Coverage for python/lsst/daf/butler/transfers/_yaml.py: 14%
199 statements
« prev ^ index » next coverage.py v7.3.2, created at 2023-10-12 09:44 +0000
« prev ^ index » next coverage.py v7.3.2, created at 2023-10-12 09:44 +0000
1# This file is part of daf_butler.
2#
3# Developed for the LSST Data Management System.
4# This product includes software developed by the LSST Project
5# (http://www.lsst.org).
6# See the COPYRIGHT file at the top-level directory of this distribution
7# for details of code ownership.
8#
9# This software is dual licensed under the GNU General Public License and also
10# under a 3-clause BSD license. Recipients may choose which of these licenses
11# to use; please see the files gpl-3.0.txt and/or bsd_license.txt,
12# respectively. If you choose the GPL option then the following text applies
13# (but note that there is still no warranty even if you opt for BSD instead):
14#
15# This program is free software: you can redistribute it and/or modify
16# it under the terms of the GNU General Public License as published by
17# the Free Software Foundation, either version 3 of the License, or
18# (at your option) any later version.
19#
20# This program is distributed in the hope that it will be useful,
21# but WITHOUT ANY WARRANTY; without even the implied warranty of
22# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
23# GNU General Public License for more details.
24#
25# You should have received a copy of the GNU General Public License
26# along with this program. If not, see <http://www.gnu.org/licenses/>.
28from __future__ import annotations
30__all__ = ["YamlRepoExportBackend", "YamlRepoImportBackend"]
32import uuid
33import warnings
34from collections import UserDict, defaultdict
35from collections.abc import Iterable, Mapping
36from datetime import datetime
37from typing import IO, TYPE_CHECKING, Any
39import astropy.time
40import yaml
41from lsst.resources import ResourcePath
42from lsst.utils import doImportType
43from lsst.utils.introspection import find_outside_stacklevel
44from lsst.utils.iteration import ensure_iterable
46from .._dataset_association import DatasetAssociation
47from .._dataset_ref import DatasetId, DatasetRef
48from .._dataset_type import DatasetType
49from .._file_dataset import FileDataset
50from .._named import NamedValueSet
51from .._timespan import Timespan
52from ..datastore import Datastore
53from ..dimensions import DimensionElement, DimensionRecord, DimensionUniverse
54from ..registry import CollectionType, Registry
55from ..registry.interfaces import ChainedCollectionRecord, CollectionRecord, RunRecord, VersionTuple
56from ..registry.versions import IncompatibleVersionError
57from ._interfaces import RepoExportBackend, RepoImportBackend
59if TYPE_CHECKING:
60 from lsst.resources import ResourcePathExpression
62EXPORT_FORMAT_VERSION = VersionTuple(1, 0, 2)
63"""Export format version.
65Files with a different major version or a newer minor version cannot be read by
66this version of the code.
67"""
70class _RefMapper(UserDict[int, uuid.UUID]):
71 """Create a local dict subclass which creates new deterministic UUID for
72 missing keys.
73 """
75 _namespace = uuid.UUID("4d4851f4-2890-4d41-8779-5f38a3f5062b")
77 def __missing__(self, key: int) -> uuid.UUID:
78 newUUID = uuid.uuid3(namespace=self._namespace, name=str(key))
79 self[key] = newUUID
80 return newUUID
83_refIntId2UUID = _RefMapper()
86def _uuid_representer(dumper: yaml.Dumper, data: uuid.UUID) -> yaml.Node:
87 """Generate YAML representation for UUID.
89 This produces a scalar node with a tag "!uuid" and value being a regular
90 string representation of UUID.
91 """
92 return dumper.represent_scalar("!uuid", str(data))
95def _uuid_constructor(loader: yaml.Loader, node: yaml.Node) -> uuid.UUID | None:
96 if node.value is not None:
97 return uuid.UUID(hex=node.value)
98 return None
101yaml.Dumper.add_representer(uuid.UUID, _uuid_representer)
102yaml.SafeLoader.add_constructor("!uuid", _uuid_constructor)
105class YamlRepoExportBackend(RepoExportBackend):
106 """A repository export implementation that saves to a YAML file.
108 Parameters
109 ----------
110 stream
111 A writeable file-like object.
112 """
114 def __init__(self, stream: IO, universe: DimensionUniverse):
115 self.stream = stream
116 self.universe = universe
117 self.data: list[dict[str, Any]] = []
119 def saveDimensionData(self, element: DimensionElement, *data: DimensionRecord) -> None:
120 # Docstring inherited from RepoExportBackend.saveDimensionData.
121 data_dicts = [record.toDict(splitTimespan=True) for record in data]
122 self.data.append(
123 {
124 "type": "dimension",
125 "element": element.name,
126 "records": data_dicts,
127 }
128 )
130 def saveCollection(self, record: CollectionRecord, doc: str | None) -> None:
131 # Docstring inherited from RepoExportBackend.saveCollections.
132 data: dict[str, Any] = {
133 "type": "collection",
134 "collection_type": record.type.name,
135 "name": record.name,
136 }
137 if doc is not None:
138 data["doc"] = doc
139 if isinstance(record, RunRecord):
140 data["host"] = record.host
141 data["timespan_begin"] = record.timespan.begin
142 data["timespan_end"] = record.timespan.end
143 elif isinstance(record, ChainedCollectionRecord):
144 data["children"] = list(record.children)
145 self.data.append(data)
147 def saveDatasets(self, datasetType: DatasetType, run: str, *datasets: FileDataset) -> None:
148 # Docstring inherited from RepoExportBackend.saveDatasets.
149 self.data.append(
150 {
151 "type": "dataset_type",
152 "name": datasetType.name,
153 "dimensions": [d.name for d in datasetType.dimensions],
154 "storage_class": datasetType.storageClass_name,
155 "is_calibration": datasetType.isCalibration(),
156 }
157 )
158 self.data.append(
159 {
160 "type": "dataset",
161 "dataset_type": datasetType.name,
162 "run": run,
163 "records": [
164 {
165 "dataset_id": [ref.id for ref in sorted(dataset.refs)],
166 "data_id": [ref.dataId.byName() for ref in sorted(dataset.refs)],
167 "path": dataset.path,
168 "formatter": dataset.formatter,
169 # TODO: look up and save other collections
170 }
171 for dataset in datasets
172 ],
173 }
174 )
176 def saveDatasetAssociations(
177 self, collection: str, collectionType: CollectionType, associations: Iterable[DatasetAssociation]
178 ) -> None:
179 # Docstring inherited from RepoExportBackend.saveDatasetAssociations.
180 if collectionType is CollectionType.TAGGED:
181 self.data.append(
182 {
183 "type": "associations",
184 "collection": collection,
185 "collection_type": collectionType.name,
186 "dataset_ids": [assoc.ref.id for assoc in associations],
187 }
188 )
189 elif collectionType is CollectionType.CALIBRATION:
190 idsByTimespan: dict[Timespan, list[DatasetId]] = defaultdict(list)
191 for association in associations:
192 assert association.timespan is not None
193 idsByTimespan[association.timespan].append(association.ref.id)
194 self.data.append(
195 {
196 "type": "associations",
197 "collection": collection,
198 "collection_type": collectionType.name,
199 "validity_ranges": [
200 {
201 "timespan": timespan,
202 "dataset_ids": dataset_ids,
203 }
204 for timespan, dataset_ids in idsByTimespan.items()
205 ],
206 }
207 )
209 def finish(self) -> None:
210 # Docstring inherited from RepoExportBackend.
211 yaml.dump(
212 {
213 "description": "Butler Data Repository Export",
214 "version": str(EXPORT_FORMAT_VERSION),
215 "universe_version": self.universe.version,
216 "universe_namespace": self.universe.namespace,
217 "data": self.data,
218 },
219 stream=self.stream,
220 sort_keys=False,
221 )
224class YamlRepoImportBackend(RepoImportBackend):
225 """A repository import implementation that reads from a YAML file.
227 Parameters
228 ----------
229 stream
230 A readable file-like object.
231 registry : `Registry`
232 The registry datasets will be imported into. Only used to retreive
233 dataset types during construction; all write happen in `register`
234 and `load`.
235 """
237 def __init__(self, stream: IO, registry: Registry):
238 # We read the file fully and convert its contents to Python objects
239 # instead of loading incrementally so we can spot some problems early;
240 # because `register` can't be put inside a transaction, we'd rather not
241 # run that at all if there's going to be problem later in `load`.
242 wrapper = yaml.safe_load(stream)
243 if wrapper["version"] == 0:
244 # Grandfather-in 'version: 0' -> 1.0.0, which is what we wrote
245 # before we really tried to do versioning here.
246 fileVersion = VersionTuple(1, 0, 0)
247 else:
248 fileVersion = VersionTuple.fromString(wrapper["version"])
249 if fileVersion.major != EXPORT_FORMAT_VERSION.major:
250 raise IncompatibleVersionError(
251 f"Cannot read repository export file with version={fileVersion} "
252 f"({EXPORT_FORMAT_VERSION.major}.x.x required)."
253 )
254 if fileVersion.minor > EXPORT_FORMAT_VERSION.minor:
255 raise IncompatibleVersionError(
256 f"Cannot read repository export file with version={fileVersion} "
257 f"< {EXPORT_FORMAT_VERSION.major}.{EXPORT_FORMAT_VERSION.minor}.x required."
258 )
259 self.runs: dict[str, tuple[str | None, Timespan]] = {}
260 self.chains: dict[str, list[str]] = {}
261 self.collections: dict[str, CollectionType] = {}
262 self.collectionDocs: dict[str, str] = {}
263 self.datasetTypes: NamedValueSet[DatasetType] = NamedValueSet()
264 self.dimensions: Mapping[DimensionElement, list[DimensionRecord]] = defaultdict(list)
265 self.tagAssociations: dict[str, list[DatasetId]] = defaultdict(list)
266 self.calibAssociations: dict[str, dict[Timespan, list[DatasetId]]] = defaultdict(dict)
267 self.refsByFileId: dict[DatasetId, DatasetRef] = {}
268 self.registry: Registry = registry
270 universe_version = wrapper.get("universe_version", 0)
271 universe_namespace = wrapper.get("universe_namespace", "daf_butler")
273 # If this is data exported before the reorganization of visits
274 # and visit systems and that new schema is in use, some filtering
275 # will be needed. The entry in the visit dimension record will be
276 # silently dropped when visit is created but the
277 # visit_system_membership must be constructed.
278 migrate_visit_system = False
279 if (
280 universe_version < 2
281 and universe_namespace == "daf_butler"
282 and "visit_system_membership" in self.registry.dimensions
283 ):
284 migrate_visit_system = True
286 datasetData = []
287 for data in wrapper["data"]:
288 if data["type"] == "dimension":
289 # convert all datetime values to astropy
290 for record in data["records"]:
291 for key in record:
292 # Some older YAML files were produced with native
293 # YAML support for datetime, we support reading that
294 # data back. Newer conversion uses _AstropyTimeToYAML
295 # class with special YAML tag.
296 if isinstance(record[key], datetime):
297 record[key] = astropy.time.Time(record[key], scale="utc")
298 element = self.registry.dimensions[data["element"]]
299 RecordClass: type[DimensionRecord] = element.RecordClass
300 self.dimensions[element].extend(RecordClass(**r) for r in data["records"])
302 if data["element"] == "visit" and migrate_visit_system:
303 # Must create the visit_system_membership records.
304 element = self.registry.dimensions["visit_system_membership"]
305 RecordClass = element.RecordClass
306 self.dimensions[element].extend(
307 RecordClass(instrument=r["instrument"], visit_system=r["visit_system"], visit=r["id"])
308 for r in data["records"]
309 )
311 elif data["type"] == "collection":
312 collectionType = CollectionType.from_name(data["collection_type"])
313 if collectionType is CollectionType.RUN:
314 self.runs[data["name"]] = (
315 data["host"],
316 Timespan(begin=data["timespan_begin"], end=data["timespan_end"]),
317 )
318 elif collectionType is CollectionType.CHAINED:
319 children = []
320 for child in data["children"]:
321 if not isinstance(child, str):
322 warnings.warn(
323 f"CHAINED collection {data['name']} includes restrictions on child "
324 "collection searches, which are no longer suppored and will be ignored.",
325 stacklevel=find_outside_stacklevel("lsst.daf.butler"),
326 )
327 # Old form with dataset type restrictions only,
328 # supported for backwards compatibility.
329 child, _ = child
330 children.append(child)
331 self.chains[data["name"]] = children
332 else:
333 self.collections[data["name"]] = collectionType
334 doc = data.get("doc")
335 if doc is not None:
336 self.collectionDocs[data["name"]] = doc
337 elif data["type"] == "run":
338 # Also support old form of saving a run with no extra info.
339 self.runs[data["name"]] = (None, Timespan(None, None))
340 elif data["type"] == "dataset_type":
341 dimensions = data["dimensions"]
342 if migrate_visit_system and "visit" in dimensions and "visit_system" in dimensions:
343 dimensions.remove("visit_system")
344 self.datasetTypes.add(
345 DatasetType(
346 data["name"],
347 dimensions=dimensions,
348 storageClass=data["storage_class"],
349 universe=self.registry.dimensions,
350 isCalibration=data.get("is_calibration", False),
351 )
352 )
353 elif data["type"] == "dataset":
354 # Save raw dataset data for a second loop, so we can ensure we
355 # know about all dataset types first.
356 datasetData.append(data)
357 elif data["type"] == "associations":
358 collectionType = CollectionType.from_name(data["collection_type"])
359 if collectionType is CollectionType.TAGGED:
360 self.tagAssociations[data["collection"]].extend(
361 [x if not isinstance(x, int) else _refIntId2UUID[x] for x in data["dataset_ids"]]
362 )
363 elif collectionType is CollectionType.CALIBRATION:
364 assocsByTimespan = self.calibAssociations[data["collection"]]
365 for d in data["validity_ranges"]:
366 if "timespan" in d:
367 assocsByTimespan[d["timespan"]] = [
368 x if not isinstance(x, int) else _refIntId2UUID[x] for x in d["dataset_ids"]
369 ]
370 else:
371 # TODO: this is for backward compatibility, should
372 # be removed at some point.
373 assocsByTimespan[Timespan(begin=d["begin"], end=d["end"])] = [
374 x if not isinstance(x, int) else _refIntId2UUID[x] for x in d["dataset_ids"]
375 ]
376 else:
377 raise ValueError(f"Unexpected calibration type for association: {collectionType.name}.")
378 else:
379 raise ValueError(f"Unexpected dictionary type: {data['type']}.")
380 # key is (dataset type name, run)
381 self.datasets: Mapping[tuple[str, str], list[FileDataset]] = defaultdict(list)
382 for data in datasetData:
383 datasetType = self.datasetTypes.get(data["dataset_type"])
384 if datasetType is None:
385 datasetType = self.registry.getDatasetType(data["dataset_type"])
386 self.datasets[data["dataset_type"], data["run"]].extend(
387 FileDataset(
388 d.get("path"),
389 [
390 DatasetRef(
391 datasetType,
392 dataId,
393 run=data["run"],
394 id=refid if not isinstance(refid, int) else _refIntId2UUID[refid],
395 )
396 for dataId, refid in zip(
397 ensure_iterable(d["data_id"]), ensure_iterable(d["dataset_id"]), strict=True
398 )
399 ],
400 formatter=doImportType(d.get("formatter")) if "formatter" in d else None,
401 )
402 for d in data["records"]
403 )
405 def register(self) -> None:
406 # Docstring inherited from RepoImportBackend.register.
407 for datasetType in self.datasetTypes:
408 self.registry.registerDatasetType(datasetType)
409 for run in self.runs:
410 self.registry.registerRun(run, doc=self.collectionDocs.get(run))
411 # No way to add extra run info to registry yet.
412 for collection, collection_type in self.collections.items():
413 self.registry.registerCollection(
414 collection, collection_type, doc=self.collectionDocs.get(collection)
415 )
416 for chain, children in self.chains.items():
417 self.registry.registerCollection(
418 chain, CollectionType.CHAINED, doc=self.collectionDocs.get(chain)
419 )
420 self.registry.setCollectionChain(chain, children)
422 def load(
423 self,
424 datastore: Datastore | None,
425 *,
426 directory: ResourcePathExpression | None = None,
427 transfer: str | None = None,
428 skip_dimensions: set | None = None,
429 ) -> None:
430 # Docstring inherited from RepoImportBackend.load.
431 for element, dimensionRecords in self.dimensions.items():
432 if skip_dimensions and element in skip_dimensions:
433 continue
434 # Using skip_existing=True here assumes that the records in the
435 # database are either equivalent or at least preferable to the ones
436 # being imported. It'd be ideal to check that, but that would mean
437 # using syncDimensionData, which is not vectorized and is hence
438 # unacceptably slo.
439 self.registry.insertDimensionData(element, *dimensionRecords, skip_existing=True)
440 # FileDatasets to ingest into the datastore (in bulk):
441 fileDatasets = []
442 for records in self.datasets.values():
443 # Make a big flattened list of all data IDs and dataset_ids, while
444 # remembering slices that associate them with the FileDataset
445 # instances they came from.
446 datasets: list[DatasetRef] = []
447 dataset_ids: list[DatasetId] = []
448 slices = []
449 for fileDataset in records:
450 start = len(datasets)
451 datasets.extend(fileDataset.refs)
452 dataset_ids.extend(ref.id for ref in fileDataset.refs)
453 stop = len(datasets)
454 slices.append(slice(start, stop))
455 # Insert all of those DatasetRefs at once.
456 # For now, we ignore the dataset_id we pulled from the file
457 # and just insert without one to get a new autoincrement value.
458 # Eventually (once we have origin in IDs) we'll preserve them.
459 resolvedRefs = self.registry._importDatasets(datasets)
460 # Populate our dictionary that maps int dataset_id values from the
461 # export file to the new DatasetRefs
462 for fileId, ref in zip(dataset_ids, resolvedRefs, strict=True):
463 self.refsByFileId[fileId] = ref
464 # Now iterate over the original records, and install the new
465 # resolved DatasetRefs to replace the unresolved ones as we
466 # reorganize the collection information.
467 for sliceForFileDataset, fileDataset in zip(slices, records, strict=True):
468 fileDataset.refs = resolvedRefs[sliceForFileDataset]
469 if directory is not None:
470 fileDataset.path = ResourcePath(directory, forceDirectory=True).join(fileDataset.path)
471 fileDatasets.append(fileDataset)
472 # Ingest everything into the datastore at once.
473 if datastore is not None and fileDatasets:
474 datastore.ingest(*fileDatasets, transfer=transfer)
475 # Associate datasets with tagged collections.
476 for collection, dataset_ids in self.tagAssociations.items():
477 self.registry.associate(collection, [self.refsByFileId[i] for i in dataset_ids])
478 # Associate datasets with calibration collections.
479 for collection, idsByTimespan in self.calibAssociations.items():
480 for timespan, dataset_ids in idsByTimespan.items():
481 self.registry.certify(collection, [self.refsByFileId[i] for i in dataset_ids], timespan)