Coverage for python/lsst/daf/butler/dimensions/_records.py: 22%
171 statements
« prev ^ index » next coverage.py v7.3.2, created at 2023-12-06 10:53 +0000
« prev ^ index » next coverage.py v7.3.2, created at 2023-12-06 10:53 +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__ = ("DimensionRecord", "SerializedDimensionRecord")
32from collections.abc import Hashable
33from typing import TYPE_CHECKING, Any, ClassVar, Optional, Tuple
35import lsst.sphgeom
36from lsst.daf.butler._compat import PYDANTIC_V2, _BaseModelCompat
37from lsst.utils.classes import immutable
38from pydantic import Field, StrictBool, StrictFloat, StrictInt, StrictStr, create_model
40from .._timespan import Timespan, TimespanDatabaseRepresentation
41from ..json import from_json_pydantic, to_json_pydantic
42from ..persistence_context import PersistenceContextVars
43from ._elements import Dimension, DimensionElement
45if TYPE_CHECKING: # Imports needed only for type annotations; may be circular.
46 from ..registry import Registry
47 from ._coordinate import DataCoordinate
48 from ._schema import DimensionElementFields
49 from ._universe import DimensionUniverse
52def _reconstructDimensionRecord(definition: DimensionElement, mapping: dict[str, Any]) -> DimensionRecord:
53 """Unpickle implementation for `DimensionRecord` subclasses.
55 For internal use by `DimensionRecord`.
56 """
57 return definition.RecordClass(**mapping)
60def _subclassDimensionRecord(definition: DimensionElement) -> type[DimensionRecord]:
61 """Create a dynamic subclass of `DimensionRecord` for the given element.
63 For internal use by `DimensionRecord`.
64 """
65 from ._schema import DimensionElementFields
67 fields = DimensionElementFields(definition)
68 slots = list(fields.standard.names)
69 if definition.spatial:
70 slots.append("region")
71 if definition.temporal:
72 slots.append(TimespanDatabaseRepresentation.NAME)
73 d = {"definition": definition, "__slots__": tuple(slots), "fields": fields}
74 return type(definition.name + ".RecordClass", (DimensionRecord,), d)
77class SpecificSerializedDimensionRecord(_BaseModelCompat, extra="forbid"):
78 """Base model for a specific serialized record content."""
81_SIMPLE_RECORD_CLASS_CACHE: dict[
82 tuple[DimensionElement, DimensionUniverse], type[SpecificSerializedDimensionRecord]
83] = {}
86def _createSimpleRecordSubclass(definition: DimensionElement) -> type[SpecificSerializedDimensionRecord]:
87 from ._schema import DimensionElementFields
89 # Cache on the definition (which hashes as the name) and the
90 # associated universe.
91 cache_key = (definition, definition.universe)
92 if cache_key in _SIMPLE_RECORD_CLASS_CACHE:
93 return _SIMPLE_RECORD_CLASS_CACHE[cache_key]
95 fields = DimensionElementFields(definition)
96 members = {}
97 # Prefer strict typing for external data
98 type_map = {
99 str: StrictStr,
100 float: StrictFloat,
101 bool: StrictBool,
102 int: StrictInt,
103 }
105 for field in fields.standard:
106 field_type = field.getPythonType()
107 field_type = type_map.get(field_type, field_type)
108 if field.nullable:
109 field_type = Optional[field_type] # type: ignore
110 members[field.name] = (field_type, ...)
111 if definition.temporal:
112 members["timespan"] = (Optional[Tuple[int, int]], ...) # type: ignore
113 if definition.spatial:
114 members["region"] = (str, ...)
116 # mypy does not seem to like create_model
117 model = create_model(
118 f"SpecificSerializedDimensionRecord{definition.name.capitalize()}",
119 __base__=SpecificSerializedDimensionRecord,
120 **members, # type: ignore
121 )
123 _SIMPLE_RECORD_CLASS_CACHE[cache_key] = model
124 return model
127# While supporting pydantic v1 and v2 keep this outside the model.
128_serialized_dimension_record_schema_extra = {
129 "examples": [
130 {
131 "definition": "detector",
132 "record": {
133 "instrument": "HSC",
134 "id": 72,
135 "full_name": "0_01",
136 "name_in_raft": "01",
137 "raft": "0",
138 "purpose": "SCIENCE",
139 },
140 }
141 ]
142}
145class SerializedDimensionRecord(_BaseModelCompat):
146 """Simplified model for serializing a `DimensionRecord`."""
148 definition: str = Field(
149 ...,
150 title="Name of dimension associated with this record.",
151 examples=["exposure"],
152 )
154 # Use strict types to prevent casting
155 record: dict[str, None | StrictInt | StrictFloat | StrictStr | StrictBool | tuple[int, int]] = Field(
156 ...,
157 title="Dimension record keys and values.",
158 examples=[
159 {
160 "definition": "exposure",
161 "record": {
162 "instrument": "LATISS",
163 "exposure": 2021050300044,
164 "obs_id": "AT_O_20210503_00044",
165 },
166 }
167 ],
168 )
170 if PYDANTIC_V2: 170 ↛ 171line 170 didn't jump to line 171
171 model_config = {
172 "json_schema_extra": _serialized_dimension_record_schema_extra, # type: ignore[typeddict-item]
173 }
174 else:
176 class Config:
177 """Local configuration overrides for model."""
179 schema_extra = _serialized_dimension_record_schema_extra
181 @classmethod
182 def direct(
183 cls,
184 *,
185 definition: str,
186 record: dict[str, None | StrictFloat | StrictStr | StrictBool | StrictInt | tuple[int, int]],
187 ) -> SerializedDimensionRecord:
188 """Construct a `SerializedDimensionRecord` directly without validators.
190 This differs from the pydantic "construct" method in that the arguments
191 are explicitly what the model requires, and it will recurse through
192 members, constructing them from their corresponding `direct` methods.
194 This method should only be called when the inputs are trusted.
195 """
196 # This method requires tuples as values of the mapping, but JSON
197 # readers will read things in as lists. Be kind and transparently
198 # transform to tuples
199 _recItems = {
200 k: v if type(v) != list else tuple(v) for k, v in record.items() # type: ignore # noqa: E721
201 }
203 # Type ignore because the ternary statement seems to confuse mypy
204 # based on conflicting inferred types of v.
205 key = (
206 definition,
207 frozenset(_recItems.items()),
208 )
209 cache = PersistenceContextVars.serializedDimensionRecordMapping.get()
210 if cache is not None and (result := cache.get(key)) is not None:
211 return result
213 node = cls.model_construct(definition=definition, record=_recItems) # type: ignore
215 if cache is not None:
216 cache[key] = node
217 return node
220@immutable
221class DimensionRecord:
222 """Base class for the Python representation of database records.
224 Parameters
225 ----------
226 **kwargs
227 Field values for this record. Unrecognized keys are ignored. If this
228 is the record for a `Dimension`, its primary key value may be provided
229 with the actual name of the field (e.g. "id" or "name"), the name of
230 the `Dimension`, or both. If this record class has a "timespan"
231 attribute, "datetime_begin" and "datetime_end" keyword arguments may
232 be provided instead of a single "timespan" keyword argument (but are
233 ignored if a "timespan" argument is provided).
235 Notes
236 -----
237 `DimensionRecord` subclasses are created dynamically for each
238 `DimensionElement` in a `DimensionUniverse`, and are accessible via the
239 `DimensionElement.RecordClass` attribute. The `DimensionRecord` base class
240 itself is pure abstract, but does not use the `abc` module to indicate this
241 because it does not have overridable methods.
243 Record classes have attributes that correspond exactly to the
244 `~DimensionElementFields.standard` fields in the related database table,
245 plus "region" and "timespan" attributes for spatial and/or temporal
246 elements (respectively).
248 Instances are usually obtained from a `Registry`, but can be constructed
249 directly from Python as well.
251 `DimensionRecord` instances are immutable.
252 """
254 # Derived classes are required to define __slots__ as well, and it's those
255 # derived-class slots that other methods on the base class expect to see
256 # when they access self.__slots__.
257 __slots__ = ("dataId",)
259 _serializedType = SerializedDimensionRecord
261 def __init__(self, **kwargs: Any):
262 # Accept either the dimension name or the actual name of its primary
263 # key field; ensure both are present in the dict for convenience below.
264 if isinstance(self.definition, Dimension):
265 v = kwargs.get(self.definition.primaryKey.name)
266 if v is None:
267 v = kwargs.get(self.definition.name)
268 if v is None:
269 raise ValueError(
270 f"No value provided for {self.definition.name}.{self.definition.primaryKey.name}."
271 )
272 kwargs[self.definition.primaryKey.name] = v
273 else:
274 v2 = kwargs.setdefault(self.definition.name, v)
275 if v != v2:
276 raise ValueError(
277 "Multiple inconsistent values for "
278 f"{self.definition.name}.{self.definition.primaryKey.name}: {v!r} != {v2!r}."
279 )
281 from ._coordinate import DataCoordinate
283 object.__setattr__(
284 self,
285 "dataId",
286 DataCoordinate.from_required_values(
287 self.definition.minimal_group,
288 tuple(kwargs[dimension] for dimension in self.definition.required.names),
289 ),
290 )
291 # Don't need the primary key value aliased to the dimension name
292 # anymore.
293 kwargs.pop(self.definition.name, None)
295 for name in self.__slots__:
296 # Note that we remove from kwargs as we go, to make sure there's
297 # nothing left at the end.
298 object.__setattr__(self, name, kwargs.pop(name, None))
299 # Support 'datetime_begin' and 'datetime_end' instead of 'timespan' for
300 # backwards compatibility, but if one is provided both must be.
301 if self.definition.temporal is not None and self.timespan is None and "datetime_begin" in kwargs:
302 object.__setattr__(
303 self,
304 "timespan",
305 Timespan(
306 kwargs.pop("datetime_begin"),
307 kwargs.pop("datetime_end"),
308 ),
309 )
311 if kwargs:
312 raise TypeError(f"Invalid fields for {self.definition} dimension record: {set(kwargs.keys())}.")
314 def __eq__(self, other: Any) -> bool:
315 if type(other) != type(self):
316 return False
317 return self.dataId == other.dataId
319 def __hash__(self) -> int:
320 return hash(self.dataId.required_values)
322 def __str__(self) -> str:
323 lines = [f"{self.definition.name}:"]
324 lines.extend(f" {name}: {getattr(self, name)!r}" for name in self.__slots__)
325 return "\n".join(lines)
327 def __repr__(self) -> str:
328 return "{}.RecordClass({})".format(
329 self.definition.name, ", ".join(f"{name}={getattr(self, name)!r}" for name in self.__slots__)
330 )
332 def __reduce__(self) -> tuple:
333 mapping = {name: getattr(self, name) for name in self.__slots__}
334 return (_reconstructDimensionRecord, (self.definition, mapping))
336 def _repr_html_(self) -> str:
337 """Override the default representation in IPython/Jupyter notebooks.
339 This gives a more readable output that understands embedded newlines.
340 """
341 return f"<pre>{self}<pre>"
343 def to_simple(self, minimal: bool = False) -> SerializedDimensionRecord:
344 """Convert this class to a simple python type.
346 This makes it suitable for serialization.
348 Parameters
349 ----------
350 minimal : `bool`, optional
351 Use minimal serialization. Has no effect on for this class.
353 Returns
354 -------
355 names : `list`
356 The names of the dimensions.
357 """
358 # The DataId is sufficient if you are willing to do a deferred
359 # query. This may not be overly useful since to reconstruct
360 # a collection of records will require repeated registry queries.
361 # For now do not implement minimal form.
362 key = (id(self.definition), self.dataId)
363 cache = PersistenceContextVars.serializedDimensionRecordMapping.get()
364 if cache is not None and (result := cache.get(key)) is not None:
365 return result
367 mapping = {name: getattr(self, name) for name in self.__slots__}
368 # If the item in mapping supports simplification update it
369 for k, v in mapping.items():
370 try:
371 mapping[k] = v.to_simple(minimal=minimal)
372 except AttributeError:
373 if isinstance(v, lsst.sphgeom.Region):
374 # YAML serialization specifies the class when it
375 # doesn't have to. This is partly for explicitness
376 # and also history. Here use a different approach.
377 # This code needs to be migrated to sphgeom
378 mapping[k] = v.encode().hex()
379 if isinstance(v, bytes):
380 # We actually can't handle serializing out to bytes for
381 # hash objects, encode it here to a hex string
382 mapping[k] = v.hex()
383 definition = self.definition.to_simple(minimal=minimal)
384 dimRec = SerializedDimensionRecord(definition=definition, record=mapping)
385 if cache is not None:
386 cache[key] = dimRec
387 return dimRec
389 @classmethod
390 def from_simple(
391 cls,
392 simple: SerializedDimensionRecord,
393 universe: DimensionUniverse | None = None,
394 registry: Registry | None = None,
395 cacheKey: Hashable | None = None,
396 ) -> DimensionRecord:
397 """Construct a new object from the simplified form.
399 This is generally data returned from the `to_simple`
400 method.
402 Parameters
403 ----------
404 simple : `SerializedDimensionRecord`
405 Value return from `to_simple`.
406 universe : `DimensionUniverse`
407 The special graph of all known dimensions of which this graph will
408 be a subset. Can be `None` if `Registry` is provided.
409 registry : `lsst.daf.butler.Registry`, optional
410 Registry from which a universe can be extracted. Can be `None`
411 if universe is provided explicitly.
412 cacheKey : `Hashable` or `None`
413 If this is not None, it will be used as a key for any cached
414 reconstruction instead of calculating a value from the serialized
415 format.
417 Returns
418 -------
419 record : `DimensionRecord`
420 Newly-constructed object.
421 """
422 if universe is None and registry is None:
423 raise ValueError("One of universe or registry is required to convert names to a DimensionGraph")
424 if universe is None and registry is not None:
425 universe = registry.dimensions
426 if universe is None:
427 # this is for mypy
428 raise ValueError("Unable to determine a usable universe")
429 # Type ignore because the ternary statement seems to confuse mypy
430 # based on conflicting inferred types of v.
431 key = cacheKey or (
432 simple.definition,
433 frozenset(simple.record.items()), # type: ignore
434 )
435 cache = PersistenceContextVars.dimensionRecords.get()
436 if cache is not None and (result := cache.get(key)) is not None:
437 return result
439 definition = DimensionElement.from_simple(simple.definition, universe=universe)
441 # Create a specialist subclass model with type validation.
442 # This allows us to do simple checks of external data (possibly
443 # sent as JSON) since for now _reconstructDimensionRecord does not
444 # do any validation.
445 record_model_cls = _createSimpleRecordSubclass(definition)
446 record_model = record_model_cls(**simple.record)
448 # Timespan and region have to be converted to native form
449 # for now assume that those keys are special
450 rec = record_model.model_dump()
452 if (ts := "timespan") in rec:
453 rec[ts] = Timespan.from_simple(rec[ts], universe=universe, registry=registry)
454 if (reg := "region") in rec:
455 encoded = bytes.fromhex(rec[reg])
456 rec[reg] = lsst.sphgeom.Region.decode(encoded)
457 if (hsh := "hash") in rec:
458 rec[hsh] = bytes.fromhex(rec[hsh].decode())
460 dimRec = _reconstructDimensionRecord(definition, rec)
461 if cache is not None:
462 cache[key] = dimRec
463 return dimRec
465 to_json = to_json_pydantic
466 from_json: ClassVar = classmethod(from_json_pydantic)
468 def toDict(self, splitTimespan: bool = False) -> dict[str, Any]:
469 """Return a vanilla `dict` representation of this record.
471 Parameters
472 ----------
473 splitTimespan : `bool`, optional
474 If `True` (`False` is default) transform any "timespan" key value
475 from a `Timespan` instance into a pair of regular
476 ("datetime_begin", "datetime_end") fields.
477 """
478 results = {name: getattr(self, name) for name in self.__slots__}
479 if splitTimespan:
480 timespan = results.pop("timespan", None)
481 if timespan is not None:
482 results["datetime_begin"] = timespan.begin
483 results["datetime_end"] = timespan.end
484 return results
486 # DimensionRecord subclasses are dynamically created, so static type
487 # checkers can't know about them or their attributes. To avoid having to
488 # put "type: ignore", everywhere, add a dummy __getattr__ that tells type
489 # checkers not to worry about missing attributes.
490 def __getattr__(self, name: str) -> Any:
491 raise AttributeError(name)
493 # Class attributes below are shadowed by instance attributes, and are
494 # present just to hold the docstrings for those instance attributes.
496 dataId: DataCoordinate
497 """A dict-like identifier for this record's primary keys
498 (`DataCoordinate`).
499 """
501 definition: ClassVar[DimensionElement]
502 """The `DimensionElement` whose records this class represents
503 (`DimensionElement`).
504 """
506 fields: ClassVar[DimensionElementFields]
507 """A categorized view of the fields in this class
508 (`DimensionElementFields`).
509 """