Coverage for tests/test_dimensions.py: 11%
444 statements
« prev ^ index » next coverage.py v7.2.7, created at 2023-08-05 01:26 +0000
« prev ^ index » next coverage.py v7.2.7, created at 2023-08-05 01:26 +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 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/>.
22import copy
23import itertools
24import math
25import os
26import pickle
27import unittest
28from collections.abc import Iterator
29from dataclasses import dataclass
30from random import Random
32import lsst.sphgeom
33from lsst.daf.butler import (
34 Config,
35 DataCoordinate,
36 DataCoordinateSequence,
37 DataCoordinateSet,
38 Dimension,
39 DimensionConfig,
40 DimensionGraph,
41 DimensionPacker,
42 DimensionUniverse,
43 NamedKeyDict,
44 NamedValueSet,
45 TimespanDatabaseRepresentation,
46 YamlRepoImportBackend,
47)
48from lsst.daf.butler.registry import RegistryConfig, _RegistryFactory
50DIMENSION_DATA_FILE = os.path.normpath(
51 os.path.join(os.path.dirname(__file__), "data", "registry", "hsc-rc2-subset.yaml")
52)
55def loadDimensionData() -> DataCoordinateSequence:
56 """Load dimension data from an export file included in the code repository.
58 Returns
59 -------
60 dataIds : `DataCoordinateSet`
61 A set containing all data IDs in the export file.
62 """
63 # Create an in-memory SQLite database and Registry just to import the YAML
64 # data and retreive it as a set of DataCoordinate objects.
65 config = RegistryConfig()
66 config["db"] = "sqlite://"
67 registry = _RegistryFactory(config).create_from_config()
68 with open(DIMENSION_DATA_FILE) as stream:
69 backend = YamlRepoImportBackend(stream, registry)
70 backend.register()
71 backend.load(datastore=None)
72 dimensions = DimensionGraph(registry.dimensions, names=["visit", "detector", "tract", "patch"])
73 return registry.queryDataIds(dimensions).expanded().toSequence()
76class ConcreteTestDimensionPacker(DimensionPacker):
77 """A concrete `DimensionPacker` for testing its base class implementations.
79 This class just returns the detector ID as-is.
80 """
82 def __init__(self, fixed: DataCoordinate, dimensions: DimensionGraph):
83 super().__init__(fixed, dimensions)
84 self._n_detectors = fixed.records["instrument"].detector_max
85 self._max_bits = (self._n_detectors - 1).bit_length()
87 @property
88 def maxBits(self) -> int:
89 # Docstring inherited from DimensionPacker.maxBits
90 return self._max_bits
92 def _pack(self, dataId: DataCoordinate) -> int:
93 # Docstring inherited from DimensionPacker._pack
94 return dataId["detector"]
96 def unpack(self, packedId: int) -> DataCoordinate:
97 # Docstring inherited from DimensionPacker.unpack
98 return DataCoordinate.standardize(
99 {
100 "instrument": self.fixed["instrument"],
101 "detector": packedId,
102 },
103 graph=self.dimensions,
104 )
107class DimensionTestCase(unittest.TestCase):
108 """Tests for dimensions.
110 All tests here rely on the content of ``config/dimensions.yaml``, either
111 to test that the definitions there are read in properly or just as generic
112 data for testing various operations.
113 """
115 def setUp(self):
116 self.universe = DimensionUniverse()
118 def checkGraphInvariants(self, graph):
119 elements = list(graph.elements)
120 for n, element in enumerate(elements):
121 # Ordered comparisons on graphs behave like sets.
122 self.assertLessEqual(element.graph, graph)
123 # Ordered comparisons on elements correspond to the ordering within
124 # a DimensionUniverse (topological, with deterministic
125 # tiebreakers).
126 for other in elements[:n]:
127 self.assertLess(other, element)
128 self.assertLessEqual(other, element)
129 for other in elements[n + 1 :]:
130 self.assertGreater(other, element)
131 self.assertGreaterEqual(other, element)
132 if isinstance(element, Dimension):
133 self.assertEqual(element.graph.required, element.required)
134 self.assertEqual(DimensionGraph(self.universe, graph.required), graph)
135 self.assertCountEqual(
136 graph.required,
137 [
138 dimension
139 for dimension in graph.dimensions
140 if not any(dimension in other.graph.implied for other in graph.elements)
141 ],
142 )
143 self.assertCountEqual(graph.implied, graph.dimensions - graph.required)
144 self.assertCountEqual(
145 graph.dimensions, [element for element in graph.elements if isinstance(element, Dimension)]
146 )
147 self.assertCountEqual(graph.dimensions, itertools.chain(graph.required, graph.implied))
148 # Check primary key traversal order: each element should follow any it
149 # requires, and element that is implied by any other in the graph
150 # follow at least one of those.
151 seen = NamedValueSet()
152 for element in graph.primaryKeyTraversalOrder:
153 with self.subTest(required=graph.required, implied=graph.implied, element=element):
154 seen.add(element)
155 self.assertLessEqual(element.graph.required, seen)
156 if element in graph.implied:
157 self.assertTrue(any(element in s.implied for s in seen))
158 self.assertCountEqual(seen, graph.elements)
160 def testConfigPresent(self):
161 config = self.universe.dimensionConfig
162 self.assertIsInstance(config, DimensionConfig)
164 def testCompatibility(self):
165 # Simple check that should always be true.
166 self.assertTrue(self.universe.isCompatibleWith(self.universe))
168 # Create a universe like the default universe but with a different
169 # version number.
170 clone = self.universe.dimensionConfig.copy()
171 clone["version"] = clone["version"] + 1_000_000 # High version number
172 universe_clone = DimensionUniverse(config=clone)
173 with self.assertLogs("lsst.daf.butler.core.dimensions", "INFO") as cm:
174 self.assertTrue(self.universe.isCompatibleWith(universe_clone))
175 self.assertIn("differing versions", "\n".join(cm.output))
177 # Create completely incompatible universe.
178 config = Config(
179 {
180 "version": 1,
181 "namespace": "compat_test",
182 "skypix": {
183 "common": "htm7",
184 "htm": {
185 "class": "lsst.sphgeom.HtmPixelization",
186 "max_level": 24,
187 },
188 },
189 "elements": {
190 "A": {
191 "keys": [
192 {
193 "name": "id",
194 "type": "int",
195 }
196 ],
197 "storage": {
198 "cls": "lsst.daf.butler.registry.dimensions.table.TableDimensionRecordStorage",
199 },
200 },
201 "B": {
202 "keys": [
203 {
204 "name": "id",
205 "type": "int",
206 }
207 ],
208 "storage": {
209 "cls": "lsst.daf.butler.registry.dimensions.table.TableDimensionRecordStorage",
210 },
211 },
212 },
213 "packers": {},
214 }
215 )
216 universe2 = DimensionUniverse(config=config)
217 self.assertFalse(universe2.isCompatibleWith(self.universe))
219 def testVersion(self):
220 self.assertEqual(self.universe.namespace, "daf_butler")
221 # Test was added starting at version 2.
222 self.assertGreaterEqual(self.universe.version, 2)
224 def testConfigRead(self):
225 self.assertEqual(
226 set(self.universe.getStaticDimensions().names),
227 {
228 "instrument",
229 "visit",
230 "visit_system",
231 "exposure",
232 "detector",
233 "physical_filter",
234 "band",
235 "subfilter",
236 "skymap",
237 "tract",
238 "patch",
239 }
240 | {f"htm{level}" for level in range(25)}
241 | {f"healpix{level}" for level in range(18)},
242 )
244 def testGraphs(self):
245 self.checkGraphInvariants(self.universe.empty)
246 for element in self.universe.getStaticElements():
247 self.checkGraphInvariants(element.graph)
249 def testInstrumentDimensions(self):
250 graph = DimensionGraph(self.universe, names=("exposure", "detector", "visit"))
251 self.assertCountEqual(
252 graph.dimensions.names,
253 ("instrument", "exposure", "detector", "visit", "physical_filter", "band"),
254 )
255 self.assertCountEqual(graph.required.names, ("instrument", "exposure", "detector", "visit"))
256 self.assertCountEqual(graph.implied.names, ("physical_filter", "band"))
257 self.assertCountEqual(
258 graph.elements.names - graph.dimensions.names, ("visit_detector_region", "visit_definition")
259 )
260 self.assertCountEqual(graph.governors.names, {"instrument"})
262 def testCalibrationDimensions(self):
263 graph = DimensionGraph(self.universe, names=("physical_filter", "detector"))
264 self.assertCountEqual(graph.dimensions.names, ("instrument", "detector", "physical_filter", "band"))
265 self.assertCountEqual(graph.required.names, ("instrument", "detector", "physical_filter"))
266 self.assertCountEqual(graph.implied.names, ("band",))
267 self.assertCountEqual(graph.elements.names, graph.dimensions.names)
268 self.assertCountEqual(graph.governors.names, {"instrument"})
270 def testObservationDimensions(self):
271 graph = DimensionGraph(self.universe, names=("exposure", "detector", "visit"))
272 self.assertCountEqual(
273 graph.dimensions.names,
274 ("instrument", "detector", "visit", "exposure", "physical_filter", "band"),
275 )
276 self.assertCountEqual(graph.required.names, ("instrument", "detector", "exposure", "visit"))
277 self.assertCountEqual(graph.implied.names, ("physical_filter", "band"))
278 self.assertCountEqual(
279 graph.elements.names - graph.dimensions.names, ("visit_detector_region", "visit_definition")
280 )
281 self.assertCountEqual(graph.spatial.names, ("observation_regions",))
282 self.assertCountEqual(graph.temporal.names, ("observation_timespans",))
283 self.assertCountEqual(graph.governors.names, {"instrument"})
284 self.assertEqual(graph.spatial.names, {"observation_regions"})
285 self.assertEqual(graph.temporal.names, {"observation_timespans"})
286 self.assertEqual(next(iter(graph.spatial)).governor, self.universe["instrument"])
287 self.assertEqual(next(iter(graph.temporal)).governor, self.universe["instrument"])
288 self.assertEqual(self.universe["visit_definition"].populated_by, self.universe["visit"])
289 self.assertEqual(self.universe["visit_system_membership"].populated_by, self.universe["visit"])
290 self.assertEqual(self.universe["visit_detector_region"].populated_by, self.universe["visit"])
291 self.assertEqual(
292 self.universe.get_elements_populated_by(self.universe["visit"]),
293 NamedValueSet(
294 {
295 self.universe["visit"],
296 self.universe["visit_definition"],
297 self.universe["visit_system_membership"],
298 self.universe["visit_detector_region"],
299 }
300 ),
301 )
303 def testSkyMapDimensions(self):
304 graph = DimensionGraph(self.universe, names=("patch",))
305 self.assertCountEqual(graph.dimensions.names, ("skymap", "tract", "patch"))
306 self.assertCountEqual(graph.required.names, ("skymap", "tract", "patch"))
307 self.assertCountEqual(graph.implied.names, ())
308 self.assertCountEqual(graph.elements.names, graph.dimensions.names)
309 self.assertCountEqual(graph.spatial.names, ("skymap_regions",))
310 self.assertCountEqual(graph.governors.names, {"skymap"})
311 self.assertEqual(graph.spatial.names, {"skymap_regions"})
312 self.assertEqual(next(iter(graph.spatial)).governor, self.universe["skymap"])
314 def testSubsetCalculation(self):
315 """Test that independent spatial and temporal options are computed
316 correctly.
317 """
318 graph = DimensionGraph(
319 self.universe, names=("visit", "detector", "tract", "patch", "htm7", "exposure")
320 )
321 self.assertCountEqual(graph.spatial.names, ("observation_regions", "skymap_regions", "htm"))
322 self.assertCountEqual(graph.temporal.names, ("observation_timespans",))
324 def testSchemaGeneration(self):
325 tableSpecs = NamedKeyDict({})
326 for element in self.universe.getStaticElements():
327 if element.hasTable and element.viewOf is None:
328 tableSpecs[element] = element.RecordClass.fields.makeTableSpec(
329 TimespanReprClass=TimespanDatabaseRepresentation.Compound,
330 )
331 for element, tableSpec in tableSpecs.items():
332 for dep in element.required:
333 with self.subTest(element=element.name, dep=dep.name):
334 if dep != element:
335 self.assertIn(dep.name, tableSpec.fields)
336 self.assertEqual(tableSpec.fields[dep.name].dtype, dep.primaryKey.dtype)
337 self.assertEqual(tableSpec.fields[dep.name].length, dep.primaryKey.length)
338 self.assertEqual(tableSpec.fields[dep.name].nbytes, dep.primaryKey.nbytes)
339 self.assertFalse(tableSpec.fields[dep.name].nullable)
340 self.assertTrue(tableSpec.fields[dep.name].primaryKey)
341 else:
342 self.assertIn(element.primaryKey.name, tableSpec.fields)
343 self.assertEqual(
344 tableSpec.fields[element.primaryKey.name].dtype, dep.primaryKey.dtype
345 )
346 self.assertEqual(
347 tableSpec.fields[element.primaryKey.name].length, dep.primaryKey.length
348 )
349 self.assertEqual(
350 tableSpec.fields[element.primaryKey.name].nbytes, dep.primaryKey.nbytes
351 )
352 self.assertFalse(tableSpec.fields[element.primaryKey.name].nullable)
353 self.assertTrue(tableSpec.fields[element.primaryKey.name].primaryKey)
354 for dep in element.implied:
355 with self.subTest(element=element.name, dep=dep.name):
356 self.assertIn(dep.name, tableSpec.fields)
357 self.assertEqual(tableSpec.fields[dep.name].dtype, dep.primaryKey.dtype)
358 self.assertFalse(tableSpec.fields[dep.name].primaryKey)
359 for foreignKey in tableSpec.foreignKeys:
360 self.assertIn(foreignKey.table, tableSpecs)
361 self.assertIn(foreignKey.table, element.graph.dimensions.names)
362 self.assertEqual(len(foreignKey.source), len(foreignKey.target))
363 for source, target in zip(foreignKey.source, foreignKey.target, strict=True):
364 self.assertIn(source, tableSpec.fields.names)
365 self.assertIn(target, tableSpecs[foreignKey.table].fields.names)
366 self.assertEqual(
367 tableSpec.fields[source].dtype, tableSpecs[foreignKey.table].fields[target].dtype
368 )
369 self.assertEqual(
370 tableSpec.fields[source].length, tableSpecs[foreignKey.table].fields[target].length
371 )
372 self.assertEqual(
373 tableSpec.fields[source].nbytes, tableSpecs[foreignKey.table].fields[target].nbytes
374 )
376 def testPickling(self):
377 # Pickling and copying should always yield the exact same object within
378 # a single process (cross-process is impossible to test here).
379 universe1 = DimensionUniverse()
380 universe2 = pickle.loads(pickle.dumps(universe1))
381 universe3 = copy.copy(universe1)
382 universe4 = copy.deepcopy(universe1)
383 self.assertIs(universe1, universe2)
384 self.assertIs(universe1, universe3)
385 self.assertIs(universe1, universe4)
386 for element1 in universe1.getStaticElements():
387 element2 = pickle.loads(pickle.dumps(element1))
388 self.assertIs(element1, element2)
389 graph1 = element1.graph
390 graph2 = pickle.loads(pickle.dumps(graph1))
391 self.assertIs(graph1, graph2)
394@dataclass
395class SplitByStateFlags:
396 """A struct that separates data IDs with different states but the same
397 values.
398 """
400 minimal: DataCoordinateSequence | None = None
401 """Data IDs that only contain values for required dimensions.
403 `DataCoordinateSequence.hasFull()` will return `True` for this if and only
404 if ``minimal.graph.implied`` has no elements.
405 `DataCoordinate.hasRecords()` will always return `False`.
406 """
408 complete: DataCoordinateSequence | None = None
409 """Data IDs that contain values for all dimensions.
411 `DataCoordinateSequence.hasFull()` will always `True` and
412 `DataCoordinate.hasRecords()` will always return `True` for this attribute.
413 """
415 expanded: DataCoordinateSequence | None = None
416 """Data IDs that contain values for all dimensions as well as records.
418 `DataCoordinateSequence.hasFull()` and `DataCoordinate.hasRecords()` will
419 always return `True` for this attribute.
420 """
422 def chain(self, n: int | None = None) -> Iterator:
423 """Iterate over the data IDs of different types.
425 Parameters
426 ----------
427 n : `int`, optional
428 If provided (`None` is default), iterate over only the ``nth``
429 data ID in each attribute.
431 Yields
432 ------
433 dataId : `DataCoordinate`
434 A data ID from one of the attributes in this struct.
435 """
436 if n is None:
437 s = slice(None, None)
438 else:
439 s = slice(n, n + 1)
440 if self.minimal is not None:
441 yield from self.minimal[s]
442 if self.complete is not None:
443 yield from self.complete[s]
444 if self.expanded is not None:
445 yield from self.expanded[s]
448class DataCoordinateTestCase(unittest.TestCase):
449 """Test `DataCoordinate`."""
451 RANDOM_SEED = 10
453 @classmethod
454 def setUpClass(cls):
455 cls.allDataIds = loadDimensionData()
457 def setUp(self):
458 self.rng = Random(self.RANDOM_SEED)
460 def randomDataIds(self, n: int, dataIds: DataCoordinateSequence | None = None):
461 """Select random data IDs from those loaded from test data.
463 Parameters
464 ----------
465 n : `int`
466 Number of data IDs to select.
467 dataIds : `DataCoordinateSequence`, optional
468 Data IDs to select from. Defaults to ``self.allDataIds``.
470 Returns
471 -------
472 selected : `DataCoordinateSequence`
473 ``n`` Data IDs randomly selected from ``dataIds`` with replacement.
474 """
475 if dataIds is None:
476 dataIds = self.allDataIds
477 return DataCoordinateSequence(
478 self.rng.sample(dataIds, n),
479 graph=dataIds.graph,
480 hasFull=dataIds.hasFull(),
481 hasRecords=dataIds.hasRecords(),
482 check=False,
483 )
485 def randomDimensionSubset(self, n: int = 3, graph: DimensionGraph | None = None) -> DimensionGraph:
486 """Generate a random `DimensionGraph` that has a subset of the
487 dimensions in a given one.
489 Parameters
490 ----------
491 n : `int`
492 Number of dimensions to select, before automatic expansion by
493 `DimensionGraph`.
494 dataIds : `DimensionGraph`, optional
495 Dimensions to select from. Defaults to ``self.allDataIds.graph``.
497 Returns
498 -------
499 selected : `DimensionGraph`
500 ``n`` or more dimensions randomly selected from ``graph`` with
501 replacement.
502 """
503 if graph is None:
504 graph = self.allDataIds.graph
505 return DimensionGraph(
506 graph.universe, names=self.rng.sample(list(graph.dimensions.names), max(n, len(graph.dimensions)))
507 )
509 def splitByStateFlags(
510 self,
511 dataIds: DataCoordinateSequence | None = None,
512 *,
513 expanded: bool = True,
514 complete: bool = True,
515 minimal: bool = True,
516 ) -> SplitByStateFlags:
517 """Given a sequence of data IDs, generate new equivalent sequences
518 containing less information.
520 Parameters
521 ----------
522 dataIds : `DataCoordinateSequence`, optional.
523 Data IDs to start from. Defaults to ``self.allDataIds``.
524 ``dataIds.hasRecords()`` and ``dataIds.hasFull()`` must both return
525 `True`.
526 expanded : `bool`, optional
527 If `True` (default) include the original data IDs that contain all
528 information in the result.
529 complete : `bool`, optional
530 If `True` (default) include data IDs for which ``hasFull()``
531 returns `True` but ``hasRecords()`` does not.
532 minimal : `bool`, optional
533 If `True` (default) include data IDS that only contain values for
534 required dimensions, for which ``hasFull()`` may not return `True`.
536 Returns
537 -------
538 split : `SplitByStateFlags`
539 A dataclass holding the indicated data IDs in attributes that
540 correspond to the boolean keyword arguments.
541 """
542 if dataIds is None:
543 dataIds = self.allDataIds
544 assert dataIds.hasFull() and dataIds.hasRecords()
545 result = SplitByStateFlags(expanded=dataIds)
546 if complete:
547 result.complete = DataCoordinateSequence(
548 [DataCoordinate.standardize(e.full.byName(), graph=dataIds.graph) for e in result.expanded],
549 graph=dataIds.graph,
550 )
551 self.assertTrue(result.complete.hasFull())
552 self.assertFalse(result.complete.hasRecords())
553 if minimal:
554 result.minimal = DataCoordinateSequence(
555 [DataCoordinate.standardize(e.byName(), graph=dataIds.graph) for e in result.expanded],
556 graph=dataIds.graph,
557 )
558 self.assertEqual(result.minimal.hasFull(), not dataIds.graph.implied)
559 self.assertFalse(result.minimal.hasRecords())
560 if not expanded:
561 result.expanded = None
562 return result
564 def testMappingInterface(self):
565 """Test that the mapping interface in `DataCoordinate` and (when
566 applicable) its ``full`` property are self-consistent and consistent
567 with the ``graph`` property.
568 """
569 for _ in range(5):
570 dimensions = self.randomDimensionSubset()
571 dataIds = self.randomDataIds(n=1).subset(dimensions)
572 split = self.splitByStateFlags(dataIds)
573 for dataId in split.chain():
574 with self.subTest(dataId=dataId):
575 self.assertEqual(list(dataId.values()), [dataId[d] for d in dataId])
576 self.assertEqual(list(dataId.values()), [dataId[d.name] for d in dataId])
577 self.assertEqual(dataId.keys(), dataId.graph.required)
578 for dataId in itertools.chain(split.complete, split.expanded):
579 with self.subTest(dataId=dataId):
580 self.assertTrue(dataId.hasFull())
581 self.assertEqual(dataId.graph.dimensions, dataId.full.keys())
582 self.assertEqual(list(dataId.full.values()), [dataId[k] for k in dataId.graph.dimensions])
584 def test_pickle(self):
585 for _ in range(5):
586 dimensions = self.randomDimensionSubset()
587 dataIds = self.randomDataIds(n=1).subset(dimensions)
588 split = self.splitByStateFlags(dataIds)
589 for data_id in split.chain():
590 s = pickle.dumps(data_id)
591 read_data_id = pickle.loads(s)
592 self.assertEqual(data_id, read_data_id)
593 self.assertEqual(data_id.hasFull(), read_data_id.hasFull())
594 self.assertEqual(data_id.hasRecords(), read_data_id.hasRecords())
595 if data_id.hasFull():
596 self.assertEqual(data_id.full, read_data_id.full)
597 if data_id.hasRecords():
598 self.assertEqual(data_id.records, read_data_id.records)
600 def test_record_attributes(self):
601 """Test that dimension records are available as attributes on expanded
602 data coordinates.
603 """
604 for _ in range(5):
605 dimensions = self.randomDimensionSubset()
606 dataIds = self.randomDataIds(n=1).subset(dimensions)
607 split = self.splitByStateFlags(dataIds)
608 for data_id in split.expanded:
609 for element in data_id.graph.elements:
610 self.assertIs(getattr(data_id, element.name), data_id.records[element.name])
611 self.assertIn(element.name, dir(data_id))
612 with self.assertRaisesRegex(AttributeError, "^not_a_dimension_name$"):
613 data_id.not_a_dimension_name
614 for data_id in itertools.chain(split.minimal, split.complete):
615 for element in data_id.graph.elements:
616 with self.assertRaisesRegex(AttributeError, "only available on expanded DataCoordinates"):
617 getattr(data_id, element.name)
618 with self.assertRaisesRegex(AttributeError, "^not_a_dimension_name$"):
619 data_id.not_a_dimension_name
621 def testEquality(self):
622 """Test that different `DataCoordinate` instances with different state
623 flags can be compared with each other and other mappings.
624 """
625 dataIds = self.randomDataIds(n=2)
626 split = self.splitByStateFlags(dataIds)
627 # Iterate over all combinations of different states of DataCoordinate,
628 # with the same underlying data ID values.
629 for a0, b0 in itertools.combinations(split.chain(0), 2):
630 self.assertEqual(a0, b0)
631 self.assertEqual(a0, b0.byName())
632 self.assertEqual(a0.byName(), b0)
633 # Same thing, for a different data ID value.
634 for a1, b1 in itertools.combinations(split.chain(1), 2):
635 self.assertEqual(a1, b1)
636 self.assertEqual(a1, b1.byName())
637 self.assertEqual(a1.byName(), b1)
638 # Iterate over all combinations of different states of DataCoordinate,
639 # with different underlying data ID values.
640 for a0, b1 in itertools.product(split.chain(0), split.chain(1)):
641 self.assertNotEqual(a0, b1)
642 self.assertNotEqual(a1, b0)
643 self.assertNotEqual(a0, b1.byName())
644 self.assertNotEqual(a0.byName(), b1)
645 self.assertNotEqual(a1, b0.byName())
646 self.assertNotEqual(a1.byName(), b0)
648 def testStandardize(self):
649 """Test constructing a DataCoordinate from many different kinds of
650 input via `DataCoordinate.standardize` and `DataCoordinate.subset`.
651 """
652 for _ in range(5):
653 dimensions = self.randomDimensionSubset()
654 dataIds = self.randomDataIds(n=1).subset(dimensions)
655 split = self.splitByStateFlags(dataIds)
656 for dataId in split.chain():
657 # Passing in any kind of DataCoordinate alone just returns
658 # that object.
659 self.assertIs(dataId, DataCoordinate.standardize(dataId))
660 # Same if we also explicitly pass the dimensions we want.
661 self.assertIs(dataId, DataCoordinate.standardize(dataId, graph=dataId.graph))
662 # Same if we pass the dimensions and some irrelevant
663 # kwargs.
664 self.assertIs(dataId, DataCoordinate.standardize(dataId, graph=dataId.graph, htm7=12))
665 # Test constructing a new data ID from this one with a
666 # subset of the dimensions.
667 # This is not possible for some combinations of
668 # dimensions if hasFull is False (see
669 # `DataCoordinate.subset` docs).
670 newDimensions = self.randomDimensionSubset(n=1, graph=dataId.graph)
671 if dataId.hasFull() or dataId.graph.required.issuperset(newDimensions.required):
672 newDataIds = [
673 dataId.subset(newDimensions),
674 DataCoordinate.standardize(dataId, graph=newDimensions),
675 DataCoordinate.standardize(dataId, graph=newDimensions, htm7=12),
676 ]
677 for newDataId in newDataIds:
678 with self.subTest(newDataId=newDataId, type=type(dataId)):
679 commonKeys = dataId.keys() & newDataId.keys()
680 self.assertTrue(commonKeys)
681 self.assertEqual(
682 [newDataId[k] for k in commonKeys],
683 [dataId[k] for k in commonKeys],
684 )
685 # This should never "downgrade" from
686 # Complete to Minimal or Expanded to Complete.
687 if dataId.hasRecords():
688 self.assertTrue(newDataId.hasRecords())
689 if dataId.hasFull():
690 self.assertTrue(newDataId.hasFull())
691 # Start from a complete data ID, and pass its values in via several
692 # different ways that should be equivalent.
693 for dataId in split.complete:
694 # Split the keys (dimension names) into two random subsets, so
695 # we can pass some as kwargs below.
696 keys1 = set(
697 self.rng.sample(list(dataId.graph.dimensions.names), len(dataId.graph.dimensions) // 2)
698 )
699 keys2 = dataId.graph.dimensions.names - keys1
700 newCompleteDataIds = [
701 DataCoordinate.standardize(dataId.full.byName(), universe=dataId.universe),
702 DataCoordinate.standardize(dataId.full.byName(), graph=dataId.graph),
703 DataCoordinate.standardize(
704 DataCoordinate.makeEmpty(dataId.graph.universe), **dataId.full.byName()
705 ),
706 DataCoordinate.standardize(
707 DataCoordinate.makeEmpty(dataId.graph.universe),
708 graph=dataId.graph,
709 **dataId.full.byName(),
710 ),
711 DataCoordinate.standardize(**dataId.full.byName(), universe=dataId.universe),
712 DataCoordinate.standardize(graph=dataId.graph, **dataId.full.byName()),
713 DataCoordinate.standardize(
714 {k: dataId[k] for k in keys1},
715 universe=dataId.universe,
716 **{k: dataId[k] for k in keys2},
717 ),
718 DataCoordinate.standardize(
719 {k: dataId[k] for k in keys1}, graph=dataId.graph, **{k: dataId[k] for k in keys2}
720 ),
721 ]
722 for newDataId in newCompleteDataIds:
723 with self.subTest(dataId=dataId, newDataId=newDataId, type=type(dataId)):
724 self.assertEqual(dataId, newDataId)
725 self.assertTrue(newDataId.hasFull())
727 def testUnion(self):
728 """Test `DataCoordinate.union`."""
729 # Make test graphs to combine; mostly random, but with a few explicit
730 # cases to make sure certain edge cases are covered.
731 graphs = [self.randomDimensionSubset(n=2) for i in range(2)]
732 graphs.append(self.allDataIds.universe["visit"].graph)
733 graphs.append(self.allDataIds.universe["detector"].graph)
734 graphs.append(self.allDataIds.universe["physical_filter"].graph)
735 graphs.append(self.allDataIds.universe["band"].graph)
736 # Iterate over all combinations, including the same graph with itself.
737 for graph1, graph2 in itertools.product(graphs, repeat=2):
738 parentDataIds = self.randomDataIds(n=1)
739 split1 = self.splitByStateFlags(parentDataIds.subset(graph1))
740 split2 = self.splitByStateFlags(parentDataIds.subset(graph2))
741 (parentDataId,) = parentDataIds
742 for lhs, rhs in itertools.product(split1.chain(), split2.chain()):
743 unioned = lhs.union(rhs)
744 with self.subTest(lhs=lhs, rhs=rhs, unioned=unioned):
745 self.assertEqual(unioned.graph, graph1.union(graph2))
746 self.assertEqual(unioned, parentDataId.subset(unioned.graph))
747 if unioned.hasFull():
748 self.assertEqual(unioned.subset(lhs.graph), lhs)
749 self.assertEqual(unioned.subset(rhs.graph), rhs)
750 if lhs.hasFull() and rhs.hasFull():
751 self.assertTrue(unioned.hasFull())
752 if lhs.graph >= unioned.graph and lhs.hasFull():
753 self.assertTrue(unioned.hasFull())
754 if lhs.hasRecords():
755 self.assertTrue(unioned.hasRecords())
756 if rhs.graph >= unioned.graph and rhs.hasFull():
757 self.assertTrue(unioned.hasFull())
758 if rhs.hasRecords():
759 self.assertTrue(unioned.hasRecords())
760 if lhs.graph.required | rhs.graph.required >= unioned.graph.dimensions:
761 self.assertTrue(unioned.hasFull())
762 if (
763 lhs.hasRecords()
764 and rhs.hasRecords()
765 and lhs.graph.elements | rhs.graph.elements >= unioned.graph.elements
766 ):
767 self.assertTrue(unioned.hasRecords())
769 def testRegions(self):
770 """Test that data IDs for a few known dimensions have the expected
771 regions.
772 """
773 for dataId in self.randomDataIds(n=4).subset(
774 DimensionGraph(self.allDataIds.universe, names=["visit"])
775 ):
776 self.assertIsNotNone(dataId.region)
777 self.assertEqual(dataId.graph.spatial.names, {"observation_regions"})
778 self.assertEqual(dataId.region, dataId.records["visit"].region)
779 for dataId in self.randomDataIds(n=4).subset(
780 DimensionGraph(self.allDataIds.universe, names=["visit", "detector"])
781 ):
782 self.assertIsNotNone(dataId.region)
783 self.assertEqual(dataId.graph.spatial.names, {"observation_regions"})
784 self.assertEqual(dataId.region, dataId.records["visit_detector_region"].region)
785 for dataId in self.randomDataIds(n=4).subset(
786 DimensionGraph(self.allDataIds.universe, names=["tract"])
787 ):
788 self.assertIsNotNone(dataId.region)
789 self.assertEqual(dataId.graph.spatial.names, {"skymap_regions"})
790 self.assertEqual(dataId.region, dataId.records["tract"].region)
791 for dataId in self.randomDataIds(n=4).subset(
792 DimensionGraph(self.allDataIds.universe, names=["patch"])
793 ):
794 self.assertIsNotNone(dataId.region)
795 self.assertEqual(dataId.graph.spatial.names, {"skymap_regions"})
796 self.assertEqual(dataId.region, dataId.records["patch"].region)
797 for data_id in self.randomDataIds(n=1).subset(
798 DimensionGraph(self.allDataIds.universe, names=["visit", "tract"])
799 ):
800 self.assertEqual(data_id.region.relate(data_id.records["visit"].region), lsst.sphgeom.WITHIN)
801 self.assertEqual(data_id.region.relate(data_id.records["tract"].region), lsst.sphgeom.WITHIN)
803 def testTimespans(self):
804 """Test that data IDs for a few known dimensions have the expected
805 timespans.
806 """
807 for dataId in self.randomDataIds(n=4).subset(
808 DimensionGraph(self.allDataIds.universe, names=["visit"])
809 ):
810 self.assertIsNotNone(dataId.timespan)
811 self.assertEqual(dataId.graph.temporal.names, {"observation_timespans"})
812 self.assertEqual(dataId.timespan, dataId.records["visit"].timespan)
813 self.assertEqual(dataId.timespan, dataId.visit.timespan)
814 # Also test the case for non-temporal DataIds.
815 for dataId in self.randomDataIds(n=4).subset(
816 DimensionGraph(self.allDataIds.universe, names=["patch"])
817 ):
818 self.assertIsNone(dataId.timespan)
820 def testIterableStatusFlags(self):
821 """Test that DataCoordinateSet and DataCoordinateSequence compute
822 their hasFull and hasRecords flags correctly from their elements.
823 """
824 dataIds = self.randomDataIds(n=10)
825 split = self.splitByStateFlags(dataIds)
826 for cls in (DataCoordinateSet, DataCoordinateSequence):
827 self.assertTrue(cls(split.expanded, graph=dataIds.graph, check=True).hasFull())
828 self.assertTrue(cls(split.expanded, graph=dataIds.graph, check=False).hasFull())
829 self.assertTrue(cls(split.expanded, graph=dataIds.graph, check=True).hasRecords())
830 self.assertTrue(cls(split.expanded, graph=dataIds.graph, check=False).hasRecords())
831 self.assertTrue(cls(split.complete, graph=dataIds.graph, check=True).hasFull())
832 self.assertTrue(cls(split.complete, graph=dataIds.graph, check=False).hasFull())
833 self.assertFalse(cls(split.complete, graph=dataIds.graph, check=True).hasRecords())
834 self.assertFalse(cls(split.complete, graph=dataIds.graph, check=False).hasRecords())
835 with self.assertRaises(ValueError):
836 cls(split.complete, graph=dataIds.graph, hasRecords=True, check=True)
837 self.assertEqual(
838 cls(split.minimal, graph=dataIds.graph, check=True).hasFull(), not dataIds.graph.implied
839 )
840 self.assertEqual(
841 cls(split.minimal, graph=dataIds.graph, check=False).hasFull(), not dataIds.graph.implied
842 )
843 self.assertFalse(cls(split.minimal, graph=dataIds.graph, check=True).hasRecords())
844 self.assertFalse(cls(split.minimal, graph=dataIds.graph, check=False).hasRecords())
845 with self.assertRaises(ValueError):
846 cls(split.minimal, graph=dataIds.graph, hasRecords=True, check=True)
847 if dataIds.graph.implied:
848 with self.assertRaises(ValueError):
849 cls(split.minimal, graph=dataIds.graph, hasFull=True, check=True)
851 def testSetOperations(self):
852 """Test for self-consistency across DataCoordinateSet's operations."""
853 c = self.randomDataIds(n=10).toSet()
854 a = self.randomDataIds(n=20).toSet() | c
855 b = self.randomDataIds(n=20).toSet() | c
856 # Make sure we don't have a particularly unlucky random seed, since
857 # that would make a lot of this test uninteresting.
858 self.assertNotEqual(a, b)
859 self.assertGreater(len(a), 0)
860 self.assertGreater(len(b), 0)
861 # The rest of the tests should not depend on the random seed.
862 self.assertEqual(a, a)
863 self.assertNotEqual(a, a.toSequence())
864 self.assertEqual(a, a.toSequence().toSet())
865 self.assertEqual(a, a.toSequence().toSet())
866 self.assertEqual(b, b)
867 self.assertNotEqual(b, b.toSequence())
868 self.assertEqual(b, b.toSequence().toSet())
869 self.assertEqual(a & b, a.intersection(b))
870 self.assertLessEqual(a & b, a)
871 self.assertLessEqual(a & b, b)
872 self.assertEqual(a | b, a.union(b))
873 self.assertGreaterEqual(a | b, a)
874 self.assertGreaterEqual(a | b, b)
875 self.assertEqual(a - b, a.difference(b))
876 self.assertLessEqual(a - b, a)
877 self.assertLessEqual(b - a, b)
878 self.assertEqual(a ^ b, a.symmetric_difference(b))
879 self.assertGreaterEqual(a ^ b, (a | b) - (a & b))
881 def testPackers(self):
882 (instrument_data_id,) = self.allDataIds.subset(
883 self.allDataIds.universe.extract(["instrument"])
884 ).toSet()
885 (detector_data_id,) = self.randomDataIds(n=1).subset(self.allDataIds.universe.extract(["detector"]))
886 packer = ConcreteTestDimensionPacker(instrument_data_id, detector_data_id.graph)
887 packed_id, max_bits = packer.pack(detector_data_id, returnMaxBits=True)
888 self.assertEqual(packed_id, detector_data_id["detector"])
889 self.assertEqual(max_bits, packer.maxBits)
890 self.assertEqual(
891 max_bits, math.ceil(math.log2(instrument_data_id.records["instrument"].detector_max))
892 )
893 self.assertEqual(packer.pack(detector_data_id), packed_id)
894 self.assertEqual(packer.pack(detector=detector_data_id["detector"]), detector_data_id["detector"])
895 self.assertEqual(packer.unpack(packed_id), detector_data_id)
898if __name__ == "__main__":
899 unittest.main()