Coverage for python/lsst/daf/butler/core/datasets/type.py: 22%
Shortcuts on this page
r m x p toggle line displays
j k next/prev highlighted chunk
0 (zero) top of page
1 (one) first highlighted chunk
Shortcuts on this page
r m x p toggle line displays
j k next/prev highlighted chunk
0 (zero) top of page
1 (one) first highlighted chunk
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/>.
22from __future__ import annotations
24__all__ = ["DatasetType", "SerializedDatasetType"]
26from copy import deepcopy
27import re
29from types import MappingProxyType
31from typing import (
32 TYPE_CHECKING,
33 Any,
34 Callable,
35 Dict,
36 Iterable,
37 List,
38 Mapping,
39 Optional,
40 Tuple,
41 Type,
42 Union,
43)
45from pydantic import BaseModel, StrictStr, StrictBool
47from ..storageClass import StorageClass, StorageClassFactory
48from ..dimensions import DimensionGraph, SerializedDimensionGraph
49from ..configSupport import LookupKey
50from ..json import from_json_pydantic, to_json_pydantic
52if TYPE_CHECKING: 52 ↛ 53line 52 didn't jump to line 53, because the condition on line 52 was never true
53 from ..dimensions import Dimension, DimensionUniverse
54 from ...registry import Registry
57def _safeMakeMappingProxyType(data: Optional[Mapping]) -> Mapping:
58 if data is None:
59 data = {}
60 return MappingProxyType(data)
63class SerializedDatasetType(BaseModel):
64 """Simplified model of a `DatasetType` suitable for serialization."""
66 name: StrictStr
67 storageClass: Optional[StrictStr] = None
68 dimensions: Optional[SerializedDimensionGraph] = None
69 parentStorageClass: Optional[StrictStr] = None
70 isCalibration: StrictBool = False
72 @classmethod
73 def direct(cls, *, name: str, storageClass: Optional[str] = None,
74 dimensions: Optional[Dict] = None,
75 parentStorageClass: Optional[str] = None, isCalibration: bool = False
76 ) -> SerializedDatasetType:
77 """Construct a `SerializedDatasetType` directly without validators.
79 This differs from PyDantics construct method in that the arguments are
80 explicitly what the model requires, and it will recurse through
81 members, constructing them from their corresponding `direct` methods.
83 This method should only be called when the inputs are trusted.
84 """
85 node = SerializedDatasetType.__new__(cls)
86 setter = object.__setattr__
87 setter(node, 'name', name)
88 setter(node, 'storageClass', storageClass)
89 setter(node, 'dimensions',
90 dimensions if dimensions is None else SerializedDimensionGraph.direct(**dimensions))
91 setter(node, 'parentStorageClass', parentStorageClass)
92 setter(node, 'isCalibration', isCalibration)
93 setter(node, '__fields_set__', {'name', 'storageClass', 'dimensions', 'parentStorageClass',
94 'isCalibration'})
95 return node
98class DatasetType:
99 r"""A named category of Datasets.
101 Defines how they are organized, related, and stored.
103 A concrete, final class whose instances represent `DatasetType`\ s.
104 `DatasetType` instances may be constructed without a `Registry`,
105 but they must be registered
106 via `Registry.registerDatasetType()` before corresponding Datasets
107 may be added.
108 `DatasetType` instances are immutable.
110 Parameters
111 ----------
112 name : `str`
113 A string name for the Dataset; must correspond to the same
114 `DatasetType` across all Registries. Names must start with an
115 upper or lowercase letter, and may contain only letters, numbers,
116 and underscores. Component dataset types should contain a single
117 period separating the base dataset type name from the component name
118 (and may be recursive).
119 dimensions : `DimensionGraph` or iterable of `Dimension`
120 Dimensions used to label and relate instances of this `DatasetType`.
121 If not a `DimensionGraph`, ``universe`` must be provided as well.
122 storageClass : `StorageClass` or `str`
123 Instance of a `StorageClass` or name of `StorageClass` that defines
124 how this `DatasetType` is persisted.
125 parentStorageClass : `StorageClass` or `str`, optional
126 Instance of a `StorageClass` or name of `StorageClass` that defines
127 how the composite parent is persisted. Must be `None` if this
128 is not a component.
129 universe : `DimensionUniverse`, optional
130 Set of all known dimensions, used to normalize ``dimensions`` if it
131 is not already a `DimensionGraph`.
132 isCalibration : `bool`, optional
133 If `True`, this dataset type may be included in
134 `~CollectionType.CALIBRATION` collections.
136 See Also
137 --------
138 :ref:`daf_butler_organizing_datasets`
139 """
141 __slots__ = ("_name", "_dimensions", "_storageClass", "_storageClassName",
142 "_parentStorageClass", "_parentStorageClassName",
143 "_isCalibration")
145 _serializedType = SerializedDatasetType
147 VALID_NAME_REGEX = re.compile("^[a-zA-Z][a-zA-Z0-9_]*(\\.[a-zA-Z][a-zA-Z0-9_]*)*$")
149 @staticmethod
150 def nameWithComponent(datasetTypeName: str, componentName: str) -> str:
151 """Form a valid DatasetTypeName from a parent and component.
153 No validation is performed.
155 Parameters
156 ----------
157 datasetTypeName : `str`
158 Base type name.
159 componentName : `str`
160 Name of component.
162 Returns
163 -------
164 compTypeName : `str`
165 Name to use for component DatasetType.
166 """
167 return "{}.{}".format(datasetTypeName, componentName)
169 def __init__(self, name: str, dimensions: Union[DimensionGraph, Iterable[Dimension]],
170 storageClass: Union[StorageClass, str],
171 parentStorageClass: Optional[Union[StorageClass, str]] = None, *,
172 universe: Optional[DimensionUniverse] = None,
173 isCalibration: bool = False):
174 if self.VALID_NAME_REGEX.match(name) is None:
175 raise ValueError(f"DatasetType name '{name}' is invalid.")
176 self._name = name
177 if not isinstance(dimensions, DimensionGraph):
178 if universe is None:
179 raise ValueError("If dimensions is not a normalized DimensionGraph, "
180 "a universe must be provided.")
181 dimensions = universe.extract(dimensions)
182 self._dimensions = dimensions
183 if name in self._dimensions.universe.getGovernorDimensions().names:
184 raise ValueError(f"Governor dimension name {name} cannot be used as a dataset type name.")
185 if not isinstance(storageClass, (StorageClass, str)):
186 raise ValueError("StorageClass argument must be StorageClass or str. "
187 f"Got {storageClass}")
188 self._storageClass: Optional[StorageClass]
189 if isinstance(storageClass, StorageClass):
190 self._storageClass = storageClass
191 self._storageClassName = storageClass.name
192 else:
193 self._storageClass = None
194 self._storageClassName = storageClass
196 self._parentStorageClass: Optional[StorageClass] = None
197 self._parentStorageClassName: Optional[str] = None
198 if parentStorageClass is not None:
199 if not isinstance(storageClass, (StorageClass, str)):
200 raise ValueError("Parent StorageClass argument must be StorageClass or str. "
201 f"Got {parentStorageClass}")
203 # Only allowed for a component dataset type
204 _, componentName = self.splitDatasetTypeName(self._name)
205 if componentName is None:
206 raise ValueError("Can not specify a parent storage class if this is not a component"
207 f" ({self._name})")
208 if isinstance(parentStorageClass, StorageClass):
209 self._parentStorageClass = parentStorageClass
210 self._parentStorageClassName = parentStorageClass.name
211 else:
212 self._parentStorageClassName = parentStorageClass
214 # Ensure that parent storage class is specified when we have
215 # a component and is not specified when we don't
216 _, componentName = self.splitDatasetTypeName(self._name)
217 if parentStorageClass is None and componentName is not None:
218 raise ValueError(f"Component dataset type '{self._name}' constructed without parent"
219 " storage class")
220 if parentStorageClass is not None and componentName is None:
221 raise ValueError(f"Parent storage class specified by {self._name} is not a composite")
222 self._isCalibration = isCalibration
224 def __repr__(self) -> str:
225 extra = ""
226 if self._parentStorageClassName:
227 extra = f", parentStorageClass={self._parentStorageClassName}"
228 if self._isCalibration:
229 extra += ", isCalibration=True"
230 return f"DatasetType({self.name!r}, {self.dimensions}, {self._storageClassName}{extra})"
232 def __eq__(self, other: Any) -> bool:
233 if not isinstance(other, type(self)):
234 return False
235 if self._name != other._name:
236 return False
237 if self._dimensions != other._dimensions:
238 return False
239 if self._storageClass is not None and other._storageClass is not None:
240 if self._storageClass != other._storageClass:
241 return False
242 else:
243 if self._storageClassName != other._storageClassName:
244 return False
245 if self._isCalibration != other._isCalibration:
246 return False
247 if self._parentStorageClass is not None and other._parentStorageClass is not None:
248 return self._parentStorageClass == other._parentStorageClass
249 else:
250 return self._parentStorageClassName == other._parentStorageClassName
252 def __hash__(self) -> int:
253 """Hash DatasetType instance.
255 This only uses StorageClass name which is it consistent with the
256 implementation of StorageClass hash method.
257 """
258 return hash((self._name, self._dimensions, self._storageClassName,
259 self._parentStorageClassName))
261 def __lt__(self, other: Any) -> bool:
262 """Sort using the dataset type name."""
263 if not isinstance(other, type(self)):
264 return NotImplemented
265 return self.name < other.name
267 @property
268 def name(self) -> str:
269 """Return a string name for the Dataset.
271 Must correspond to the same `DatasetType` across all Registries.
272 """
273 return self._name
275 @property
276 def dimensions(self) -> DimensionGraph:
277 r"""Return the `Dimension`\ s fir this dataset type.
279 The dimensions label and relate instances of this
280 `DatasetType` (`DimensionGraph`).
281 """
282 return self._dimensions
284 @property
285 def storageClass(self) -> StorageClass:
286 """Return `StorageClass` instance associated with this dataset type.
288 The `StorageClass` defines how this `DatasetType`
289 is persisted. Note that if DatasetType was constructed with a name
290 of a StorageClass then Butler has to be initialized before using
291 this property.
292 """
293 if self._storageClass is None:
294 self._storageClass = StorageClassFactory().getStorageClass(self._storageClassName)
295 return self._storageClass
297 @property
298 def parentStorageClass(self) -> Optional[StorageClass]:
299 """Return the storage class of the composite containing this component.
301 Note that if DatasetType was constructed with a name of a
302 StorageClass then Butler has to be initialized before using this
303 property. Can be `None` if this is not a component of a composite.
304 Must be defined if this is a component.
305 """
306 if self._parentStorageClass is None and self._parentStorageClassName is None:
307 return None
308 if self._parentStorageClass is None and self._parentStorageClassName is not None:
309 self._parentStorageClass = StorageClassFactory().getStorageClass(self._parentStorageClassName)
310 return self._parentStorageClass
312 def isCalibration(self) -> bool:
313 """Return if datasets of this type can be in calibration collections.
315 Returns
316 -------
317 flag : `bool`
318 `True` if datasets of this type may be included in calibration
319 collections.
320 """
321 return self._isCalibration
323 @staticmethod
324 def splitDatasetTypeName(datasetTypeName: str) -> Tuple[str, Optional[str]]:
325 """Return the root name and the component from a composite name.
327 Parameters
328 ----------
329 datasetTypeName : `str`
330 The name of the dataset type, can include a component using
331 a "."-separator.
333 Returns
334 -------
335 rootName : `str`
336 Root name without any components.
337 componentName : `str`
338 The component if it has been specified, else `None`.
340 Notes
341 -----
342 If the dataset type name is ``a.b.c`` this method will return a
343 root name of ``a`` and a component name of ``b.c``.
344 """
345 comp = None
346 root = datasetTypeName
347 if "." in root:
348 # If there is doubt, the component is after the first "."
349 root, comp = root.split(".", maxsplit=1)
350 return root, comp
352 def nameAndComponent(self) -> Tuple[str, Optional[str]]:
353 """Return the root name of this dataset type and any component.
355 Returns
356 -------
357 rootName : `str`
358 Root name for this `DatasetType` without any components.
359 componentName : `str`
360 The component if it has been specified, else `None`.
361 """
362 return self.splitDatasetTypeName(self.name)
364 def component(self) -> Optional[str]:
365 """Return the component name (if defined).
367 Returns
368 -------
369 comp : `str`
370 Name of component part of DatasetType name. `None` if this
371 `DatasetType` is not associated with a component.
372 """
373 _, comp = self.nameAndComponent()
374 return comp
376 def componentTypeName(self, component: str) -> str:
377 """Derive a component dataset type from a composite.
379 Parameters
380 ----------
381 component : `str`
382 Name of component
384 Returns
385 -------
386 derived : `str`
387 Compound name of this `DatasetType` and the component.
389 Raises
390 ------
391 KeyError
392 Requested component is not supported by this `DatasetType`.
393 """
394 if component in self.storageClass.allComponents():
395 return self.nameWithComponent(self.name, component)
396 raise KeyError("Requested component ({}) not understood by this DatasetType".format(component))
398 def makeCompositeDatasetType(self) -> DatasetType:
399 """Return a composite dataset type from the component.
401 Returns
402 -------
403 composite : `DatasetType`
404 The composite dataset type.
406 Raises
407 ------
408 RuntimeError
409 Raised if this dataset type is not a component dataset type.
410 """
411 if not self.isComponent():
412 raise RuntimeError(f"DatasetType {self.name} must be a component to form the composite")
413 composite_name, _ = self.nameAndComponent()
414 if self.parentStorageClass is None:
415 raise ValueError("Parent storage class is not set. "
416 f"Unable to create composite type from {self.name}")
417 return DatasetType(composite_name, dimensions=self.dimensions,
418 storageClass=self.parentStorageClass)
420 def makeComponentDatasetType(self, component: str) -> DatasetType:
421 """Return a component dataset type from a composite.
423 Assumes the same dimensions as the parent.
425 Parameters
426 ----------
427 component : `str`
428 Name of component
430 Returns
431 -------
432 datasetType : `DatasetType`
433 A new DatasetType instance.
434 """
435 # The component could be a read/write or read component
436 return DatasetType(self.componentTypeName(component), dimensions=self.dimensions,
437 storageClass=self.storageClass.allComponents()[component],
438 parentStorageClass=self.storageClass)
440 def makeAllComponentDatasetTypes(self) -> List[DatasetType]:
441 """Return all component dataset types for this composite.
443 Returns
444 -------
445 all : `list` of `DatasetType`
446 All the component dataset types. If this is not a composite
447 then returns an empty list.
448 """
449 return [self.makeComponentDatasetType(componentName)
450 for componentName in self.storageClass.allComponents()]
452 def isComponent(self) -> bool:
453 """Return whether this `DatasetType` refers to a component.
455 Returns
456 -------
457 isComponent : `bool`
458 `True` if this `DatasetType` is a component, `False` otherwise.
459 """
460 if self.component():
461 return True
462 return False
464 def isComposite(self) -> bool:
465 """Return whether this `DatasetType` is a composite.
467 Returns
468 -------
469 isComposite : `bool`
470 `True` if this `DatasetType` is a composite type, `False`
471 otherwise.
472 """
473 return self.storageClass.isComposite()
475 def _lookupNames(self) -> Tuple[LookupKey, ...]:
476 """Return name keys to use for lookups in configurations.
478 The names are returned in order of priority.
480 Returns
481 -------
482 names : `tuple` of `LookupKey`
483 Tuple of the `DatasetType` name and the `StorageClass` name.
484 If the name includes a component the name with the component
485 is first, then the name without the component and finally
486 the storage class name and the storage class name of the
487 composite.
488 """
489 rootName, componentName = self.nameAndComponent()
490 lookups: Tuple[LookupKey, ...] = (LookupKey(name=self.name),)
491 if componentName is not None:
492 lookups = lookups + (LookupKey(name=rootName),)
494 if self.dimensions:
495 # Dimensions are a lower priority than dataset type name
496 lookups = lookups + (LookupKey(dimensions=self.dimensions),)
498 storageClasses = self.storageClass._lookupNames()
499 if componentName is not None and self.parentStorageClass is not None:
500 storageClasses += self.parentStorageClass._lookupNames()
502 return lookups + storageClasses
504 def to_simple(self, minimal: bool = False) -> SerializedDatasetType:
505 """Convert this class to a simple python type.
507 This makes it suitable for serialization.
509 Parameters
510 ----------
511 minimal : `bool`, optional
512 Use minimal serialization. Requires Registry to convert
513 back to a full type.
515 Returns
516 -------
517 simple : `SerializedDatasetType`
518 The object converted to a class suitable for serialization.
519 """
520 as_dict: Dict[str, Any]
521 if minimal:
522 # Only needs the name.
523 as_dict = {"name": self.name}
524 else:
525 # Convert to a dict form
526 as_dict = {"name": self.name,
527 "storageClass": self._storageClassName,
528 "isCalibration": self._isCalibration,
529 "dimensions": self.dimensions.to_simple(),
530 }
532 if self._parentStorageClassName is not None:
533 as_dict["parentStorageClass"] = self._parentStorageClassName
534 return SerializedDatasetType(**as_dict)
536 @classmethod
537 def from_simple(cls, simple: SerializedDatasetType,
538 universe: Optional[DimensionUniverse] = None,
539 registry: Optional[Registry] = None) -> DatasetType:
540 """Construct a new object from the simplified form.
542 This is usually data returned from the `to_simple` method.
544 Parameters
545 ----------
546 simple : `SerializedDatasetType`
547 The value returned by `to_simple()`.
548 universe : `DimensionUniverse`
549 The special graph of all known dimensions of which this graph will
550 be a subset. Can be `None` if a registry is provided.
551 registry : `lsst.daf.butler.Registry`, optional
552 Registry to use to convert simple name of a DatasetType to
553 a full `DatasetType`. Can be `None` if a full description of
554 the type is provided along with a universe.
556 Returns
557 -------
558 datasetType : `DatasetType`
559 Newly-constructed object.
560 """
561 if simple.storageClass is None:
562 # Treat this as minimalist representation
563 if registry is None:
564 raise ValueError(f"Unable to convert a DatasetType name '{simple}' to DatasetType"
565 " without a Registry")
566 return registry.getDatasetType(simple.name)
568 if universe is None and registry is None:
569 raise ValueError("One of universe or registry must be provided.")
571 if universe is None and registry is not None:
572 # registry should not be none by now but test helps mypy
573 universe = registry.dimensions
575 if universe is None:
576 # this is for mypy
577 raise ValueError("Unable to determine a usable universe")
579 if simple.dimensions is None:
580 # mypy hint
581 raise ValueError(f"Dimensions must be specified in {simple}")
583 return cls(name=simple.name,
584 dimensions=DimensionGraph.from_simple(simple.dimensions, universe=universe),
585 storageClass=simple.storageClass,
586 isCalibration=simple.isCalibration,
587 parentStorageClass=simple.parentStorageClass,
588 universe=universe)
590 to_json = to_json_pydantic
591 from_json = classmethod(from_json_pydantic)
593 def __reduce__(self) -> Tuple[Callable, Tuple[Type[DatasetType],
594 Tuple[str, DimensionGraph, str, Optional[str]],
595 Dict[str, bool]]]:
596 """Support pickling.
598 StorageClass instances can not normally be pickled, so we pickle
599 StorageClass name instead of instance.
600 """
601 return _unpickle_via_factory, (self.__class__, (self.name, self.dimensions, self._storageClassName,
602 self._parentStorageClassName),
603 {"isCalibration": self._isCalibration})
605 def __deepcopy__(self, memo: Any) -> DatasetType:
606 """Support for deep copy method.
608 Normally ``deepcopy`` will use pickle mechanism to make copies.
609 We want to avoid that to support (possibly degenerate) use case when
610 DatasetType is constructed with StorageClass instance which is not
611 registered with StorageClassFactory (this happens in unit tests).
612 Instead we re-implement ``__deepcopy__`` method.
613 """
614 return DatasetType(name=deepcopy(self.name, memo),
615 dimensions=deepcopy(self.dimensions, memo),
616 storageClass=deepcopy(self._storageClass or self._storageClassName, memo),
617 parentStorageClass=deepcopy(self._parentStorageClass
618 or self._parentStorageClassName, memo),
619 isCalibration=deepcopy(self._isCalibration, memo))
622def _unpickle_via_factory(factory: Callable, args: Any, kwargs: Any) -> DatasetType:
623 """Unpickle something by calling a factory.
625 Allows subclasses to unpickle using `__reduce__` with keyword
626 arguments as well as positional arguments.
627 """
628 return factory(*args, **kwargs)