl.TableSpec( 

228 fields=[ 

229 ddl.FieldSpec(name="dataset_id", dtype=datasetIdColumnType, primaryKey=True), 

230 ddl.FieldSpec(name="path", dtype=String, length=256, nullable=False), 

231 ddl.FieldSpec(name="formatter", dtype=String, length=128, nullable=False), 

232 ddl.FieldSpec(name="storage_class", dtype=String, length=64, nullable=False), 

233 # Use empty string to indicate no component 

234 ddl.FieldSpec(name="component", dtype=String, length=32, primaryKey=True), 

235 # TODO: should checksum be Base64Bytes instead? 

236 ddl.FieldSpec(name="checksum", dtype=String, length=128, nullable=True), 

237 ddl.FieldSpec(name="file_size", dtype=BigInteger, nullable=True), 

238 ], 

239 unique=frozenset(), 

240 indexes=[tuple(["path"])], 

241 ) 

242 

243 def __init__( 

244 self, 

245 config: Union[DatastoreConfig, str], 

246 bridgeManager: DatastoreRegistryBridgeManager, 

247 butlerRoot: str = None, 

248 ): 

249 super().__init__(config, bridgeManager) 

250 if "root" not in self.config: 250 ↛ 251line 250 didn't jump to line 251, because the condition on line 250 was never true

251 raise ValueError("No root directory specified in configuration") 

252 

253 self._bridgeManager = bridgeManager 

254 

255 # Name ourselves either using an explicit name or a name 

256 # derived from the (unexpanded) root 

257 if "name" in self.config: 

258 self.name = self.config["name"] 

259 else: 

260 # We use the unexpanded root in the name to indicate that this 

261 # datastore can be moved without having to update registry. 

262 self.name = "{}@{}".format(type(self).__name__, self.config["root"]) 

263 

264 # Support repository relocation in config 

265 # Existence of self.root is checked in subclass 

266 self.root = ResourcePath( 

267 replaceRoot(self.config["root"], butlerRoot), forceDirectory=True, forceAbsolute=True 

268 ) 

269 

270 self.locationFactory = LocationFactory(self.root) 

271 self.formatterFactory = FormatterFactory() 

272 

273 # Now associate formatters with storage classes 

274 self.formatterFactory.registerFormatters(self.config["formatters"], universe=bridgeManager.universe) 

275 

276 # Read the file naming templates 

277 self.templates = FileTemplates(self.config["templates"], universe=bridgeManager.universe) 

278 

279 # See if composites should be disassembled 

280 self.composites = CompositesMap(self.config["composites"], universe=bridgeManager.universe) 

281 

282 tableName = self.config["records", "table"] 

283 try: 

284 # Storage of paths and formatters, keyed by dataset_id 

285 self._table = bridgeManager.opaque.register( 

286 tableName, self.makeTableSpec(bridgeManager.datasetIdColumnType) 

287 ) 

288 # Interface to Registry. 

289 self._bridge = bridgeManager.register(self.name) 

290 except ReadOnlyDatabaseError: 

291 # If the database is read only and we just tried and failed to 

292 # create a table, it means someone is trying to create a read-only 

293 # butler client for an empty repo. That should be okay, as long 

294 # as they then try to get any datasets before some other client 

295 # creates the table. Chances are they'rejust validating 

296 # configuration. 

297 pass 

298 

299 # Determine whether checksums should be used - default to False 

300 self.useChecksum = self.config.get("checksum", False) 

301 

302 # Determine whether we can fall back to configuration if a 

303 # requested dataset is not known to registry 

304 self.trustGetRequest = self.config.get("trust_get_request", False) 

305 

306 # Create a cache manager 

307 self.cacheManager: AbstractDatastoreCacheManager 

308 if "cached" in self.config: 308 ↛ 311line 308 didn't jump to line 311, because the condition on line 308 was never false

309 self.cacheManager = DatastoreCacheManager(self.config["cached"], universe=bridgeManager.universe) 

310 else: 

311 self.cacheManager = DatastoreDisabledCacheManager("", universe=bridgeManager.universe) 

312 

313 # Check existence and create directory structure if necessary 

314 if not self.root.exists(): 

315 if "create" not in self.config or not self.config["create"]: 315 ↛ 316line 315 didn't jump to line 316, because the condition on line 315 was never true

316 raise ValueError(f"No valid root and not allowed to create one at: {self.root}") 

317 try: 

318 self.root.mkdir() 

319 except Exception as e: 

320 raise ValueError( 

321 f"Can not create datastore root '{self.root}', check permissions. Got error: {e}" 

322 ) from e 

323 

324 def __str__(self) -> str: 

325 return str(self.root) 

326 

327 @property 

328 def bridge(self) -> DatastoreRegistryBridge: 

329 return self._bridge 

330 

331 def _artifact_exists(self, location: Location) -> bool: 

332 """Check that an artifact exists in this datastore at the specified 

333 location. 

334 

335 Parameters 

336 ---------- 

337 location : `Location` 

338 Expected location of the artifact associated with this datastore. 

339 

340 Returns 

341 ------- 

342 exists : `bool` 

343 True if the location can be found, false otherwise. 

344 """ 

345 log.debug("Checking if resource exists: %s", location.uri) 

346 return location.uri.exists() 

347 

348 def _delete_artifact(self, location: Location) -> None: 

349 """Delete the artifact from the datastore. 

350 

351 Parameters 

352 ---------- 

353 location : `Location` 

354 Location of the artifact associated with this datastore. 

355 """ 

356 if location.pathInStore.isabs(): 356 ↛ 357line 356 didn't jump to line 357, because the condition on line 356 was never true

357 raise RuntimeError(f"Cannot delete artifact with absolute uri {location.uri}.") 

358 

359 try: 

360 location.uri.remove() 

361 except FileNotFoundError: 

362 log.debug("File %s did not exist and so could not be deleted.", location.uri) 

363 raise 

364 except Exception as e: 

365 log.critical("Failed to delete file: %s (%s)", location.uri, e) 

366 raise 

367 log.debug("Successfully deleted file: %s", location.uri) 

368 

369 def addStoredItemInfo(self, refs: Iterable[DatasetRef], infos: Iterable[StoredFileInfo]) -> None: 

370 # Docstring inherited from GenericBaseDatastore 

371 records = [info.rebase(ref).to_record() for ref, info in zip(refs, infos)] 

372 self._table.insert(*records) 

373 

374 def getStoredItemsInfo(self, ref: DatasetIdRef) -> List[StoredFileInfo]: 

375 # Docstring inherited from GenericBaseDatastore 

376 

377 # Look for the dataset_id -- there might be multiple matches 

378 # if we have disassembled the dataset. 

379 records = self._table.fetch(dataset_id=ref.id) 

380 return [StoredFileInfo.from_record(record) for record in records] 

381 

382 def _get_stored_records_associated_with_refs( 

383 self, refs: Iterable[DatasetIdRef] 

384 ) -> Dict[DatasetId, List[StoredFileInfo]]: 

385 """Retrieve all records associated with the provided refs. 

386 

387 Parameters 

388 ---------- 

389 refs : iterable of `DatasetIdRef` 

390 The refs for which records are to be retrieved. 

391 

392 Returns 

393 ------- 

394 records : `dict` of [`DatasetId`, `list` of `StoredFileInfo`] 

395 The matching records indexed by the ref ID. The number of entries 

396 in the dict can be smaller than the number of requested refs. 

397 """ 

398 records = self._table.fetch(dataset_id=[ref.id for ref in refs]) 

399 

400 # Uniqueness is dataset_id + component so can have multiple records 

401 # per ref. 

402 records_by_ref = defaultdict(list) 

403 for record in records: 

404 records_by_ref[record["dataset_id"]].append(StoredFileInfo.from_record(record)) 

405 return records_by_ref 

406 

407 def _refs_associated_with_artifacts( 

408 self, paths: List[Union[str, ResourcePath]] 

409 ) -> Dict[str, Set[DatasetId]]: 

410 """Return paths and associated dataset refs. 

411 

412 Parameters 

413 ---------- 

414 paths : `list` of `str` or `lsst.resources.ResourcePath` 

415 All the paths to include in search. 

416 

417 Returns 

418 ------- 

419 mapping : `dict` of [`str`, `set` [`DatasetId`]] 

420 Mapping of each path to a set of associated database IDs. 

421 """ 

422 records = self._table.fetch(path=[str(path) for path in paths]) 

423 result = defaultdict(set) 

424 for row in records: 

425 result[row["path"]].add(row["dataset_id"]) 

426 return result 

427 

428 def _registered_refs_per_artifact(self, pathInStore: ResourcePath) -> Set[DatasetId]: 

429 """Return all dataset refs associated with the supplied path. 

430 

431 Parameters 

432 ---------- 

433 pathInStore : `lsst.resources.ResourcePath` 

434 Path of interest in the data store. 

435 

436 Returns 

437 ------- 

438 ids : `set` of `int` 

439 All `DatasetRef` IDs associated with this path. 

440 """ 

441 records = list(self._table.fetch(path=str(pathInStore))) 

442 ids = {r["dataset_id"] for r in records} 

443 return ids 

444 

445 def removeStoredItemInfo(self, ref: DatasetIdRef) -> None: 

446 # Docstring inherited from GenericBaseDatastore 

447 self._table.delete(["dataset_id"], {"dataset_id": ref.id}) 

448 

449 def _get_dataset_locations_info(self, ref: DatasetIdRef) -> List[Tuple[Location, StoredFileInfo]]: 

450 r"""Find all the `Location`\ s of the requested dataset in the 

451 `Datastore` and the associated stored file information. 

452 

453 Parameters 

454 ---------- 

455 ref : `DatasetRef` 

456 Reference to the required `Dataset`. 

457 

458 Returns 

459 ------- 

460 results : `list` [`tuple` [`Location`, `StoredFileInfo` ]] 

461 Location of the dataset within the datastore and 

462 stored information about each file and its formatter. 

463 """ 

464 # Get the file information (this will fail if no file) 

465 records = self.getStoredItemsInfo(ref) 

466 

467 # Use the path to determine the location -- we need to take 

468 # into account absolute URIs in the datastore record 

469 return [(r.file_location(self.locationFactory), r) for r in records] 

470 

471 def _can_remove_dataset_artifact(self, ref: DatasetIdRef, location: Location) -> bool: 

472 """Check that there is only one dataset associated with the 

473 specified artifact. 

474 

475 Parameters 

476 ---------- 

477 ref : `DatasetRef` or `FakeDatasetRef` 

478 Dataset to be removed. 

479 location : `Location` 

480 The location of the artifact to be removed. 

481 

482 Returns 

483 ------- 

484 can_remove : `Bool` 

485 True if the artifact can be safely removed. 

486 """ 

487 # Can't ever delete absolute URIs. 

488 if location.pathInStore.isabs(): 

489 return False 

490 

491 # Get all entries associated with this path 

492 allRefs = self._registered_refs_per_artifact(location.pathInStore) 

493 if not allRefs: 

494 raise RuntimeError(f"Datastore inconsistency error. {location.pathInStore} not in registry") 

495 

496 # Remove these refs from all the refs and if there is nothing left 

497 # then we can delete 

498 remainingRefs = allRefs - {ref.id} 

499 

500 if remainingRefs: 

501 return False 

502 return True 

503 

504 def _get_expected_dataset_locations_info(self, ref: DatasetRef) -> List[Tuple[Location, StoredFileInfo]]: 

505 """Predict the location and related file information of the requested 

506 dataset in this datastore. 

507 

508 Parameters 

509 ---------- 

510 ref : `DatasetRef` 

511 Reference to the required `Dataset`. 

512 

513 Returns 

514 ------- 

515 results : `list` [`tuple` [`Location`, `StoredFileInfo` ]] 

516 Expected Location of the dataset within the datastore and 

517 placeholder information about each file and its formatter. 

518 

519 Notes 

520 ----- 

521 Uses the current configuration to determine how we would expect the 

522 datastore files to have been written if we couldn't ask registry. 

523 This is safe so long as there has been no change to datastore 

524 configuration between writing the dataset and wanting to read it. 

525 Will not work for files that have been ingested without using the 

526 standard file template or default formatter. 

527 """ 

528 

529 # If we have a component ref we always need to ask the questions 

530 # of the composite. If the composite is disassembled this routine 

531 # should return all components. If the composite was not 

532 # disassembled the composite is what is stored regardless of 

533 # component request. Note that if the caller has disassembled 

534 # a composite there is no way for this guess to know that 

535 # without trying both the composite and component ref and seeing 

536 # if there is something at the component Location even without 

537 # disassembly being enabled. 

538 if ref.datasetType.isComponent(): 

539 ref = ref.makeCompositeRef() 

540 

541 # See if the ref is a composite that should be disassembled 

542 doDisassembly = self.composites.shouldBeDisassembled(ref) 

543 

544 all_info: List[Tuple[Location, Formatter, StorageClass, Optional[str]]] = [] 

545 

546 if doDisassembly: 

547 for component, componentStorage in ref.datasetType.storageClass.components.items(): 

548 compRef = ref.makeComponentRef(component) 

549 location, formatter = self._determine_put_formatter_location(compRef) 

550 all_info.append((location, formatter, componentStorage, component)) 

551 

552 else: 

553 # Always use the composite ref if no disassembly 

554 location, formatter = self._determine_put_formatter_location(ref) 

555 all_info.append((location, formatter, ref.datasetType.storageClass, None)) 

556 

557 # Convert the list of tuples to have StoredFileInfo as second element 

558 return [ 

559 ( 

560 location, 

561 StoredFileInfo( 

562 formatter=formatter, 

563 path=location.pathInStore.path, 

564 storageClass=storageClass, 

565 component=component, 

566 checksum=None, 

567 file_size=-1, 

568 dataset_id=ref.getCheckedId(), 

569 ), 

570 ) 

571 for location, formatter, storageClass, component in all_info 

572 ] 

573 

574 def _prepare_for_get( 

575 self, ref: DatasetRef, parameters: Optional[Mapping[str, Any]] = None 

576 ) -> List[DatastoreFileGetInformation]: 

577 """Check parameters for ``get`` and obtain formatter and 

578 location. 

579 

580 Parameters 

581 ---------- 

582 ref : `DatasetRef` 

583 Reference to the required Dataset. 

584 parameters : `dict` 

585 `StorageClass`-specific parameters that specify, for example, 

586 a slice of the dataset to be loaded. 

587 

588 Returns 

589 ------- 

590 getInfo : `list` [`DatastoreFileGetInformation`] 

591 Parameters needed to retrieve each file. 

592 """ 

593 log.debug("Retrieve %s from %s with parameters %s", ref, self.name, parameters) 

594 

595 # Get file metadata and internal metadata 

596 fileLocations = self._get_dataset_locations_info(ref) 

597 if not fileLocations: 

598 if not self.trustGetRequest: 

599 raise FileNotFoundError(f"Could not retrieve dataset {ref}.") 

600 # Assume the dataset is where we think it should be 

601 fileLocations = self._get_expected_dataset_locations_info(ref) 

602 

603 # The storage class we want to use eventually 

604 refStorageClass = ref.datasetType.storageClass 

605 

606 if len(fileLocations) > 1: 

607 disassembled = True 

608 

609 # If trust is involved it is possible that there will be 

610 # components listed here that do not exist in the datastore. 

611 # Explicitly check for file artifact existence and filter out any 

612 # that are missing. 

613 if self.trustGetRequest: 

614 fileLocations = [loc for loc in fileLocations if loc[0].uri.exists()] 

615 

616 # For now complain only if we have no components at all. One 

617 # component is probably a problem but we can punt that to the 

618 # assembler. 

619 if not fileLocations: 619 ↛ 620line 619 didn't jump to line 620, because the condition on line 619 was never true

620 raise FileNotFoundError(f"None of the component files for dataset {ref} exist.") 

621 

622 else: 

623 disassembled = False 

624 

625 # Is this a component request? 

626 refComponent = ref.datasetType.component() 

627 

628 fileGetInfo = [] 

629 for location, storedFileInfo in fileLocations: 

630 

631 # The storage class used to write the file 

632 writeStorageClass = storedFileInfo.storageClass 

633 

634 # If this has been disassembled we need read to match the write 

635 if disassembled: 

636 readStorageClass = writeStorageClass 

637 else: 

638 readStorageClass = refStorageClass 

639 

640 formatter = get_instance_of( 

641 storedFileInfo.formatter, 

642 FileDescriptor( 

643 location, 

644 readStorageClass=readStorageClass, 

645 storageClass=writeStorageClass, 

646 parameters=parameters, 

647 ), 

648 ref.dataId, 

649 ) 

650 

651 formatterParams, notFormatterParams = formatter.segregateParameters() 

652 

653 # Of the remaining parameters, extract the ones supported by 

654 # this StorageClass (for components not all will be handled) 

655 assemblerParams = readStorageClass.filterParameters(notFormatterParams) 

656 

657 # The ref itself could be a component if the dataset was 

658 # disassembled by butler, or we disassembled in datastore and 

659 # components came from the datastore records 

660 component = storedFileInfo.component if storedFileInfo.component else refComponent 

661 

662 fileGetInfo.append( 

663 DatastoreFileGetInformation( 

664 location, 

665 formatter, 

666 storedFileInfo, 

667 assemblerParams, 

668 formatterParams, 

669 component, 

670 readStorageClass, 

671 ) 

672 ) 

673 

674 return fileGetInfo 

675 

676 def _prepare_for_put(self, inMemoryDataset: Any, ref: DatasetRef) -> Tuple[Location, Formatter]: 

677 """Check the arguments for ``put`` and obtain formatter and 

678 location. 

679 

680 Parameters 

681 ---------- 

682 inMemoryDataset : `object` 

683 The dataset to store. 

684 ref : `DatasetRef` 

685 Reference to the associated Dataset. 

686 

687 Returns 

688 ------- 

689 location : `Location` 

690 The location to write the dataset. 

691 formatter : `Formatter` 

692 The `Formatter` to use to write the dataset. 

693 

694 Raises 

695 ------ 

696 TypeError 

697 Supplied object and storage class are inconsistent. 

698 DatasetTypeNotSupportedError 

699 The associated `DatasetType` is not handled by this datastore. 

700 """ 

701 self._validate_put_parameters(inMemoryDataset, ref) 

702 return self._determine_put_formatter_location(ref) 

703 

704 def _determine_put_formatter_location(self, ref: DatasetRef) -> Tuple[Location, Formatter]: 

705 """Calculate the formatter and output location to use for put. 

706 

707 Parameters 

708 ---------- 

709 ref : `DatasetRef` 

710 Reference to the associated Dataset. 

711 

712 Returns 

713 ------- 

714 location : `Location` 

715 The location to write the dataset. 

716 formatter : `Formatter` 

717 The `Formatter` to use to write the dataset. 

718 """ 

719 # Work out output file name 

720 try: 

721 template = self.templates.getTemplate(ref) 

722 except KeyError as e: 

723 raise DatasetTypeNotSupportedError(f"Unable to find template for {ref}") from e 

724 

725 # Validate the template to protect against filenames from different 

726 # dataIds returning the same and causing overwrite confusion. 

727 template.validateTemplate(ref) 

728 

729 location = self.locationFactory.fromPath(template.format(ref)) 

730 

731 # Get the formatter based on the storage class 

732 storageClass = ref.datasetType.storageClass 

733 try: 

734 formatter = self.formatterFactory.getFormatter( 

735 ref, FileDescriptor(location, storageClass=storageClass), ref.dataId 

736 ) 

737 except KeyError as e: 

738 raise DatasetTypeNotSupportedError( 

739 f"Unable to find formatter for {ref} in datastore {self.name}" 

740 ) from e 

741 

742 # Now that we know the formatter, update the location 

743 location = formatter.makeUpdatedLocation(location) 

744 

745 return location, formatter 

746 

747 def _overrideTransferMode(self, *datasets: FileDataset, transfer: Optional[str] = None) -> Optional[str]: 

748 # Docstring inherited from base class 

749 if transfer != "auto": 

750 return transfer 

751 

752 # See if the paths are within the datastore or not 

753 inside = [self._pathInStore(d.path) is not None for d in datasets] 

754 

755 if all(inside): 

756 transfer = None 

757 elif not any(inside): 757 ↛ 766line 757 didn't jump to line 766, because the condition on line 757 was never false

758 # Allow ResourcePath to use its own knowledge 

759 transfer = "auto" 

760 else: 

761 # This can happen when importing from a datastore that 

762 # has had some datasets ingested using "direct" mode. 

763 # Also allow ResourcePath to sort it out but warn about it. 

764 # This can happen if you are importing from a datastore 

765 # that had some direct transfer datasets. 

766 log.warning( 

767 "Some datasets are inside the datastore and some are outside. Using 'split' " 

768 "transfer mode. This assumes that the files outside the datastore are " 

769 "still accessible to the new butler since they will not be copied into " 

770 "the target datastore." 

771 ) 

772 transfer = "split" 

773 

774 return transfer 

775 

776 def _pathInStore(self, path: ResourcePathExpression) -> Optional[str]: 

777 """Return path relative to datastore root 

778 

779 Parameters 

780 ---------- 

781 path : `lsst.resources.ResourcePathExpression` 

782 Path to dataset. Can be absolute URI. If relative assumed to 

783 be relative to the datastore. Returns path in datastore 

784 or raises an exception if the path it outside. 

785 

786 Returns 

787 ------- 

788 inStore : `str` 

789 Path relative to datastore root. Returns `None` if the file is 

790 outside the root. 

791 """ 

792 # Relative path will always be relative to datastore 

793 pathUri = ResourcePath(path, forceAbsolute=False) 

794 return pathUri.relative_to(self.root) 

795 

796 def _standardizeIngestPath( 

797 self, path: Union[str, ResourcePath], *, transfer: Optional[str] = None 

798 ) -> Union[str, ResourcePath]: 

799 """Standardize the path of a to-be-ingested file. 

800 

801 Parameters 

802 ---------- 

803 path : `str` or `lsst.resources.ResourcePath` 

804 Path of a file to be ingested. This parameter is not expected 

805 to be all the types that can be used to construct a 

806 `~lsst.resources.ResourcePath`. 

807 transfer : `str`, optional 

808 How (and whether) the dataset should be added to the datastore. 

809 See `ingest` for details of transfer modes. 

810 This implementation is provided only so 

811 `NotImplementedError` can be raised if the mode is not supported; 

812 actual transfers are deferred to `_extractIngestInfo`. 

813 

814 Returns 

815 ------- 

816 path : `str` or `lsst.resources.ResourcePath` 

817 New path in what the datastore considers standard form. If an 

818 absolute URI was given that will be returned unchanged. 

819 

820 Notes 

821 ----- 

822 Subclasses of `FileDatastore` can implement this method instead 

823 of `_prepIngest`. It should not modify the data repository or given 

824 file in any way. 

825 

826 Raises 

827 ------ 

828 NotImplementedError 

829 Raised if the datastore does not support the given transfer mode 

830 (including the case where ingest is not supported at all). 

831 FileNotFoundError 

832 Raised if one of the given files does not exist. 

833 """ 

834 if transfer not in (None, "direct", "split") + self.root.transferModes: 834 ↛ 835line 834 didn't jump to line 835, because the condition on line 834 was never true

835 raise NotImplementedError(f"Transfer mode {transfer} not supported.") 

836 

837 # A relative URI indicates relative to datastore root 

838 srcUri = ResourcePath(path, forceAbsolute=False) 

839 if not srcUri.isabs(): 

840 srcUri = self.root.join(path) 

841 

842 if not srcUri.exists(): 

843 raise FileNotFoundError( 

844 f"Resource at {srcUri} does not exist; note that paths to ingest " 

845 f"are assumed to be relative to {self.root} unless they are absolute." 

846 ) 

847 

848 if transfer is None: 

849 relpath = srcUri.relative_to(self.root) 

850 if not relpath: 

851 raise RuntimeError( 

852 f"Transfer is none but source file ({srcUri}) is not within datastore ({self.root})" 

853 ) 

854 

855 # Return the relative path within the datastore for internal 

856 # transfer 

857 path = relpath 

858 

859 return path 

860 

861 def _extractIngestInfo( 

862 self, 

863 path: ResourcePathExpression, 

864 ref: DatasetRef, 

865 *, 

866 formatter: Union[Formatter, Type[Formatter]], 

867 transfer: Optional[str] = None, 

868 record_validation_info: bool = True, 

869 ) -> StoredFileInfo: 

870 """Relocate (if necessary) and extract `StoredFileInfo` from a 

871 to-be-ingested file. 

872 

873 Parameters 

874 ---------- 

875 path : `lsst.resources.ResourcePathExpression` 

876 URI or path of a file to be ingested. 

877 ref : `DatasetRef` 

878 Reference for the dataset being ingested. Guaranteed to have 

879 ``dataset_id not None`. 

880 formatter : `type` or `Formatter` 

881 `Formatter` subclass to use for this dataset or an instance. 

882 transfer : `str`, optional 

883 How (and whether) the dataset should be added to the datastore. 

884 See `ingest` for details of transfer modes. 

885 record_validation_info : `bool`, optional 

886 If `True`, the default, the datastore can record validation 

887 information associated with the file. If `False` the datastore 

888 will not attempt to track any information such as checksums 

889 or file sizes. This can be useful if such information is tracked 

890 in an external system or if the file is to be compressed in place. 

891 It is up to the datastore whether this parameter is relevant. 

892 

893 Returns 

894 ------- 

895 info : `StoredFileInfo` 

896 Internal datastore record for this file. This will be inserted by 

897 the caller; the `_extractIngestInfo` is only responsible for 

898 creating and populating the struct. 

899 

900 Raises 

901 ------ 

902 FileNotFoundError 

903 Raised if one of the given files does not exist. 

904 FileExistsError 

905 Raised if transfer is not `None` but the (internal) location the 

906 file would be moved to is already occupied. 

907 """ 

908 if self._transaction is None: 908 ↛ 909line 908 didn't jump to line 909, because the condition on line 908 was never true

909 raise RuntimeError("Ingest called without transaction enabled") 

910 

911 # Create URI of the source path, do not need to force a relative 

912 # path to absolute. 

913 srcUri = ResourcePath(path, forceAbsolute=False) 

914 

915 # Track whether we have read the size of the source yet 

916 have_sized = False 

917 

918 tgtLocation: Optional[Location] 

919 if transfer is None or transfer == "split": 

920 # A relative path is assumed to be relative to the datastore 

921 # in this context 

922 if not srcUri.isabs(): 

923 tgtLocation = self.locationFactory.fromPath(srcUri.ospath) 

924 else: 

925 # Work out the path in the datastore from an absolute URI 

926 # This is required to be within the datastore. 

927 pathInStore = srcUri.relative_to(self.root) 

928 if pathInStore is None and transfer is None: 928 ↛ 929line 928 didn't jump to line 929, because the condition on line 928 was never true

929 raise RuntimeError( 

930 f"Unexpectedly learned that {srcUri} is not within datastore {self.root}" 

931 ) 

932 if pathInStore: 932 ↛ 934line 932 didn't jump to line 934, because the condition on line 932 was never false

933 tgtLocation = self.locationFactory.fromPath(pathInStore) 

934 elif transfer == "split": 

935 # Outside the datastore but treat that as a direct ingest 

936 # instead. 

937 tgtLocation = None 

938 else: 

939 raise RuntimeError(f"Unexpected transfer mode encountered: {transfer} for URI {srcUri}") 

940 elif transfer == "direct": 940 ↛ 945line 940 didn't jump to line 945, because the condition on line 940 was never true

941 # Want to store the full URI to the resource directly in 

942 # datastore. This is useful for referring to permanent archive 

943 # storage for raw data. 

944 # Trust that people know what they are doing. 

945 tgtLocation = None 

946 else: 

947 # Work out the name we want this ingested file to have 

948 # inside the datastore 

949 tgtLocation = self._calculate_ingested_datastore_name(srcUri, ref, formatter) 

950 if not tgtLocation.uri.dirname().exists(): 

951 log.debug("Folder %s does not exist yet.", tgtLocation.uri.dirname()) 

952 tgtLocation.uri.dirname().mkdir() 

953 

954 # if we are transferring from a local file to a remote location 

955 # it may be more efficient to get the size and checksum of the 

956 # local file rather than the transferred one 

957 if record_validation_info and srcUri.isLocal: 

958 size = srcUri.size() 

959 checksum = self.computeChecksum(srcUri) if self.useChecksum else None 

960 have_sized = True 

961 

962 # Transfer the resource to the destination. 

963 # Allow overwrite of an existing file. This matches the behavior 

964 # of datastore.put() in that it trusts that registry would not 

965 # be asking to overwrite unless registry thought that the 

966 # overwrite was allowed. 

967 tgtLocation.uri.transfer_from( 

968 srcUri, transfer=transfer, transaction=self._transaction, overwrite=True 

969 ) 

970 

971 if tgtLocation is None: 971 ↛ 973line 971 didn't jump to line 973, because the condition on line 971 was never true

972 # This means we are using direct mode 

973 targetUri = srcUri 

974 targetPath = str(srcUri) 

975 else: 

976 targetUri = tgtLocation.uri 

977 targetPath = tgtLocation.pathInStore.path 

978 

979 # the file should exist in the datastore now 

980 if record_validation_info: 

981 if not have_sized: 

982 size = targetUri.size() 

983 checksum = self.computeChecksum(targetUri) if self.useChecksum else None 

984 else: 

985 # Not recording any file information. 

986 size = -1 

987 checksum = None 

988 

989 return StoredFileInfo( 

990 formatter=formatter, 

991 path=targetPath, 

992 storageClass=ref.datasetType.storageClass, 

993 component=ref.datasetType.component(), 

994 file_size=size, 

995 checksum=checksum, 

996 dataset_id=ref.getCheckedId(), 

997 ) 

998 

999 def _prepIngest(self, *datasets: FileDataset, transfer: Optional[str] = None) -> _IngestPrepData: 

1000 # Docstring inherited from Datastore._prepIngest. 

1001 filtered = [] 

1002 for dataset in datasets: 

1003 acceptable = [ref for ref in dataset.refs if self.constraints.isAcceptable(ref)] 

1004 if not acceptable: 

1005 continue 

1006 else: 

1007 dataset.refs = acceptable 

1008 if dataset.formatter is None: 

1009 dataset.formatter = self.formatterFactory.getFormatterClass(dataset.refs[0]) 

1010 else: 

1011 assert isinstance(dataset.formatter, (type, str)) 

1012 formatter_class = get_class_of(dataset.formatter) 

1013 if not issubclass(formatter_class, Formatter): 1013 ↛ 1014line 1013 didn't jump to line 1014, because the condition on line 1013 was never true

1014 raise TypeError(f"Requested formatter {dataset.formatter} is not a Formatter class.") 

1015 dataset.formatter = formatter_class 

1016 dataset.path = self._standardizeIngestPath(dataset.path, transfer=transfer) 

1017 filtered.append(dataset) 

1018 return _IngestPrepData(filtered) 

1019 

1020 @transactional 

1021 def _finishIngest( 

1022 self, 

1023 prepData: Datastore.IngestPrepData, 

1024 *, 

1025 transfer: Optional[str] = None, 

1026 record_validation_info: bool = True, 

1027 ) -> None: 

1028 # Docstring inherited from Datastore._finishIngest. 

1029 refsAndInfos = [] 

1030 progress = Progress("lsst.daf.butler.datastores.FileDatastore.ingest", level=logging.DEBUG) 

1031 for dataset in progress.wrap(prepData.datasets, desc="Ingesting dataset files"): 

1032 # Do ingest as if the first dataset ref is associated with the file 

1033 info = self._extractIngestInfo( 

1034 dataset.path, 

1035 dataset.refs[0], 

1036 formatter=dataset.formatter, 

1037 transfer=transfer, 

1038 record_validation_info=record_validation_info, 

1039 ) 

1040 refsAndInfos.extend([(ref, info) for ref in dataset.refs]) 

1041 self._register_datasets(refsAndInfos) 

1042 

1043 def _calculate_ingested_datastore_name( 

1044 self, srcUri: ResourcePath, ref: DatasetRef, formatter: Union[Formatter, Type[Formatter]] 

1045 ) -> Location: 

1046 """Given a source URI and a DatasetRef, determine the name the 

1047 dataset will have inside datastore. 

1048 

1049 Parameters 

1050 ---------- 

1051 srcUri : `lsst.resources.ResourcePath` 

1052 URI to the source dataset file. 

1053 ref : `DatasetRef` 

1054 Ref associated with the newly-ingested dataset artifact. This 

1055 is used to determine the name within the datastore. 

1056 formatter : `Formatter` or Formatter class. 

1057 Formatter to use for validation. Can be a class or an instance. 

1058 

1059 Returns 

1060 ------- 

1061 location : `Location` 

1062 Target location for the newly-ingested dataset. 

1063 """ 

1064 # Ingesting a file from outside the datastore. 

1065 # This involves a new name. 

1066 template = self.templates.getTemplate(ref) 

1067 location = self.locationFactory.fromPath(template.format(ref)) 

1068 

1069 # Get the extension 

1070 ext = srcUri.getExtension() 

1071 

1072 # Update the destination to include that extension 

1073 location.updateExtension(ext) 

1074 

1075 # Ask the formatter to validate this extension 

1076 formatter.validateExtension(location) 

1077 

1078 return location 

1079 

1080 def _write_in_memory_to_artifact(self, inMemoryDataset: Any, ref: DatasetRef) -> StoredFileInfo: 

1081 """Write out in memory dataset to datastore. 

1082 

1083 Parameters 

1084 ---------- 

1085 inMemoryDataset : `object` 

1086 Dataset to write to datastore. 

1087 ref : `DatasetRef` 

1088 Registry information associated with this dataset. 

1089 

1090 Returns 

1091 ------- 

1092 info : `StoredFileInfo` 

1093 Information describing the artifact written to the datastore. 

1094 """ 

1095 # May need to coerce the in memory dataset to the correct 

1096 # python type. 

1097 inMemoryDataset = ref.datasetType.storageClass.coerce_type(inMemoryDataset) 

1098 

1099 location, formatter = self._prepare_for_put(inMemoryDataset, ref) 

1100 uri = location.uri 

1101 

1102 if not uri.dirname().exists(): 

1103 log.debug("Folder %s does not exist yet so creating it.", uri.dirname()) 

1104 uri.dirname().mkdir() 

1105 

1106 if self._transaction is None: 1106 ↛ 1107line 1106 didn't jump to line 1107, because the condition on line 1106 was never true

1107 raise RuntimeError("Attempting to write artifact without transaction enabled") 

1108 

1109 def _removeFileExists(uri: ResourcePath) -> None: 

1110 """Remove a file and do not complain if it is not there. 

1111 

1112 This is important since a formatter might fail before the file 

1113 is written and we should not confuse people by writing spurious 

1114 error messages to the log. 

1115 """ 

1116 try: 

1117 uri.remove() 

1118 except FileNotFoundError: 

1119 pass 

1120 

1121 # Register a callback to try to delete the uploaded data if 

1122 # something fails below 

1123 self._transaction.registerUndo("artifactWrite", _removeFileExists, uri) 

1124 

1125 data_written = False 

1126 if not uri.isLocal: 

1127 # This is a remote URI. Some datasets can be serialized directly 

1128 # to bytes and sent to the remote datastore without writing a 

1129 # file. If the dataset is intended to be saved to the cache 

1130 # a file is always written and direct write to the remote 

1131 # datastore is bypassed. 

1132 if not self.cacheManager.should_be_cached(ref): 

1133 try: 

1134 serializedDataset = formatter.toBytes(inMemoryDataset) 

1135 except NotImplementedError: 

1136 # Fallback to the file writing option. 

1137 pass 

1138 except Exception as e: 

1139 raise RuntimeError( 

1140 f"Failed to serialize dataset {ref} of type {type(inMemoryDataset)} to bytes." 

1141 ) from e 

1142 else: 

1143 log.debug("Writing bytes directly to %s", uri) 

1144 uri.write(serializedDataset, overwrite=True) 

1145 log.debug("Successfully wrote bytes directly to %s", uri) 

1146 data_written = True 

1147 

1148 if not data_written: 

1149 # Did not write the bytes directly to object store so instead 

1150 # write to temporary file. Always write to a temporary even if 

1151 # using a local file system -- that gives us atomic writes. 

1152 # If a process is killed as the file is being written we do not 

1153 # want it to remain in the correct place but in corrupt state. 

1154 # For local files write to the output directory not temporary dir. 

1155 prefix = uri.dirname() if uri.isLocal else None 

1156 with ResourcePath.temporary_uri(suffix=uri.getExtension(), prefix=prefix) as temporary_uri: 

1157 # Need to configure the formatter to write to a different 

1158 # location and that needs us to overwrite internals 

1159 log.debug("Writing dataset to temporary location at %s", temporary_uri) 

1160 with formatter._updateLocation(Location(None, temporary_uri)): 

1161 try: 

1162 formatter.write(inMemoryDataset) 

1163 except Exception as e: 

1164 raise RuntimeError( 

1165 f"Failed to serialize dataset {ref} of type" 

1166 f" {type(inMemoryDataset)} to " 

1167 f"temporary location {temporary_uri}" 

1168 ) from e 

1169 

1170 # Use move for a local file since that becomes an efficient 

1171 # os.rename. For remote resources we use copy to allow the 

1172 # file to be cached afterwards. 

1173 transfer = "move" if uri.isLocal else "copy" 

1174 

1175 uri.transfer_from(temporary_uri, transfer=transfer, overwrite=True) 

1176 

1177 if transfer == "copy": 

1178 # Cache if required 

1179 self.cacheManager.move_to_cache(temporary_uri, ref) 

1180 

1181 log.debug("Successfully wrote dataset to %s via a temporary file.", uri) 

1182 

1183 # URI is needed to resolve what ingest case are we dealing with 

1184 return self._extractIngestInfo(uri, ref, formatter=formatter) 

1185 

1186 def _read_artifact_into_memory( 

1187 self, 

1188 getInfo: DatastoreFileGetInformation, 

1189 ref: DatasetRef, 

1190 isComponent: bool = False, 

1191 cache_ref: Optional[DatasetRef] = None, 

1192 ) -> Any: 

1193 """Read the artifact from datastore into in memory object. 

1194 

1195 Parameters 

1196 ---------- 

1197 getInfo : `DatastoreFileGetInformation` 

1198 Information about the artifact within the datastore. 

1199 ref : `DatasetRef` 

1200 The registry information associated with this artifact. 

1201 isComponent : `bool` 

1202 Flag to indicate if a component is being read from this artifact. 

1203 cache_ref : `DatasetRef`, optional 

1204 The DatasetRef to use when looking up the file in the cache. 

1205 This ref must have the same ID as the supplied ref but can 

1206 be a parent ref or component ref to indicate to the cache whether 

1207 a composite file is being requested from the cache or a component 

1208 file. Without this the cache will default to the supplied ref but 

1209 it can get confused with read-only derived components for 

1210 disassembled composites. 

1211 

1212 Returns 

1213 ------- 

1214 inMemoryDataset : `object` 

1215 The artifact as a python object. 

1216 """ 

1217 location = getInfo.location 

1218 uri = location.uri 

1219 log.debug("Accessing data from %s", uri) 

1220 

1221 if cache_ref is None: 

1222 cache_ref = ref 

1223 if cache_ref.id != ref.id: 1223 ↛ 1224line 1223 didn't jump to line 1224, because the condition on line 1223 was never true

1224 raise ValueError( 

1225 "The supplied cache dataset ref refers to a different dataset than expected:" 

1226 f" {ref.id} != {cache_ref.id}" 

1227 ) 

1228 

1229 # Cannot recalculate checksum but can compare size as a quick check 

1230 # Do not do this if the size is negative since that indicates 

1231 # we do not know. 

1232 recorded_size = getInfo.info.file_size 

1233 resource_size = uri.size() 

1234 if recorded_size >= 0 and resource_size != recorded_size: 1234 ↛ 1235line 1234 didn't jump to line 1235, because the condition on line 1234 was never true

1235 raise RuntimeError( 

1236 "Integrity failure in Datastore. " 

1237 f"Size of file {uri} ({resource_size}) " 

1238 f"does not match size recorded in registry of {recorded_size}" 

1239 ) 

1240 

1241 # For the general case we have choices for how to proceed. 

1242 # 1. Always use a local file (downloading the remote resource to a 

1243 # temporary file if needed). 

1244 # 2. Use a threshold size and read into memory and use bytes. 

1245 # Use both for now with an arbitrary hand off size. 

1246 # This allows small datasets to be downloaded from remote object 

1247 # stores without requiring a temporary file. 

1248 

1249 formatter = getInfo.formatter 

1250 nbytes_max = 10_000_000 # Arbitrary number that we can tune 

1251 if resource_size <= nbytes_max and formatter.can_read_bytes(): 

1252 with self.cacheManager.find_in_cache(cache_ref, uri.getExtension()) as cached_file: 

1253 if cached_file is not None: 

1254 desired_uri = cached_file 

1255 msg = f" (cached version of {uri})" 

1256 else: 

1257 desired_uri = uri 

1258 msg = "" 

1259 with time_this(log, msg="Reading bytes from %s%s", args=(desired_uri, msg)): 

1260 serializedDataset = desired_uri.read() 

1261 log.debug( 

1262 "Deserializing %s from %d bytes from location %s with formatter %s", 

1263 f"component {getInfo.component}" if isComponent else "", 

1264 len(serializedDataset), 

1265 uri, 

1266 formatter.name(), 

1267 ) 

1268 try: 

1269 result = formatter.fromBytes( 

1270 serializedDataset, component=getInfo.component if isComponent else None 

1271 ) 

1272 except Exception as e: 

1273 raise ValueError( 

1274 f"Failure from formatter '{formatter.name()}' for dataset {ref.id}" 

1275 f" ({ref.datasetType.name} from {uri}): {e}" 

1276 ) from e 

1277 else: 

1278 # Read from file. 

1279 

1280 # Have to update the Location associated with the formatter 

1281 # because formatter.read does not allow an override. 

1282 # This could be improved. 

1283 location_updated = False 

1284 msg = "" 

1285 

1286 # First check in cache for local version. 

1287 # The cache will only be relevant for remote resources but 

1288 # no harm in always asking. Context manager ensures that cache 

1289 # file is not deleted during cache expiration. 

1290 with self.cacheManager.find_in_cache(cache_ref, uri.getExtension()) as cached_file: 

1291 if cached_file is not None: 

1292 msg = f"(via cache read of remote file {uri})" 

1293 uri = cached_file 

1294 location_updated = True 

1295 

1296 with uri.as_local() as local_uri: 

1297 

1298 can_be_cached = False 

1299 if uri != local_uri: 1299 ↛ 1301line 1299 didn't jump to line 1301, because the condition on line 1299 was never true

1300 # URI was remote and file was downloaded 

1301 cache_msg = "" 

1302 location_updated = True 

1303 

1304 if self.cacheManager.should_be_cached(cache_ref): 

1305 # In this scenario we want to ask if the downloaded 

1306 # file should be cached but we should not cache 

1307 # it until after we've used it (to ensure it can't 

1308 # be expired whilst we are using it). 

1309 can_be_cached = True 

1310 

1311 # Say that it is "likely" to be cached because 

1312 # if the formatter read fails we will not be 

1313 # caching this file. 

1314 cache_msg = " and likely cached" 

1315 

1316 msg = f"(via download to local file{cache_msg})" 

1317 

1318 # Calculate the (possibly) new location for the formatter 

1319 # to use. 

1320 newLocation = Location(*local_uri.split()) if location_updated else None 

1321 

1322 log.debug( 

1323 "Reading%s from location %s %s with formatter %s", 

1324 f" component {getInfo.component}" if isComponent else "", 

1325 uri, 

1326 msg, 

1327 formatter.name(), 

1328 ) 

1329 try: 

1330 with formatter._updateLocation(newLocation): 

1331 with time_this( 

1332 log, 

1333 msg="Reading%s from location %s %s with formatter %s", 

1334 args=( 

1335 f" component {getInfo.component}" if isComponent else "", 

1336 uri, 

1337 msg, 

1338 formatter.name(), 

1339 ), 

1340 ): 

1341 result = formatter.read(component=getInfo.component if isComponent else None) 

1342 except Exception as e: 

1343 raise ValueError( 

1344 f"Failure from formatter '{formatter.name()}' for dataset {ref.id}" 

1345 f" ({ref.datasetType.name} from {uri}): {e}" 

1346 ) from e 

1347 

1348 # File was read successfully so can move to cache 

1349 if can_be_cached: 1349 ↛ 1350line 1349 didn't jump to line 1350, because the condition on line 1349 was never true

1350 self.cacheManager.move_to_cache(local_uri, cache_ref) 

1351 

1352 return self._post_process_get( 

1353 result, getInfo.readStorageClass, getInfo.assemblerParams, isComponent=isComponent 

1354 ) 

1355 

1356 def knows(self, ref: DatasetRef) -> bool: 

1357 """Check if the dataset is known to the datastore. 

1358 

1359 Does not check for existence of any artifact. 

1360 

1361 Parameters 

1362 ---------- 

1363 ref : `DatasetRef` 

1364 Reference to the required dataset. 

1365 

1366 Returns 

1367 ------- 

1368 exists : `bool` 

1369 `True` if the dataset is known to the datastore. 

1370 """ 

1371 fileLocations = self._get_dataset_locations_info(ref) 

1372 if fileLocations: 

1373 return True 

1374 return False 

1375 

1376 def _process_mexists_records( 

1377 self, 

1378 id_to_ref: Dict[DatasetId, DatasetRef], 

1379 records: Dict[DatasetId, List[StoredFileInfo]], 

1380 all_required: bool, 

1381 artifact_existence: Optional[Dict[ResourcePath, bool]] = None, 

1382 ) -> Dict[DatasetRef, bool]: 

1383 """Helper function for mexists that checks the given records. 

1384 

1385 Parameters 

1386 ---------- 

1387 id_to_ref : `dict` of [`DatasetId`, `DatasetRef`] 

1388 Mapping of the dataset ID to the dataset ref itself. 

1389 records : `dict` of [`DatasetId`, `list` of `StoredFileInfo`] 

1390 Records as generally returned by 

1391 ``_get_stored_records_associated_with_refs``. 

1392 all_required : `bool` 

1393 Flag to indicate whether existence requires all artifacts 

1394 associated with a dataset ID to exist or not for existence. 

1395 artifact_existence : `dict` [`lsst.resources.ResourcePath`, `bool`] 

1396 Optional mapping of datastore artifact to existence. Updated by 

1397 this method with details of all artifacts tested. Can be `None` 

1398 if the caller is not interested. 

1399 

1400 Returns 

1401 ------- 

1402 existence : `dict` of [`DatasetRef`, `bool`] 

1403 Mapping from dataset to boolean indicating existence. 

1404 """ 

1405 # The URIs to be checked and a mapping of those URIs to 

1406 # the dataset ID. 

1407 uris_to_check: List[ResourcePath] = [] 

1408 location_map: Dict[ResourcePath, DatasetId] = {} 

1409 

1410 location_factory = self.locationFactory 

1411 

1412 uri_existence: Dict[ResourcePath, bool] = {} 

1413 for ref_id, infos in records.items(): 

1414 # Key is the dataset Id, value is list of StoredItemInfo 

1415 uris = [info.file_location(location_factory).uri for info in infos] 

1416 location_map.update({uri: ref_id for uri in uris}) 

1417 

1418 # Check the local cache directly for a dataset corresponding 

1419 # to the remote URI. 

1420 if self.cacheManager.file_count > 0: 

1421 ref = id_to_ref[ref_id] 

1422 for uri, storedFileInfo in zip(uris, infos): 

1423 check_ref = ref 

1424 if not ref.datasetType.isComponent() and (component := storedFileInfo.component): 1424 ↛ 1425line 1424 didn't jump to line 1425, because the condition on line 1424 was never true

1425 check_ref = ref.makeComponentRef(component) 

1426 if self.cacheManager.known_to_cache(check_ref, uri.getExtension()): 

1427 # Proxy for URI existence. 

1428 uri_existence[uri] = True 

1429 else: 

1430 uris_to_check.append(uri) 

1431 else: 

1432 # Check all of them. 

1433 uris_to_check.extend(uris) 

1434 

1435 if artifact_existence is not None: 

1436 # If a URI has already been checked remove it from the list 

1437 # and immediately add the status to the output dict. 

1438 filtered_uris_to_check = [] 

1439 for uri in uris_to_check: 

1440 if uri in artifact_existence: 

1441 uri_existence[uri] = artifact_existence[uri] 

1442 else: 

1443 filtered_uris_to_check.append(uri) 

1444 uris_to_check = filtered_uris_to_check 

1445 

1446 # Results. 

1447 dataset_existence: Dict[DatasetRef, bool] = {} 

1448 

1449 uri_existence.update(ResourcePath.mexists(uris_to_check)) 

1450 for uri, exists in uri_existence.items(): 

1451 dataset_id = location_map[uri] 

1452 ref = id_to_ref[dataset_id] 

1453 

1454 # Disassembled composite needs to check all locations. 

1455 # all_required indicates whether all need to exist or not. 

1456 if ref in dataset_existence: 

1457 if all_required: 

1458 exists = dataset_existence[ref] and exists 

1459 else: 

1460 exists = dataset_existence[ref] or exists 

1461 dataset_existence[ref] = exists 

1462 

1463 if artifact_existence is not None: 

1464 artifact_existence.update(uri_existence) 

1465 

1466 return dataset_existence 

1467 

1468 def mexists( 

1469 self, refs: Iterable[DatasetRef], artifact_existence: Optional[Dict[ResourcePath, bool]] = None 

1470 ) -> Dict[DatasetRef, bool]: 

1471 """Check the existence of multiple datasets at once. 

1472 

1473 Parameters 

1474 ---------- 

1475 refs : iterable of `DatasetRef` 

1476 The datasets to be checked. 

1477 artifact_existence : `dict` [`lsst.resources.ResourcePath`, `bool`] 

1478 Optional mapping of datastore artifact to existence. Updated by 

1479 this method with details of all artifacts tested. Can be `None` 

1480 if the caller is not interested. 

1481 

1482 Returns 

1483 ------- 

1484 existence : `dict` of [`DatasetRef`, `bool`] 

1485 Mapping from dataset to boolean indicating existence. 

1486 

1487 Notes 

1488 ----- 

1489 To minimize potentially costly remote existence checks, the local 

1490 cache is checked as a proxy for existence. If a file for this 

1491 `DatasetRef` does exist no check is done for the actual URI. This 

1492 could result in possibly unexpected behavior if the dataset itself 

1493 has been removed from the datastore by another process whilst it is 

1494 still in the cache. 

1495 """ 

1496 chunk_size = 10_000 

1497 dataset_existence: Dict[DatasetRef, bool] = {} 

1498 log.debug("Checking for the existence of multiple artifacts in datastore in chunks of %d", chunk_size) 

1499 n_found_total = 0 

1500 n_checked = 0 

1501 n_chunks = 0 

1502 for chunk in chunk_iterable(refs, chunk_size=chunk_size): 

1503 chunk_result = self._mexists(chunk, artifact_existence) 

1504 if log.isEnabledFor(VERBOSE): 

1505 n_results = len(chunk_result) 

1506 n_checked += n_results 

1507 # Can treat the booleans as 0, 1 integers and sum them. 

1508 n_found = sum(chunk_result.values()) 

1509 n_found_total += n_found 

1510 log.verbose( 

1511 "Number of datasets found in datastore for chunk %d = %d/%d (running total: %d/%d)", 

1512 n_chunks, 

1513 n_found, 

1514 n_results, 

1515 n_found_total, 

1516 n_checked, 

1517 ) 

1518 dataset_existence.update(chunk_result) 

1519 n_chunks += 1 

1520 

1521 return dataset_existence 

1522 

1523 def _mexists( 

1524 self, refs: Iterable[DatasetRef], artifact_existence: Optional[Dict[ResourcePath, bool]] = None 

1525 ) -> Dict[DatasetRef, bool]: 

1526 """Check the existence of multiple datasets at once. 

1527 

1528 Parameters 

1529 ---------- 

1530 refs : iterable of `DatasetRef` 

1531 The datasets to be checked. 

1532 

1533 Returns 

1534 ------- 

1535 existence : `dict` of [`DatasetRef`, `bool`] 

1536 Mapping from dataset to boolean indicating existence. 

1537 """ 

1538 # Need a mapping of dataset_id to dataset ref since the API 

1539 # works with dataset_id 

1540 id_to_ref = {ref.getCheckedId(): ref for ref in refs} 

1541 

1542 # Set of all IDs we are checking for. 

1543 requested_ids = set(id_to_ref.keys()) 

1544 

1545 # The records themselves. Could be missing some entries. 

1546 records = self._get_stored_records_associated_with_refs(refs) 

1547 

1548 dataset_existence = self._process_mexists_records( 

1549 id_to_ref, records, True, artifact_existence=artifact_existence 

1550 ) 

1551 

1552 # Set of IDs that have been handled. 

1553 handled_ids = {ref.id for ref in dataset_existence.keys()} 

1554 

1555 missing_ids = requested_ids - handled_ids 

1556 if missing_ids: 

1557 if not self.trustGetRequest: 

1558 # Must assume these do not exist 

1559 for missing in missing_ids: 

1560 dataset_existence[id_to_ref[missing]] = False 

1561 else: 

1562 log.debug( 

1563 "%d out of %d datasets were not known to datastore during initial existence check.", 

1564 len(missing_ids), 

1565 len(requested_ids), 

1566 ) 

1567 

1568 # Construct data structure identical to that returned 

1569 # by _get_stored_records_associated_with_refs() but using 

1570 # guessed names. 

1571 records = {} 

1572 for missing in missing_ids: 

1573 expected = self._get_expected_dataset_locations_info(id_to_ref[missing]) 

1574 records[missing] = [info for _, info in expected] 

1575 

1576 dataset_existence.update( 

1577 self._process_mexists_records( 

1578 id_to_ref, records, False, artifact_existence=artifact_existence 

1579 ) 

1580 ) 

1581 

1582 return dataset_existence 

1583 

1584 def exists(self, ref: DatasetRef) -> bool: 

1585 """Check if the dataset exists in the datastore. 

1586 

1587 Parameters 

1588 ---------- 

1589 ref : `DatasetRef` 

1590 Reference to the required dataset. 

1591 

1592 Returns 

1593 ------- 

1594 exists : `bool` 

1595 `True` if the entity exists in the `Datastore`. 

1596 

1597 Notes 

1598 ----- 

1599 The local cache is checked as a proxy for existence in the remote 

1600 object store. It is possible that another process on a different 

1601 compute node could remove the file from the object store even 

1602 though it is present in the local cache. 

1603 """ 

1604 fileLocations = self._get_dataset_locations_info(ref) 

1605 

1606 # if we are being asked to trust that registry might not be correct 

1607 # we ask for the expected locations and check them explicitly 

1608 if not fileLocations: 

1609 if not self.trustGetRequest: 

1610 return False 

1611 

1612 # First check the cache. If it is not found we must check 

1613 # the datastore itself. Assume that any component in the cache 

1614 # means that the dataset does exist somewhere. 

1615 if self.cacheManager.known_to_cache(ref): 1615 ↛ 1616line 1615 didn't jump to line 1616, because the condition on line 1615 was never true

1616 return True 

1617 

1618 # When we are guessing a dataset location we can not check 

1619 # for the existence of every component since we can not 

1620 # know if every component was written. Instead we check 

1621 # for the existence of any of the expected locations. 

1622 for location, _ in self._get_expected_dataset_locations_info(ref): 

1623 if self._artifact_exists(location): 

1624 return True 

1625 return False 

1626 

1627 # All listed artifacts must exist. 

1628 for location, storedFileInfo in fileLocations: 

1629 # Checking in cache needs the component ref. 

1630 check_ref = ref 

1631 if not ref.datasetType.isComponent() and (component := storedFileInfo.component): 

1632 check_ref = ref.makeComponentRef(component) 

1633 if self.cacheManager.known_to_cache(check_ref, location.getExtension()): 

1634 continue 

1635 

1636 if not self._artifact_exists(location): 

1637 return False 

1638 

1639 return True 

1640 

1641 def getURIs(self, ref: DatasetRef, predict: bool = False) -> DatasetRefURIs: 

1642 """Return URIs associated with dataset. 

1643 

1644 Parameters 

1645 ---------- 

1646 ref : `DatasetRef` 

1647 Reference to the required dataset. 

1648 predict : `bool`, optional 

1649 If the datastore does not know about the dataset, should it 

1650 return a predicted URI or not? 

1651 

1652 Returns 

1653 ------- 

1654 uris : `DatasetRefURIs` 

1655 The URI to the primary artifact associated with this dataset (if 

1656 the dataset was disassembled within the datastore this may be 

1657 `None`), and the URIs to any components associated with the dataset 

1658 artifact. (can be empty if there are no components). 

1659 """ 

1660 # if this has never been written then we have to guess 

1661 if not self.exists(ref): 

1662 if not predict: 

1663 raise FileNotFoundError("Dataset {} not in this datastore".format(ref)) 

1664 

1665 return self._predict_URIs(ref) 

1666 

1667 # If this is a ref that we have written we can get the path. 

1668 # Get file metadata and internal metadata 

1669 fileLocations = self._get_dataset_locations_info(ref) 

1670 

1671 return self._locations_to_URI(ref, fileLocations) 

1672 

1673 def getURI(self, ref: DatasetRef, predict: bool = False) -> ResourcePath: 

1674 """URI to the Dataset. 

1675 

1676 Parameters 

1677 ---------- 

1678 ref : `DatasetRef` 

1679 Reference to the required Dataset. 

1680 predict : `bool` 

1681 If `True`, allow URIs to be returned of datasets that have not 

1682 been written. 

1683 

1684 Returns 

1685 ------- 

1686 uri : `str` 

1687 URI pointing to the dataset within the datastore. If the 

1688 dataset does not exist in the datastore, and if ``predict`` is 

1689 `True`, the URI will be a prediction and will include a URI 

1690 fragment "#predicted". 

1691 If the datastore does not have entities that relate well 

1692 to the concept of a URI the returned URI will be 

1693 descriptive. The returned URI is not guaranteed to be obtainable. 

1694 

1695 Raises 

1696 ------ 

1697 FileNotFoundError 

1698 Raised if a URI has been requested for a dataset that does not 

1699 exist and guessing is not allowed. 

1700 RuntimeError 

1701 Raised if a request is made for a single URI but multiple URIs 

1702 are associated with this dataset. 

1703 

1704 Notes 

1705 ----- 

1706 When a predicted URI is requested an attempt will be made to form 

1707 a reasonable URI based on file templates and the expected formatter. 

1708 """ 

1709 primary, components = self.getURIs(ref, predict) 

1710 if primary is None or components: 1710 ↛ 1711line 1710 didn't jump to line 1711, because the condition on line 1710 was never true

1711 raise RuntimeError( 

1712 f"Dataset ({ref}) includes distinct URIs for components. Use Datastore.getURIs() instead." 

1713 ) 

1714 return primary 

1715 

1716 def _predict_URIs( 

1717 self, 

1718 ref: DatasetRef, 

1719 ) -> DatasetRefURIs: 

1720 """Predict the URIs of a dataset ref. 

1721 

1722 Parameters 

1723 ---------- 

1724 ref : `DatasetRef` 

1725 Reference to the required Dataset. 

1726 

1727 Returns 

1728 ------- 

1729 URI : DatasetRefUris 

1730 Primary and component URIs. URIs will contain a URI fragment 

1731 "#predicted". 

1732 """ 

1733 uris = DatasetRefURIs() 

1734 

1735 if self.composites.shouldBeDisassembled(ref): 

1736 

1737 for component, _ in ref.datasetType.storageClass.components.items(): 

1738 comp_ref = ref.makeComponentRef(component) 

1739 comp_location, _ = self._determine_put_formatter_location(comp_ref) 

1740 

1741 # Add the "#predicted" URI fragment to indicate this is a 

1742 # guess 

1743 uris.componentURIs[component] = ResourcePath(comp_location.uri.geturl() + "#predicted") 

1744 

1745 else: 

1746 

1747 location, _ = self._determine_put_formatter_location(ref) 

1748 

1749 # Add the "#predicted" URI fragment to indicate this is a guess 

1750 uris.primaryURI = ResourcePath(location.uri.geturl() + "#predicted") 

1751 

1752 return uris 

1753 

1754 def getManyURIs( 

1755 self, 

1756 refs: Iterable[DatasetRef], 

1757 predict: bool = False, 

1758 allow_missing: bool = False, 

1759 ) -> Dict[DatasetRef, DatasetRefURIs]: 

1760 # Docstring inherited 

1761 

1762 uris: Dict[DatasetRef, DatasetRefURIs] = {} 

1763 

1764 records = self._get_stored_records_associated_with_refs(refs) 

1765 records_keys = records.keys() 

1766 

1767 existing_refs = (ref for ref in refs if ref.id in records_keys) 

1768 missing_refs = (ref for ref in refs if ref.id not in records_keys) 

1769 

1770 for ref in missing_refs: 

1771 

1772 # if this has never been written then we have to guess 

1773 if not predict: 

1774 if not allow_missing: 

1775 raise FileNotFoundError("Dataset {} not in this datastore.".format(ref)) 

1776 else: 

1777 uris[ref] = self._predict_URIs(ref) 

1778 

1779 for ref in existing_refs: 

1780 file_infos = records[ref.getCheckedId()] 

1781 file_locations = [(i.file_location(self.locationFactory), i) for i in file_infos] 

1782 uris[ref] = self._locations_to_URI(ref, file_locations) 

1783 

1784 return uris 

1785 

1786 def _locations_to_URI( 

1787 self, 

1788 ref: DatasetRef, 

1789 file_locations: Sequence[Tuple[Location, StoredFileInfo]], 

1790 ) -> DatasetRefURIs: 

1791 """Convert one or more file locations associated with a DatasetRef 

1792 to a DatasetRefURIs. 

1793 

1794 Parameters 

1795 ---------- 

1796 ref : `DatasetRef` 

1797 Reference to the dataset. 

1798 file_locations : Sequence[Tuple[Location, StoredFileInfo]] 

1799 Each item in the sequence is the location of the dataset within the 

1800 datastore and stored information about the file and its formatter. 

1801 If there is only one item in the sequence then it is treated as the 

1802 primary URI. If there is more than one item then they are treated 

1803 as component URIs. If there are no items then an error is raised 

1804 unless ``self.trustGetRequest`` is `True`. 

1805 

1806 Returns 

1807 ------- 

1808 uris: DatasetRefURIs 

1809 Represents the primary URI or component URIs described by the 

1810 inputs. 

1811 

1812 Raises 

1813 ------ 

1814 RuntimeError 

1815 If no file locations are passed in and ``self.trustGetRequest`` is 

1816 `False`. 

1817 FileNotFoundError 

1818 If the a passed-in URI does not exist, and ``self.trustGetRequest`` 

1819 is `False`. 

1820 RuntimeError 

1821 If a passed in `StoredFileInfo`'s ``component`` is `None` (this is 

1822 unexpected). 

1823 """ 

1824 

1825 guessing = False 

1826 uris = DatasetRefURIs() 

1827 

1828 if not file_locations: 

1829 if not self.trustGetRequest: 1829 ↛ 1830line 1829 didn't jump to line 1830, because the condition on line 1829 was never true

1830 raise RuntimeError(f"Unexpectedly got no artifacts for dataset {ref}") 

1831 file_locations = self._get_expected_dataset_locations_info(ref) 

1832 guessing = True 

1833 

1834 if len(file_locations) == 1: 

1835 # No disassembly so this is the primary URI 

1836 uris.primaryURI = file_locations[0][0].uri 

1837 if guessing and not uris.primaryURI.exists(): 1837 ↛ 1838line 1837 didn't jump to line 1838, because the condition on line 1837 was never true

1838 raise FileNotFoundError(f"Expected URI ({uris.primaryURI}) does not exist") 

1839 else: 

1840 for location, file_info in file_locations: 

1841 if file_info.component is None: 1841 ↛ 1842line 1841 didn't jump to line 1842, because the condition on line 1841 was never true

1842 raise RuntimeError(f"Unexpectedly got no component name for a component at {location}") 

1843 if guessing and not location.uri.exists(): 1843 ↛ 1847line 1843 didn't jump to line 1847, because the condition on line 1843 was never true

1844 # If we are trusting then it is entirely possible for 

1845 # some components to be missing. In that case we skip 

1846 # to the next component. 

1847 if self.trustGetRequest: 

1848 continue 

1849 raise FileNotFoundError(f"Expected URI ({location.uri}) does not exist") 

1850 uris.componentURIs[file_info.component] = location.uri 

1851 

1852 return uris 

1853 

1854 def retrieveArtifacts( 

1855 self, 

1856 refs: Iterable[DatasetRef], 

1857 destination: ResourcePath, 

1858 transfer: str = "auto", 

1859 preserve_path: bool = True, 

1860 overwrite: bool = False, 

1861 ) -> List[ResourcePath]: 

1862 """Retrieve the file artifacts associated with the supplied refs. 

1863 

1864 Parameters 

1865 ---------- 

1866 refs : iterable of `DatasetRef` 

1867 The datasets for which file artifacts are to be retrieved. 

1868 A single ref can result in multiple files. The refs must 

1869 be resolved. 

1870 destination : `lsst.resources.ResourcePath` 

1871 Location to write the file artifacts. 

1872 transfer : `str`, optional 

1873 Method to use to transfer the artifacts. Must be one of the options 

1874 supported by `lsst.resources.ResourcePath.transfer_from()`. 

1875 "move" is not allowed. 

1876 preserve_path : `bool`, optional 

1877 If `True` the full path of the file artifact within the datastore 

1878 is preserved. If `False` the final file component of the path 

1879 is used. 

1880 overwrite : `bool`, optional 

1881 If `True` allow transfers to overwrite existing files at the 

1882 destination. 

1883 

1884 Returns 

1885 ------- 

1886 targets : `list` of `lsst.resources.ResourcePath` 

1887 URIs of file artifacts in destination location. Order is not 

1888 preserved. 

1889 """ 

1890 if not destination.isdir(): 1890 ↛ 1891line 1890 didn't jump to line 1891, because the condition on line 1890 was never true

1891 raise ValueError(f"Destination location must refer to a directory. Given {destination}") 

1892 

1893 if transfer == "move": 

1894 raise ValueError("Can not move artifacts out of datastore. Use copy instead.") 

1895 

1896 # Source -> Destination 

1897 # This also helps filter out duplicate DatasetRef in the request 

1898 # that will map to the same underlying file transfer. 

1899 to_transfer: Dict[ResourcePath, ResourcePath] = {} 

1900 

1901 for ref in refs: 

1902 locations = self._get_dataset_locations_info(ref) 

1903 for location, _ in locations: 

1904 source_uri = location.uri 

1905 target_path: ResourcePathExpression 

1906 if preserve_path: 

1907 target_path = location.pathInStore 

1908 if target_path.isabs(): 1908 ↛ 1911line 1908 didn't jump to line 1911, because the condition on line 1908 was never true

1909 # This is an absolute path to an external file. 

1910 # Use the full path. 

1911 target_path = target_path.relativeToPathRoot 

1912 else: 

1913 target_path = source_uri.basename() 

1914 target_uri = destination.join(target_path) 

1915 to_transfer[source_uri] = target_uri 

1916 

1917 # In theory can now parallelize the transfer 

1918 log.debug("Number of artifacts to transfer to %s: %d", str(destination), len(to_transfer)) 

1919 for source_uri, target_uri in to_transfer.items(): 

1920 target_uri.transfer_from(source_uri, transfer=transfer, overwrite=overwrite) 

1921 

1922 return list(to_transfer.values()) 

1923 

1924 def get(self, ref: DatasetRef, parameters: Optional[Mapping[str, Any]] = None) -> Any: 

1925 """Load an InMemoryDataset from the store. 

1926 

1927 Parameters 

1928 ---------- 

1929 ref : `DatasetRef` 

1930 Reference to the required Dataset. 

1931 parameters : `dict` 

1932 `StorageClass`-specific parameters that specify, for example, 

1933 a slice of the dataset to be loaded. 

1934 

1935 Returns 

1936 ------- 

1937 inMemoryDataset : `object` 

1938 Requested dataset or slice thereof as an InMemoryDataset. 

1939 

1940 Raises 

1941 ------ 

1942 FileNotFoundError 

1943 Requested dataset can not be retrieved. 

1944 TypeError 

1945 Return value from formatter has unexpected type. 

1946 ValueError 

1947 Formatter failed to process the dataset. 

1948 """ 

1949 allGetInfo = self._prepare_for_get(ref, parameters) 

1950 refComponent = ref.datasetType.component() 

1951 

1952 # Supplied storage class for the component being read 

1953 refStorageClass = ref.datasetType.storageClass 

1954 

1955 # Create mapping from component name to related info 

1956 allComponents = {i.component: i for i in allGetInfo} 

1957 

1958 # By definition the dataset is disassembled if we have more 

1959 # than one record for it. 

1960 isDisassembled = len(allGetInfo) > 1 

1961 

1962 # Look for the special case where we are disassembled but the 

1963 # component is a derived component that was not written during 

1964 # disassembly. For this scenario we need to check that the 

1965 # component requested is listed as a derived component for the 

1966 # composite storage class 

1967 isDisassembledReadOnlyComponent = False 

1968 if isDisassembled and refComponent: 

1969 # The composite storage class should be accessible through 

1970 # the component dataset type 

1971 compositeStorageClass = ref.datasetType.parentStorageClass 

1972 

1973 # In the unlikely scenario where the composite storage 

1974 # class is not known, we can only assume that this is a 

1975 # normal component. If that assumption is wrong then the 

1976 # branch below that reads a persisted component will fail 

1977 # so there is no need to complain here. 

1978 if compositeStorageClass is not None: 1978 ↛ 1981line 1978 didn't jump to line 1981, because the condition on line 1978 was never false

1979 isDisassembledReadOnlyComponent = refComponent in compositeStorageClass.derivedComponents 

1980 

1981 if isDisassembled and not refComponent: 

1982 # This was a disassembled dataset spread over multiple files 

1983 # and we need to put them all back together again. 

1984 # Read into memory and then assemble 

1985 

1986 # Check that the supplied parameters are suitable for the type read 

1987 refStorageClass.validateParameters(parameters) 

1988 

1989 # We want to keep track of all the parameters that were not used 

1990 # by formatters. We assume that if any of the component formatters 

1991 # use a parameter that we do not need to apply it again in the 

1992 # assembler. 

1993 usedParams = set() 

1994 

1995 components: Dict[str, Any] = {} 

1996 for getInfo in allGetInfo: 

1997 # assemblerParams are parameters not understood by the 

1998 # associated formatter. 

1999 usedParams.update(set(getInfo.formatterParams)) 

2000 

2001 component = getInfo.component 

2002 

2003 if component is None: 2003 ↛ 2004line 2003 didn't jump to line 2004, because the condition on line 2003 was never true

2004 raise RuntimeError(f"Internal error in datastore assembly of {ref}") 

2005 

2006 # We do not want the formatter to think it's reading 

2007 # a component though because it is really reading a 

2008 # standalone dataset -- always tell reader it is not a 

2009 # component. 

2010 components[component] = self._read_artifact_into_memory( 

2011 getInfo, ref.makeComponentRef(component), isComponent=False 

2012 ) 

2013 

2014 inMemoryDataset = ref.datasetType.storageClass.delegate().assemble(components) 

2015 

2016 # Any unused parameters will have to be passed to the assembler 

2017 if parameters: 

2018 unusedParams = {k: v for k, v in parameters.items() if k not in usedParams} 

2019 else: 

2020 unusedParams = {} 

2021 

2022 # Process parameters 

2023 return ref.datasetType.storageClass.delegate().handleParameters( 

2024 inMemoryDataset, parameters=unusedParams 

2025 ) 

2026 

2027 elif isDisassembledReadOnlyComponent: 

2028 

2029 compositeStorageClass = ref.datasetType.parentStorageClass 

2030 if compositeStorageClass is None: 2030 ↛ 2031line 2030 didn't jump to line 2031, because the condition on line 2030 was never true

2031 raise RuntimeError( 

2032 f"Unable to retrieve derived component '{refComponent}' since" 

2033 "no composite storage class is available." 

2034 ) 

2035 

2036 if refComponent is None: 2036 ↛ 2038line 2036 didn't jump to line 2038, because the condition on line 2036 was never true

2037 # Mainly for mypy 

2038 raise RuntimeError(f"Internal error in datastore {self.name}: component can not be None here") 

2039 

2040 # Assume that every derived component can be calculated by 

2041 # forwarding the request to a single read/write component. 

2042 # Rather than guessing which rw component is the right one by 

2043 # scanning each for a derived component of the same name, 

2044 # we ask the storage class delegate directly which one is best to 

2045 # use. 

2046 compositeDelegate = compositeStorageClass.delegate() 

2047 forwardedComponent = compositeDelegate.selectResponsibleComponent( 

2048 refComponent, set(allComponents) 

2049 ) 

2050 

2051 # Select the relevant component 

2052 rwInfo = allComponents[forwardedComponent] 

2053 

2054 # For now assume that read parameters are validated against 

2055 # the real component and not the requested component 

2056 forwardedStorageClass = rwInfo.formatter.fileDescriptor.readStorageClass 

2057 forwardedStorageClass.validateParameters(parameters) 

2058 

2059 # The reference to use for the caching must refer to the forwarded 

2060 # component and not the derived component. 

2061 cache_ref = ref.makeCompositeRef().makeComponentRef(forwardedComponent) 

2062 

2063 # Unfortunately the FileDescriptor inside the formatter will have 

2064 # the wrong write storage class so we need to create a new one 

2065 # given the immutability constraint. 

2066 writeStorageClass = rwInfo.info.storageClass 

2067 

2068 # We may need to put some thought into parameters for read 

2069 # components but for now forward them on as is 

2070 readFormatter = type(rwInfo.formatter)( 

2071 FileDescriptor( 

2072 rwInfo.location, 

2073 readStorageClass=refStorageClass, 

2074 storageClass=writeStorageClass, 

2075 parameters=parameters, 

2076 ), 

2077 ref.dataId, 

2078 ) 

2079 

2080 # The assembler can not receive any parameter requests for a 

2081 # derived component at this time since the assembler will 

2082 # see the storage class of the derived component and those 

2083 # parameters will have to be handled by the formatter on the 

2084 # forwarded storage class. 

2085 assemblerParams: Dict[str, Any] = {} 

2086 

2087 # Need to created a new info that specifies the derived 

2088 # component and associated storage class 

2089 readInfo = DatastoreFileGetInformation( 

2090 rwInfo.location, 

2091 readFormatter, 

2092 rwInfo.info, 

2093 assemblerParams, 

2094 {}, 

2095 refComponent, 

2096 refStorageClass, 

2097 ) 

2098 

2099 return self._read_artifact_into_memory(readInfo, ref, isComponent=True, cache_ref=cache_ref) 

2100 

2101 else: 

2102 # Single file request or component from that composite file 

2103 for lookup in (refComponent, None): 2103 ↛ 2108line 2103 didn't jump to line 2108, because the loop on line 2103 didn't complete

2104 if lookup in allComponents: 2104 ↛ 2103line 2104 didn't jump to line 2103, because the condition on line 2104 was never false

2105 getInfo = allComponents[lookup] 

2106 break 

2107 else: 

2108 raise FileNotFoundError( 

2109 f"Component {refComponent} not found for ref {ref} in datastore {self.name}" 

2110 ) 

2111 

2112 # Do not need the component itself if already disassembled 

2113 if isDisassembled: 

2114 isComponent = False 

2115 else: 

2116 isComponent = getInfo.component is not None 

2117 

2118 # For a component read of a composite we want the cache to 

2119 # be looking at the composite ref itself. 

2120 cache_ref = ref.makeCompositeRef() if isComponent else ref 

2121 

2122 # For a disassembled component we can validate parametersagainst 

2123 # the component storage class directly 

2124 if isDisassembled: 

2125 refStorageClass.validateParameters(parameters) 

2126 else: 

2127 # For an assembled composite this could be a derived 

2128 # component derived from a real component. The validity 

2129 # of the parameters is not clear. For now validate against 

2130 # the composite storage class 

2131 getInfo.formatter.fileDescriptor.storageClass.validateParameters(parameters) 

2132 

2133 return self._read_artifact_into_memory(getInfo, ref, isComponent=isComponent, cache_ref=cache_ref) 

2134 

2135 @transactional 

2136 def put(self, inMemoryDataset: Any, ref: DatasetRef) -> None: 

2137 """Write a InMemoryDataset with a given `DatasetRef` to the store. 

2138 

2139 Parameters 

2140 ---------- 

2141 inMemoryDataset : `object` 

2142 The dataset to store. 

2143 ref : `DatasetRef` 

2144 Reference to the associated Dataset. 

2145 

2146 Raises 

2147 ------ 

2148 TypeError 

2149 Supplied object and storage class are inconsistent. 

2150 DatasetTypeNotSupportedError 

2151 The associated `DatasetType` is not handled by this datastore. 

2152 

2153 Notes 

2154 ----- 

2155 If the datastore is configured to reject certain dataset types it 

2156 is possible that the put will fail and raise a 

2157 `DatasetTypeNotSupportedError`. The main use case for this is to 

2158 allow `ChainedDatastore` to put to multiple datastores without 

2159 requiring that every datastore accepts the dataset. 

2160 """ 

2161 

2162 doDisassembly = self.composites.shouldBeDisassembled(ref) 

2163 # doDisassembly = True 

2164 

2165 artifacts = [] 

2166 if doDisassembly: 

2167 components = ref.datasetType.storageClass.delegate().disassemble(inMemoryDataset) 

2168 if components is None: 2168 ↛ 2169line 2168 didn't jump to line 2169, because the condition on line 2168 was never true

2169 raise RuntimeError( 

2170 f"Inconsistent configuration: dataset type {ref.datasetType.name} " 

2171 f"with storage class {ref.datasetType.storageClass.name} " 

2172 "is configured to be disassembled, but cannot be." 

2173 ) 

2174 for component, componentInfo in components.items(): 

2175 # Don't recurse because we want to take advantage of 

2176 # bulk insert -- need a new DatasetRef that refers to the 

2177 # same dataset_id but has the component DatasetType 

2178 # DatasetType does not refer to the types of components 

2179 # So we construct one ourselves. 

2180 compRef = ref.makeComponentRef(component) 

2181 storedInfo = self._write_in_memory_to_artifact(componentInfo.component, compRef) 

2182 artifacts.append((compRef, storedInfo)) 

2183 else: 

2184 # Write the entire thing out 

2185 storedInfo = self._write_in_memory_to_artifact(inMemoryDataset, ref) 

2186 artifacts.append((ref, storedInfo)) 

2187 

2188 self._register_datasets(artifacts) 

2189 

2190 @transactional 

2191 def trash(self, ref: Union[DatasetRef, Iterable[DatasetRef]], ignore_errors: bool = True) -> None: 

2192 # At this point can safely remove these datasets from the cache 

2193 # to avoid confusion later on. If they are not trashed later 

2194 # the cache will simply be refilled. 

2195 self.cacheManager.remove_from_cache(ref) 

2196 

2197 # If we are in trust mode there will be nothing to move to 

2198 # the trash table and we will have to try to delete the file 

2199 # immediately. 

2200 if self.trustGetRequest: 

2201 # Try to keep the logic below for a single file trash. 

2202 if isinstance(ref, DatasetRef): 

2203 refs = {ref} 

2204 else: 

2205 # Will recreate ref at the end of this branch. 

2206 refs = set(ref) 

2207 

2208 # Determine which datasets are known to datastore directly. 

2209 id_to_ref = {ref.getCheckedId(): ref for ref in refs} 

2210 existing_ids = self._get_stored_records_associated_with_refs(refs) 

2211 existing_refs = {id_to_ref[ref_id] for ref_id in existing_ids} 

2212 

2213 missing = refs - existing_refs 

2214 if missing: 

2215 # Do an explicit existence check on these refs. 

2216 # We only care about the artifacts at this point and not 

2217 # the dataset existence. 

2218 artifact_existence: Dict[ResourcePath, bool] = {} 

2219 _ = self.mexists(missing, artifact_existence) 

2220 uris = [uri for uri, exists in artifact_existence.items() if exists] 

2221 

2222 # FUTURE UPGRADE: Implement a parallelized bulk remove. 

2223 log.debug("Removing %d artifacts from datastore that are unknown to datastore", len(uris)) 

2224 for uri in uris: 

2225 try: 

2226 uri.remove() 

2227 except Exception as e: 

2228 if ignore_errors: 

2229 log.debug("Artifact %s could not be removed: %s", uri, e) 

2230 continue 

2231 raise 

2232 

2233 # There is no point asking the code below to remove refs we 

2234 # know are missing so update it with the list of existing 

2235 # records. Try to retain one vs many logic. 

2236 if not existing_refs: 

2237 # Nothing more to do since none of the datasets were 

2238 # known to the datastore record table. 

2239 return 

2240 ref = list(existing_refs) 

2241 if len(ref) == 1: 

2242 ref = ref[0] 

2243 

2244 # Get file metadata and internal metadata 

2245 if not isinstance(ref, DatasetRef): 

2246 log.debug("Doing multi-dataset trash in datastore %s", self.name) 

2247 # Assumed to be an iterable of refs so bulk mode enabled. 

2248 try: 

2249 self.bridge.moveToTrash(ref) 

2250 except Exception as e: 

2251 if ignore_errors: 

2252 log.warning("Unexpected issue moving multiple datasets to trash: %s", e) 

2253 else: 

2254 raise 

2255 return 

2256 

2257 log.debug("Trashing dataset %s in datastore %s", ref, self.name) 

2258 

2259 fileLocations = self._get_dataset_locations_info(ref) 

2260 

2261 if not fileLocations: 

2262 err_msg = f"Requested dataset to trash ({ref}) is not known to datastore {self.name}" 

2263 if ignore_errors: 

2264 log.warning(err_msg) 

2265 return 

2266 else: 

2267 raise FileNotFoundError(err_msg) 

2268 

2269 for location, storedFileInfo in fileLocations: 

2270 if not self._artifact_exists(location): 2270 ↛ 2271line 2270 didn't jump to line 2271

2271 err_msg = ( 

2272 f"Dataset is known to datastore {self.name} but " 

2273 f"associated artifact ({location.uri}) is missing" 

2274 ) 

2275 if ignore_errors: 

2276 log.warning(err_msg) 

2277 return 

2278 else: 

2279 raise FileNotFoundError(err_msg) 

2280 

2281 # Mark dataset as trashed 

2282 try: 

2283 self.bridge.moveToTrash([ref]) 

2284 except Exception as e: 

2285 if ignore_errors: 

2286 log.warning( 

2287 "Attempted to mark dataset (%s) to be trashed in datastore %s " 

2288 "but encountered an error: %s", 

2289 ref, 

2290 self.name, 

2291 e, 

2292 ) 

2293 pass 

2294 else: 

2295 raise 

2296 

2297 @transactional 

2298 def emptyTrash(self, ignore_errors: bool = True) -> None: 

2299 """Remove all datasets from the trash. 

2300 

2301 Parameters 

2302 ---------- 

2303 ignore_errors : `bool` 

2304 If `True` return without error even if something went wrong. 

2305 Problems could occur if another process is simultaneously trying 

2306 to delete. 

2307 """ 

2308 log.debug("Emptying trash in datastore %s", self.name) 

2309 

2310 # Context manager will empty trash iff we finish it without raising. 

2311 # It will also automatically delete the relevant rows from the 

2312 # trash table and the records table. 

2313 with self.bridge.emptyTrash( 

2314 self._table, record_class=StoredFileInfo, record_column="path" 

2315 ) as trash_data: 

2316 # Removing the artifacts themselves requires that the files are 

2317 # not also associated with refs that are not to be trashed. 

2318 # Therefore need to do a query with the file paths themselves 

2319 # and return all the refs associated with them. Can only delete 

2320 # a file if the refs to be trashed are the only refs associated 

2321 # with the file. 

2322 # This requires multiple copies of the trashed items 

2323 trashed, artifacts_to_keep = trash_data 

2324 

2325 if artifacts_to_keep is None: 

2326 # The bridge is not helping us so have to work it out 

2327 # ourselves. This is not going to be as efficient. 

2328 trashed = list(trashed) 

2329 

2330 # The instance check is for mypy since up to this point it 

2331 # does not know the type of info. 

2332 path_map = self._refs_associated_with_artifacts( 

2333 [info.path for _, info in trashed if isinstance(info, StoredFileInfo)] 

2334 ) 

2335 

2336 for ref, info in trashed: 

2337 

2338 # Mypy needs to know this is not the base class 

2339 assert isinstance(info, StoredFileInfo), f"Unexpectedly got info of class {type(info)}" 

2340 

2341 # Check for mypy 

2342 assert ref.id is not None, f"Internal logic error in emptyTrash with ref {ref}/{info}" 

2343 

2344 path_map[info.path].remove(ref.id) 

2345 if not path_map[info.path]: 2345 ↛ 2336line 2345 didn't jump to line 2336, because the condition on line 2345 was never false

2346 del path_map[info.path] 

2347 

2348 artifacts_to_keep = set(path_map) 

2349 

2350 for ref, info in trashed: 

2351 

2352 # Should not happen for this implementation but need 

2353 # to keep mypy happy. 

2354 assert info is not None, f"Internal logic error in emptyTrash with ref {ref}." 

2355 

2356 # Mypy needs to know this is not the base class 

2357 assert isinstance(info, StoredFileInfo), f"Unexpectedly got info of class {type(info)}" 

2358 

2359 # Check for mypy 

2360 assert ref.id is not None, f"Internal logic error in emptyTrash with ref {ref}/{info}" 

2361 

2362 if info.path in artifacts_to_keep: 

2363 # This is a multi-dataset artifact and we are not 

2364 # removing all associated refs. 

2365 continue 

2366 

2367 # Only trashed refs still known to datastore will be returned. 

2368 location = info.file_location(self.locationFactory) 

2369 

2370 # Point of no return for this artifact 

2371 log.debug("Removing artifact %s from datastore %s", location.uri, self.name) 

2372 try: 

2373 self._delete_artifact(location) 

2374 except FileNotFoundError: 

2375 # If the file itself has been deleted there is nothing 

2376 # we can do about it. It is possible that trash has 

2377 # been run in parallel in another process or someone 

2378 # decided to delete the file. It is unlikely to come 

2379 # back and so we should still continue with the removal 

2380 # of the entry from the trash table. It is also possible 

2381 # we removed it in a previous iteration if it was 

2382 # a multi-dataset artifact. The delete artifact method 

2383 # will log a debug message in this scenario. 

2384 # Distinguishing file missing before trash started and 

2385 # file already removed previously as part of this trash 

2386 # is not worth the distinction with regards to potential 

2387 # memory cost. 

2388 pass 

2389 except Exception as e: 

2390 if ignore_errors: 

2391 # Use a debug message here even though it's not 

2392 # a good situation. In some cases this can be 

2393 # caused by a race between user A and user B 

2394 # and neither of them has permissions for the 

2395 # other's files. Butler does not know about users 

2396 # and trash has no idea what collections these 

2397 # files were in (without guessing from a path). 

2398 log.debug( 

2399 "Encountered error removing artifact %s from datastore %s: %s", 

2400 location.uri, 

2401 self.name, 

2402 e, 

2403 ) 

2404 else: 

2405 raise 

2406 

2407 @transactional 

2408 def transfer_from( 

2409 self, 

2410 source_datastore: Datastore, 

2411 refs: Iterable[DatasetRef], 

2412 local_refs: Optional[Iterable[DatasetRef]] = None, 

2413 transfer: str = "auto", 

2414 artifact_existence: Optional[Dict[ResourcePath, bool]] = None, 

2415 ) -> None: 

2416 # Docstring inherited 

2417 if type(self) is not type(source_datastore): 

2418 raise TypeError( 

2419 f"Datastore mismatch between this datastore ({type(self)}) and the " 

2420 f"source datastore ({type(source_datastore)})." 

2421 ) 

2422 

2423 # Be explicit for mypy 

2424 if not isinstance(source_datastore, FileDatastore): 2424 ↛ 2425line 2424 didn't jump to line 2425, because the condition on line 2424 was never true

2425 raise TypeError( 

2426 "Can only transfer to a FileDatastore from another FileDatastore, not" 

2427 f" {type(source_datastore)}" 

2428 ) 

2429 

2430 # Stop early if "direct" transfer mode is requested. That would 

2431 # require that the URI inside the source datastore should be stored 

2432 # directly in the target datastore, which seems unlikely to be useful 

2433 # since at any moment the source datastore could delete the file. 

2434 if transfer in ("direct", "split"): 

2435 raise ValueError( 

2436 f"Can not transfer from a source datastore using {transfer} mode since" 

2437 " those files are controlled by the other datastore." 

2438 ) 

2439 

2440 # Empty existence lookup if none given. 

2441 if artifact_existence is None: 

2442 artifact_existence = {} 

2443 

2444 # We will go through the list multiple times so must convert 

2445 # generators to lists. 

2446 refs = list(refs) 

2447 

2448 if local_refs is None: 

2449 local_refs = refs 

2450 else: 

2451 local_refs = list(local_refs) 

2452 

2453 # In order to handle disassembled composites the code works 

2454 # at the records level since it can assume that internal APIs 

2455 # can be used. 

2456 # - If the record already exists in the destination this is assumed 

2457 # to be okay. 

2458 # - If there is no record but the source and destination URIs are 

2459 # identical no transfer is done but the record is added. 

2460 # - If the source record refers to an absolute URI currently assume 

2461 # that that URI should remain absolute and will be visible to the 

2462 # destination butler. May need to have a flag to indicate whether 

2463 # the dataset should be transferred. This will only happen if 

2464 # the detached Butler has had a local ingest. 

2465 

2466 # What we really want is all the records in the source datastore 

2467 # associated with these refs. Or derived ones if they don't exist 

2468 # in the source. 

2469 source_records = source_datastore._get_stored_records_associated_with_refs(refs) 

2470 

2471 # The source dataset_ids are the keys in these records 

2472 source_ids = set(source_records) 

2473 log.debug("Number of datastore records found in source: %d", len(source_ids)) 

2474 

2475 # The not None check is to appease mypy 

2476 requested_ids = set(ref.id for ref in refs if ref.id is not None) 

2477 missing_ids = requested_ids - source_ids 

2478 

2479 # Missing IDs can be okay if that datastore has allowed 

2480 # gets based on file existence. Should we transfer what we can 

2481 # or complain about it and warn? 

2482 if missing_ids and not source_datastore.trustGetRequest: 2482 ↛ 2483line 2482 didn't jump to line 2483, because the condition on line 2482 was never true

2483 raise ValueError( 

2484 f"Some datasets are missing from source datastore {source_datastore}: {missing_ids}" 

2485 ) 

2486 

2487 # Need to map these missing IDs to a DatasetRef so we can guess 

2488 # the details. 

2489 if missing_ids: 

2490 log.info( 

2491 "Number of expected datasets missing from source datastore records: %d out of %d", 

2492 len(missing_ids), 

2493 len(requested_ids), 

2494 ) 

2495 id_to_ref = {ref.id: ref for ref in refs if ref.id in missing_ids} 

2496 

2497 # This should be chunked in case we end up having to check 

2498 # the file store since we need some log output to show 

2499 # progress. 

2500 for missing_ids_chunk in chunk_iterable(missing_ids, chunk_size=10_000): 

2501 records = {} 

2502 for missing in missing_ids_chunk: 

2503 # Ask the source datastore where the missing artifacts 

2504 # should be. An execution butler might not know about the 

2505 # artifacts even if they are there. 

2506 expected = source_datastore._get_expected_dataset_locations_info(id_to_ref[missing]) 

2507 records[missing] = [info for _, info in expected] 

2508 

2509 # Call the mexist helper method in case we have not already 

2510 # checked these artifacts such that artifact_existence is 

2511 # empty. This allows us to benefit from parallelism. 

2512 # datastore.mexists() itself does not give us access to the 

2513 # derived datastore record. 

2514 log.verbose("Checking existence of %d datasets unknown to datastore", len(records)) 

2515 ref_exists = source_datastore._process_mexists_records( 

2516 id_to_ref, records, False, artifact_existence=artifact_existence 

2517 ) 

2518 

2519 # Now go through the records and propagate the ones that exist. 

2520 location_factory = source_datastore.locationFactory 

2521 for missing, record_list in records.items(): 

2522 # Skip completely if the ref does not exist. 

2523 ref = id_to_ref[missing] 

2524 if not ref_exists[ref]: 

2525 log.warning("Asked to transfer dataset %s but no file artifacts exist for it.", ref) 

2526 continue 

2527 # Check for file artifact to decide which parts of a 

2528 # disassembled composite do exist. If there is only a 

2529 # single record we don't even need to look because it can't 

2530 # be a composite and must exist. 

2531 if len(record_list) == 1: 

2532 dataset_records = record_list 

2533 else: 

2534 dataset_records = [ 

2535 record 

2536 for record in record_list 

2537 if artifact_existence[record.file_location(location_factory).uri] 

2538 ] 

2539 assert len(dataset_records) > 0, "Disassembled composite should have had some files." 

2540 

2541 # Rely on source_records being a defaultdict. 

2542 source_records[missing].extend(dataset_records) 

2543 

2544 # See if we already have these records 

2545 target_records = self._get_stored_records_associated_with_refs(local_refs) 

2546 

2547 # The artifacts to register 

2548 artifacts = [] 

2549 

2550 # Refs that already exist 

2551 already_present = [] 

2552 

2553 # Now can transfer the artifacts 

2554 for source_ref, target_ref in zip(refs, local_refs): 

2555 if target_ref.id in target_records: 

2556 # Already have an artifact for this. 

2557 already_present.append(target_ref) 

2558 continue 

2559 

2560 # mypy needs to know these are always resolved refs 

2561 for info in source_records[source_ref.getCheckedId()]: 

2562 source_location = info.file_location(source_datastore.locationFactory) 

2563 target_location = info.file_location(self.locationFactory) 

2564 if source_location == target_location: 2564 ↛ 2568line 2564 didn't jump to line 2568, because the condition on line 2564 was never true

2565 # Either the dataset is already in the target datastore 

2566 # (which is how execution butler currently runs) or 

2567 # it is an absolute URI. 

2568 if source_location.pathInStore.isabs(): 

2569 # Just because we can see the artifact when running 

2570 # the transfer doesn't mean it will be generally 

2571 # accessible to a user of this butler. For now warn 

2572 # but assume it will be accessible. 

2573 log.warning( 

2574 "Transfer request for an outside-datastore artifact has been found at %s", 

2575 source_location, 

2576 ) 

2577 else: 

2578 # Need to transfer it to the new location. 

2579 # Assume we should always overwrite. If the artifact 

2580 # is there this might indicate that a previous transfer 

2581 # was interrupted but was not able to be rolled back 

2582 # completely (eg pre-emption) so follow Datastore default 

2583 # and overwrite. 

2584 target_location.uri.transfer_from( 

2585 source_location.uri, transfer=transfer, overwrite=True, transaction=self._transaction 

2586 ) 

2587 

2588 artifacts.append((target_ref, info)) 

2589 

2590 self._register_datasets(artifacts) 

2591 

2592 if already_present: 

2593 n_skipped = len(already_present) 

2594 log.info( 

2595 "Skipped transfer of %d dataset%s already present in datastore", 

2596 n_skipped, 

2597 "" if n_skipped == 1 else "s", 

2598 ) 

2599 

2600 @transactional 

2601 def forget(self, refs: Iterable[DatasetRef]) -> None: 

2602 # Docstring inherited. 

2603 refs = list(refs) 

2604 self.bridge.forget(refs) 

2605 self._table.delete(["dataset_id"], *[{"dataset_id": ref.getCheckedId()} for ref in refs]) 

2606 

2607 def validateConfiguration( 

2608 self, entities: Iterable[Union[DatasetRef, DatasetType, StorageClass]], logFailures: bool = False 

2609 ) -> None: 

2610 """Validate some of the configuration for this datastore. 

2611 

2612 Parameters 

2613 ---------- 

2614 entities : iterable of `DatasetRef`, `DatasetType`, or `StorageClass` 

2615 Entities to test against this configuration. Can be differing 

2616 types. 

2617 logFailures : `bool`, optional 

2618 If `True`, output a log message for every validation error 

2619 detected. 

2620 

2621 Raises 

2622 ------ 

2623 DatastoreValidationError 

2624 Raised if there is a validation problem with a configuration. 

2625 All the problems are reported in a single exception. 

2626 

2627 Notes 

2628 ----- 

2629 This method checks that all the supplied entities have valid file 

2630 templates and also have formatters defined. 

2631 """ 

2632 

2633 templateFailed = None 

2634 try: 

2635 self.templates.validateTemplates(entities, logFailures=logFailures) 

2636 except FileTemplateValidationError as e: 

2637 templateFailed = str(e) 

2638 

2639 formatterFailed = [] 

2640 for entity in entities: 

2641 try: 

2642 self.formatterFactory.getFormatterClass(entity) 

2643 except KeyError as e: 

2644 formatterFailed.append(str(e)) 

2645 if logFailures: 2645 ↛ 2640line 2645 didn't jump to line 2640, because the condition on line 2645 was never false

2646 log.critical("Formatter failure: %s", e) 

2647 

2648 if templateFailed or formatterFailed: 

2649 messages = [] 

2650 if templateFailed: 2650 ↛ 2651line 2650 didn't jump to line 2651, because the condition on line 2650 was never true

2651 messages.append(templateFailed) 

2652 if formatterFailed: 2652 ↛ 2654line 2652 didn't jump to line 2654, because the condition on line 2652 was never false

2653 messages.append(",".join(formatterFailed)) 

2654 msg = ";\n".join(messages) 

2655 raise DatastoreValidationError(msg) 

2656 

2657 def getLookupKeys(self) -> Set[LookupKey]: 

2658 # Docstring is inherited from base class 

2659 return ( 

2660 self.templates.getLookupKeys() 

2661 | self.formatterFactory.getLookupKeys() 

2662 | self.constraints.getLookupKeys() 

2663 ) 

2664 

2665 def validateKey(self, lookupKey: LookupKey, entity: Union[DatasetRef, DatasetType, StorageClass]) -> None: 

2666 # Docstring is inherited from base class 

2667 # The key can be valid in either formatters or templates so we can 

2668 # only check the template if it exists 

2669 if lookupKey in self.templates: 

2670 try: 

2671 self.templates[lookupKey].validateTemplate(entity) 

2672 except FileTemplateValidationError as e: 

2673 raise DatastoreValidationError(e) from e 

2674 

2675 def export( 

2676 self, 

2677 refs: Iterable[DatasetRef], 

2678 *, 

2679 directory: Optional[ResourcePathExpression] = None, 

2680 transfer: Optional[str] = "auto", 

2681 ) -> Iterable[FileDataset]: 

2682 # Docstring inherited from Datastore.export. 

2683 if transfer is not None and directory is None: 2683 ↛ 2684line 2683 didn't jump to line 2684, because the condition on line 2683 was never true

2684 raise RuntimeError(f"Cannot export using transfer mode {transfer} with no export directory given") 

2685 

2686 if transfer == "move": 2686 ↛ 2687line 2686 didn't jump to line 2687, because the condition on line 2686 was never true

2687 raise RuntimeError("Can not export by moving files out of datastore.") 

2688 elif transfer == "direct": 2688 ↛ 2692line 2688 didn't jump to line 2692, because the condition on line 2688 was never true

2689 # For an export, treat this as equivalent to None. We do not 

2690 # want an import to risk using absolute URIs to datasets owned 

2691 # by another datastore. 

2692 log.info("Treating 'direct' transfer mode as in-place export.") 

2693 transfer = None 

2694 

2695 # Force the directory to be a URI object 

2696 directoryUri: Optional[ResourcePath] = None 

2697 if directory is not None: 2697 ↛ 2700line 2697 didn't jump to line 2700, because the condition on line 2697 was never false

2698 directoryUri = ResourcePath(directory, forceDirectory=True) 

2699 

2700 if transfer is not None and directoryUri is not None: 2700 ↛ 2705line 2700 didn't jump to line 2705, because the condition on line 2700 was never false

2701 # mypy needs the second test 

2702 if not directoryUri.exists(): 2702 ↛ 2703line 2702 didn't jump to line 2703, because the condition on line 2702 was never true

2703 raise FileNotFoundError(f"Export location {directory} does not exist") 

2704 

2705 progress = Progress("lsst.daf.butler.datastores.FileDatastore.export", level=logging.DEBUG) 

2706 for ref in progress.wrap(refs, "Exporting dataset files"): 

2707 fileLocations = self._get_dataset_locations_info(ref) 

2708 if not fileLocations: 2708 ↛ 2709line 2708 didn't jump to line 2709, because the condition on line 2708 was never true

2709 raise FileNotFoundError(f"Could not retrieve dataset {ref}.") 

2710 # For now we can not export disassembled datasets 

2711 if len(fileLocations) > 1: 

2712 raise NotImplementedError(f"Can not export disassembled datasets such as {ref}") 

2713 location, storedFileInfo = fileLocations[0] 

2714 

2715 pathInStore = location.pathInStore.path 

2716 if transfer is None: 2716 ↛ 2720line 2716 didn't jump to line 2720, because the condition on line 2716 was never true

2717 # TODO: do we also need to return the readStorageClass somehow? 

2718 # We will use the path in store directly. If this is an 

2719 # absolute URI, preserve it. 

2720 if location.pathInStore.isabs(): 

2721 pathInStore = str(location.uri) 

2722 elif transfer == "direct": 2722 ↛ 2724line 2722 didn't jump to line 2724, because the condition on line 2722 was never true

2723 # Use full URIs to the remote store in the export 

2724 pathInStore = str(location.uri) 

2725 else: 

2726 # mypy needs help 

2727 assert directoryUri is not None, "directoryUri must be defined to get here" 

2728 storeUri = ResourcePath(location.uri) 

2729 

2730 # if the datastore has an absolute URI to a resource, we 

2731 # have two options: 

2732 # 1. Keep the absolute URI in the exported YAML 

2733 # 2. Allocate a new name in the local datastore and transfer 

2734 # it. 

2735 # For now go with option 2 

2736 if location.pathInStore.isabs(): 2736 ↛ 2737line 2736 didn't jump to line 2737, because the condition on line 2736 was never true

2737 template = self.templates.getTemplate(ref) 

2738 newURI = ResourcePath(template.format(ref), forceAbsolute=False) 

2739 pathInStore = str(newURI.updatedExtension(location.pathInStore.getExtension())) 

2740 

2741 exportUri = directoryUri.join(pathInStore) 

2742 exportUri.transfer_from(storeUri, transfer=transfer) 

2743 

2744 yield FileDataset(refs=[ref], path=pathInStore, formatter=storedFileInfo.formatter) 

2745 

2746 @staticmethod 

2747 def computeChecksum( 

2748 uri: ResourcePath, algorithm: str = "blake2b", block_size: int = 8192 

2749 ) -> Optional[str]: 

2750 """Compute the checksum of the supplied file. 

2751 

2752 Parameters 

2753 ---------- 

2754 uri : `lsst.resources.ResourcePath` 

2755 Name of resource to calculate checksum from. 

2756 algorithm : `str`, optional 

2757 Name of algorithm to use. Must be one of the algorithms supported 

2758 by :py:class`hashlib`. 

2759 block_size : `int` 

2760 Number of bytes to read from file at one time. 

2761 

2762 Returns 

2763 ------- 

2764 hexdigest : `str` 

2765 Hex digest of the file. 

2766 

2767 Notes 

2768 ----- 

2769 Currently returns None if the URI is for a remote resource. 

2770 """ 

2771 if algorithm not in hashlib.algorithms_guaranteed: 2771 ↛ 2772line 2771 didn't jump to line 2772, because the condition on line 2771 was never true

2772 raise NameError("The specified algorithm '{}' is not supported by hashlib".format(algorithm)) 

2773 

2774 if not uri.isLocal: 2774 ↛ 2775line 2774 didn't jump to line 2775, because the condition on line 2774 was never true

2775 return None 

2776 

2777 hasher = hashlib.new(algorithm) 

2778 

2779 with uri.as_local() as local_uri: 

2780 with open(local_uri.ospath, "rb") as f: 

2781 for chunk in iter(lambda: f.read(block_size), b""): 

2782 hasher.update(chunk) 

2783 

2784 return hasher.hexdigest() 

2785 

2786 def needs_expanded_data_ids( 

2787 self, 

2788 transfer: Optional[str], 

2789 entity: Optional[Union[DatasetRef, DatasetType, StorageClass]] = None, 

2790 ) -> bool: 

2791 # Docstring inherited. 

2792 # This _could_ also use entity to inspect whether the filename template 

2793 # involves placeholders other than the required dimensions for its 

2794 # dataset type, but that's not necessary for correctness; it just 

2795 # enables more optimizations (perhaps only in theory). 

2796 return transfer not in ("direct", None) 

2797 

2798 def import_records(self, data: Mapping[str, DatastoreRecordData]) -> None: 

2799 # Docstring inherited from the base class. 

2800 record_data = data.get(self.name) 

2801 if not record_data: 2801 ↛ 2802line 2801 didn't jump to line 2802, because the condition on line 2801 was never true

2802 return 

2803 

2804 self._bridge.insert(FakeDatasetRef(dataset_id) for dataset_id in record_data.records.keys()) 

2805 

2806 # TODO: Verify that there are no unexpected table names in the dict? 

2807 unpacked_records = [] 

2808 for dataset_data in record_data.records.values(): 

2809 records = dataset_data.get(self._table.name) 

2810 if records: 2810 ↛ 2808line 2810 didn't jump to line 2808, because the condition on line 2810 was never false

2811 for info in records: 

2812 assert isinstance(info, StoredFileInfo), "Expecting StoredFileInfo records" 

2813 unpacked_records.append(info.to_record()) 

2814 if unpacked_records: 

2815 self._table.insert(*unpacked_records) 

2816 

2817 def export_records(self, refs: Iterable[DatasetIdRef]) -> Mapping[str, DatastoreRecordData]: 

2818 # Docstring inherited from the base class. 

2819 exported_refs = list(self._bridge.check(refs)) 

2820 ids = {ref.getCheckedId() for ref in exported_refs} 

2821 records: defaultdict[DatasetId, defaultdict[str, List[StoredDatastoreItemInfo]]] = defaultdict( 

2822 lambda: defaultdict(list), {id: defaultdict(list) for id in ids} 

2823 ) 

2824 for row in self._table.fetch(dataset_id=ids): 

2825 info: StoredDatastoreItemInfo = StoredFileInfo.from_record(row) 

2826 records[info.dataset_id][self._table.name].append(info) 

2827 

2828 record_data = DatastoreRecordData(records=records) 

2829 return {self.name: record_data}