Coverage for python/lsst/daf/butler/registry/datasets/byDimensions/tables.py: 97%
70 statements
« prev ^ index » next coverage.py v7.3.2, created at 2023-10-27 09:43 +0000
« prev ^ index » next coverage.py v7.3.2, created at 2023-10-27 09:43 +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
30from .... import ddl
32__all__ = (
33 "addDatasetForeignKey",
34 "makeCalibTableName",
35 "makeCalibTableSpec",
36 "makeStaticTableSpecs",
37 "makeTagTableName",
38 "makeTagTableSpec",
39 "StaticDatasetTablesTuple",
40)
42from collections import namedtuple
43from typing import Any
45import sqlalchemy
47from ...._dataset_type import DatasetType
48from ...._timespan import TimespanDatabaseRepresentation
49from ....dimensions import DimensionUniverse, GovernorDimension, addDimensionForeignKey
50from ...interfaces import CollectionManager, VersionTuple
52DATASET_TYPE_NAME_LENGTH = 128
55StaticDatasetTablesTuple = namedtuple(
56 "StaticDatasetTablesTuple",
57 [
58 "dataset_type",
59 "dataset",
60 ],
61)
64def addDatasetForeignKey(
65 tableSpec: ddl.TableSpec,
66 dtype: type,
67 *,
68 name: str = "dataset",
69 onDelete: str | None = None,
70 constraint: bool = True,
71 **kwargs: Any,
72) -> ddl.FieldSpec:
73 """Add a foreign key column for datasets and (optionally) a constraint to
74 a table.
76 This is an internal interface for the ``byDimensions`` package; external
77 code should use `DatasetRecordStorageManager.addDatasetForeignKey` instead.
79 Parameters
80 ----------
81 tableSpec : `ddl.TableSpec`
82 Specification for the table that should reference the dataset
83 table. Will be modified in place.
84 dtype: `type`
85 Type of the column, same as the column type of the PK column of
86 a referenced table (``dataset.id``).
87 name: `str`, optional
88 A name to use for the prefix of the new field; the full name is
89 ``{name}_id``.
90 onDelete: `str`, optional
91 One of "CASCADE" or "SET NULL", indicating what should happen to
92 the referencing row if the collection row is deleted. `None`
93 indicates that this should be an integrity error.
94 constraint: `bool`, optional
95 If `False` (`True` is default), add a field that can be joined to
96 the dataset primary key, but do not add a foreign key constraint.
97 **kwargs
98 Additional keyword arguments are forwarded to the `ddl.FieldSpec`
99 constructor (only the ``name`` and ``dtype`` arguments are
100 otherwise provided).
102 Returns
103 -------
104 idSpec : `ddl.FieldSpec`
105 Specification for the ID field.
106 """
107 idFieldSpec = ddl.FieldSpec(f"{name}_id", dtype=dtype, **kwargs)
108 tableSpec.fields.add(idFieldSpec)
109 if constraint:
110 tableSpec.foreignKeys.append(
111 ddl.ForeignKeySpec("dataset", source=(idFieldSpec.name,), target=("id",), onDelete=onDelete)
112 )
113 return idFieldSpec
116def makeStaticTableSpecs(
117 collections: type[CollectionManager],
118 universe: DimensionUniverse,
119 dtype: type,
120 autoincrement: bool,
121 schema_version: VersionTuple,
122) -> StaticDatasetTablesTuple:
123 """Construct all static tables used by the classes in this package.
125 Static tables are those that are present in all Registries and do not
126 depend on what DatasetTypes have been registered.
128 Parameters
129 ----------
130 collections: `CollectionManager`
131 Manager object for the collections in this `Registry`.
132 universe : `DimensionUniverse`
133 Universe graph containing all dimensions known to this `Registry`.
134 dtype: `type`
135 Type of the dataset ID (primary key) column.
136 autoincrement: `bool`
137 If `True` then dataset ID column will be auto-incrementing.
139 Returns
140 -------
141 specs : `StaticDatasetTablesTuple`
142 A named tuple containing `ddl.TableSpec` instances.
143 """
144 ingest_date_type: type
145 ingest_date_default: Any = None
146 if schema_version.major > 1:
147 ingest_date_type = ddl.AstropyTimeNsecTai
148 else:
149 ingest_date_type = sqlalchemy.TIMESTAMP
150 # New code provides explicit values for ingest_data, but we keep
151 # default just to be consistent with the existing schema.
152 ingest_date_default = sqlalchemy.sql.func.now()
154 specs = StaticDatasetTablesTuple(
155 dataset_type=ddl.TableSpec(
156 fields=[
157 ddl.FieldSpec(
158 name="id",
159 dtype=sqlalchemy.BigInteger,
160 autoincrement=True,
161 primaryKey=True,
162 doc=(
163 "Autoincrement ID that uniquely identifies a dataset "
164 "type in other tables. Python code outside the "
165 "`Registry` class should never interact with this; "
166 "its existence is considered an implementation detail."
167 ),
168 ),
169 ddl.FieldSpec(
170 name="name",
171 dtype=sqlalchemy.String,
172 length=DATASET_TYPE_NAME_LENGTH,
173 nullable=False,
174 doc="String name that uniquely identifies a dataset type.",
175 ),
176 ddl.FieldSpec(
177 name="storage_class",
178 dtype=sqlalchemy.String,
179 length=64,
180 nullable=False,
181 doc=(
182 "Name of the storage class associated with all "
183 "datasets of this type. Storage classes are "
184 "generally associated with a Python class, and are "
185 "enumerated in butler configuration."
186 ),
187 ),
188 ddl.FieldSpec(
189 name="dimensions_key",
190 dtype=sqlalchemy.BigInteger,
191 nullable=False,
192 doc="Unique key for the set of dimensions that identifies datasets of this type.",
193 ),
194 ddl.FieldSpec(
195 name="tag_association_table",
196 dtype=sqlalchemy.String,
197 length=128,
198 nullable=False,
199 doc=(
200 "Name of the table that holds associations between "
201 "datasets of this type and most types of collections."
202 ),
203 ),
204 ddl.FieldSpec(
205 name="calibration_association_table",
206 dtype=sqlalchemy.String,
207 length=128,
208 nullable=True,
209 doc=(
210 "Name of the table that holds associations between "
211 "datasets of this type and CALIBRATION collections. "
212 "NULL values indicate dataset types with "
213 "isCalibration=False."
214 ),
215 ),
216 ],
217 unique=[("name",)],
218 ),
219 dataset=ddl.TableSpec(
220 fields=[
221 ddl.FieldSpec(
222 name="id",
223 dtype=dtype,
224 autoincrement=autoincrement,
225 primaryKey=True,
226 doc="A unique field used as the primary key for dataset.",
227 ),
228 ddl.FieldSpec(
229 name="dataset_type_id",
230 dtype=sqlalchemy.BigInteger,
231 nullable=False,
232 doc="Reference to the associated entry in the dataset_type table.",
233 ),
234 ddl.FieldSpec(
235 name="ingest_date",
236 dtype=ingest_date_type,
237 default=ingest_date_default,
238 nullable=False,
239 doc="Time of dataset ingestion.",
240 ),
241 # Foreign key field/constraint to run added below.
242 ],
243 foreignKeys=[
244 ddl.ForeignKeySpec("dataset_type", source=("dataset_type_id",), target=("id",)),
245 ],
246 ),
247 )
248 # Add foreign key fields programmatically.
249 collections.addRunForeignKey(specs.dataset, onDelete="CASCADE", nullable=False)
250 return specs
253def makeTagTableName(datasetType: DatasetType, dimensionsKey: int) -> str:
254 """Construct the name for a dynamic (DatasetType-dependent) tag table used
255 by the classes in this package.
257 Parameters
258 ----------
259 datasetType : `DatasetType`
260 Dataset type to construct a name for. Multiple dataset types may
261 share the same table.
262 dimensionsKey : `int`
263 Integer key used to save ``datasetType.dimensions`` to the database.
265 Returns
266 -------
267 name : `str`
268 Name for the table.
269 """
270 return f"dataset_tags_{dimensionsKey:08d}"
273def makeCalibTableName(datasetType: DatasetType, dimensionsKey: int) -> str:
274 """Construct the name for a dynamic (DatasetType-dependent) tag + validity
275 range table used by the classes in this package.
277 Parameters
278 ----------
279 datasetType : `DatasetType`
280 Dataset type to construct a name for. Multiple dataset types may
281 share the same table.
282 dimensionsKey : `int`
283 Integer key used to save ``datasetType.dimensions`` to the database.
285 Returns
286 -------
287 name : `str`
288 Name for the table.
289 """
290 assert datasetType.isCalibration()
291 return f"dataset_calibs_{dimensionsKey:08d}"
294def makeTagTableSpec(
295 datasetType: DatasetType, collections: type[CollectionManager], dtype: type, *, constraints: bool = True
296) -> ddl.TableSpec:
297 """Construct the specification for a dynamic (DatasetType-dependent) tag
298 table used by the classes in this package.
300 Parameters
301 ----------
302 datasetType : `DatasetType`
303 Dataset type to construct a spec for. Multiple dataset types may
304 share the same table.
305 collections : `type` [ `CollectionManager` ]
306 `CollectionManager` subclass that can be used to construct foreign keys
307 to the run and/or collection tables.
308 dtype : `type`
309 Type of the FK column, same as the column type of the PK column of
310 a referenced table (``dataset.id``).
311 constraints : `bool`, optional
312 If `False` (`True` is default), do not define foreign key constraints.
314 Returns
315 -------
316 spec : `ddl.TableSpec`
317 Specification for the table.
318 """
319 tableSpec = ddl.TableSpec(
320 fields=[
321 # Foreign key fields to dataset, collection, and usually dimension
322 # tables added below.
323 # The dataset_type_id field here would be redundant with the one
324 # in the main monolithic dataset table, but we need it here for an
325 # important unique constraint.
326 ddl.FieldSpec("dataset_type_id", dtype=sqlalchemy.BigInteger, nullable=False),
327 ]
328 )
329 if constraints:
330 tableSpec.foreignKeys.append(
331 ddl.ForeignKeySpec("dataset_type", source=("dataset_type_id",), target=("id",))
332 )
333 # We'll also have a unique constraint on dataset type, collection, and data
334 # ID. We only include the required part of the data ID, as that's
335 # sufficient and saves us from worrying about nulls in the constraint.
336 constraint = ["dataset_type_id"]
337 # Add foreign key fields to dataset table (part of the primary key)
338 addDatasetForeignKey(tableSpec, dtype, primaryKey=True, onDelete="CASCADE", constraint=constraints)
339 # Add foreign key fields to collection table (part of the primary key and
340 # the data ID unique constraint).
341 collectionFieldSpec = collections.addCollectionForeignKey(
342 tableSpec, primaryKey=True, onDelete="CASCADE", constraint=constraints
343 )
344 constraint.append(collectionFieldSpec.name)
345 # Add foreign key constraint to the collection_summary_dataset_type table.
346 if constraints:
347 tableSpec.foreignKeys.append(
348 ddl.ForeignKeySpec(
349 "collection_summary_dataset_type",
350 source=(collectionFieldSpec.name, "dataset_type_id"),
351 target=(collectionFieldSpec.name, "dataset_type_id"),
352 )
353 )
354 for dimension in datasetType.dimensions.required:
355 fieldSpec = addDimensionForeignKey(
356 tableSpec, dimension=dimension, nullable=False, primaryKey=False, constraint=constraints
357 )
358 constraint.append(fieldSpec.name)
359 # If this is a governor dimension, add a foreign key constraint to the
360 # collection_summary_<dimension> table.
361 if isinstance(dimension, GovernorDimension) and constraints:
362 tableSpec.foreignKeys.append(
363 ddl.ForeignKeySpec(
364 f"collection_summary_{dimension.name}",
365 source=(collectionFieldSpec.name, fieldSpec.name),
366 target=(collectionFieldSpec.name, fieldSpec.name),
367 )
368 )
369 # Actually add the unique constraint.
370 tableSpec.unique.add(tuple(constraint))
371 return tableSpec
374def makeCalibTableSpec(
375 datasetType: DatasetType,
376 collections: type[CollectionManager],
377 TimespanReprClass: type[TimespanDatabaseRepresentation],
378 dtype: type,
379) -> ddl.TableSpec:
380 """Construct the specification for a dynamic (DatasetType-dependent) tag +
381 validity range table used by the classes in this package.
383 Parameters
384 ----------
385 datasetType : `DatasetType`
386 Dataset type to construct a spec for. Multiple dataset types may
387 share the same table.
388 collections : `type` [ `CollectionManager` ]
389 `CollectionManager` subclass that can be used to construct foreign keys
390 to the run and/or collection tables.
391 dtype: `type`
392 Type of the FK column, same as the column type of the PK column of
393 a referenced table (``dataset.id``).
395 Returns
396 -------
397 spec : `ddl.TableSpec`
398 Specification for the table.
399 """
400 tableSpec = ddl.TableSpec(
401 fields=[
402 # This table has no natural primary key, compound or otherwise, so
403 # we add an autoincrement key. We may use this field a bit
404 # internally, but its presence is an implementation detail and it
405 # shouldn't appear as a foreign key in any other tables.
406 ddl.FieldSpec("id", dtype=sqlalchemy.BigInteger, autoincrement=True, primaryKey=True),
407 # Foreign key fields to dataset, collection, and usually dimension
408 # tables added below. The dataset_type_id field here is redundant
409 # with the one in the main monolithic dataset table, but this bit
410 # of denormalization lets us define what should be a much more
411 # useful index.
412 ddl.FieldSpec("dataset_type_id", dtype=sqlalchemy.BigInteger, nullable=False),
413 ],
414 foreignKeys=[
415 ddl.ForeignKeySpec("dataset_type", source=("dataset_type_id",), target=("id",)),
416 ],
417 )
418 # Record fields that should go in the temporal lookup index/constraint,
419 # starting with the dataset type.
420 index: list[str | type[TimespanDatabaseRepresentation]] = ["dataset_type_id"]
421 # Add foreign key fields to dataset table (not part of the temporal
422 # lookup/constraint).
423 addDatasetForeignKey(tableSpec, dtype, nullable=False, onDelete="CASCADE")
424 # Add foreign key fields to collection table (part of the temporal lookup
425 # index/constraint).
426 collectionFieldSpec = collections.addCollectionForeignKey(tableSpec, nullable=False, onDelete="CASCADE")
427 index.append(collectionFieldSpec.name)
428 # Add foreign key constraint to the collection_summary_dataset_type table.
429 tableSpec.foreignKeys.append(
430 ddl.ForeignKeySpec(
431 "collection_summary_dataset_type",
432 source=(collectionFieldSpec.name, "dataset_type_id"),
433 target=(collectionFieldSpec.name, "dataset_type_id"),
434 )
435 )
436 # Add dimension fields (part of the temporal lookup index.constraint).
437 for dimension in datasetType.dimensions.required:
438 fieldSpec = addDimensionForeignKey(tableSpec, dimension=dimension, nullable=False, primaryKey=False)
439 index.append(fieldSpec.name)
440 # If this is a governor dimension, add a foreign key constraint to the
441 # collection_summary_<dimension> table.
442 if isinstance(dimension, GovernorDimension):
443 tableSpec.foreignKeys.append(
444 ddl.ForeignKeySpec(
445 f"collection_summary_{dimension.name}",
446 source=(collectionFieldSpec.name, fieldSpec.name),
447 target=(collectionFieldSpec.name, fieldSpec.name),
448 )
449 )
450 # Add validity-range field(s) (part of the temporal lookup
451 # index/constraint).
452 tsFieldSpecs = TimespanReprClass.makeFieldSpecs(nullable=False)
453 for fieldSpec in tsFieldSpecs:
454 tableSpec.fields.add(fieldSpec)
455 if TimespanReprClass.hasExclusionConstraint(): 455 ↛ 460line 455 didn't jump to line 460, because the condition on line 455 was never true
456 # This database's timespan representation can define a database-level
457 # constraint that prevents overlapping validity ranges for entries with
458 # the same DatasetType, collection, and data ID.
459 # This also creates an index.
460 index.append(TimespanReprClass)
461 tableSpec.exclusion.add(tuple(index))
462 else:
463 # No database-level constraint possible. We'll have to simulate that
464 # in our DatasetRecordStorage.certify() implementation, and just create
465 # a regular index here in the hope that helps with lookups.
466 index.extend(fieldSpec.name for fieldSpec in tsFieldSpecs)
467 tableSpec.indexes.add(ddl.IndexSpec(*index)) # type: ignore
468 return tableSpec