Coverage for python/lsst/daf/butler/core/datasets/type.py : 18%

Hot-keys 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"]
26from copy import deepcopy
27import re
29from types import MappingProxyType
31from typing import (
32 TYPE_CHECKING,
33 Any,
34 Iterable,
35 List,
36 Mapping,
37 Optional,
38 Tuple,
39 Type,
40 Union,
41)
44from ..storageClass import StorageClass, StorageClassFactory
45from ..dimensions import DimensionGraph
46from ..configSupport import LookupKey
48if TYPE_CHECKING: 48 ↛ 49line 48 didn't jump to line 49, because the condition on line 48 was never true
49 from ..dimensions import Dimension, DimensionUniverse
52def _safeMakeMappingProxyType(data: Optional[Mapping]) -> Mapping:
53 if data is None:
54 data = {}
55 return MappingProxyType(data)
58class DatasetType:
59 r"""A named category of Datasets that defines how they are organized,
60 related, and stored.
62 A concrete, final class whose instances represent `DatasetType`\ s.
63 `DatasetType` instances may be constructed without a `Registry`,
64 but they must be registered
65 via `Registry.registerDatasetType()` before corresponding Datasets
66 may be added.
67 `DatasetType` instances are immutable.
69 Parameters
70 ----------
71 name : `str`
72 A string name for the Dataset; must correspond to the same
73 `DatasetType` across all Registries. Names must start with an
74 upper or lowercase letter, and may contain only letters, numbers,
75 and underscores. Component dataset types should contain a single
76 period separating the base dataset type name from the component name
77 (and may be recursive).
78 dimensions : `DimensionGraph` or iterable of `Dimension`
79 Dimensions used to label and relate instances of this `DatasetType`.
80 If not a `DimensionGraph`, ``universe`` must be provided as well.
81 storageClass : `StorageClass` or `str`
82 Instance of a `StorageClass` or name of `StorageClass` that defines
83 how this `DatasetType` is persisted.
84 parentStorageClass : `StorageClass` or `str`, optional
85 Instance of a `StorageClass` or name of `StorageClass` that defines
86 how the composite parent is persisted. Must be `None` if this
87 is not a component. Mandatory if it is a component but can be the
88 special temporary placeholder
89 (`DatasetType.PlaceholderParentStorageClass`) to allow
90 construction with an intent to finalize later.
91 universe : `DimensionUniverse`, optional
92 Set of all known dimensions, used to normalize ``dimensions`` if it
93 is not already a `DimensionGraph`.
94 isCalibration : `bool`, optional
95 If `True`, this dataset type may be included in
96 `~CollectionType.CALIBRATION` collections.
97 """
99 __slots__ = ("_name", "_dimensions", "_storageClass", "_storageClassName",
100 "_parentStorageClass", "_parentStorageClassName",
101 "_isCalibration")
103 VALID_NAME_REGEX = re.compile("^[a-zA-Z][a-zA-Z0-9_]*(\\.[a-zA-Z][a-zA-Z0-9_]*)*$")
105 PlaceholderParentStorageClass = StorageClass("PlaceHolder")
106 """Placeholder StorageClass that can be used temporarily for a
107 component.
109 This can be useful in pipeline construction where we are creating
110 dataset types without a registry.
111 """
113 @staticmethod
114 def nameWithComponent(datasetTypeName: str, componentName: str) -> str:
115 """Form a valid DatasetTypeName from a parent and component.
117 No validation is performed.
119 Parameters
120 ----------
121 datasetTypeName : `str`
122 Base type name.
123 componentName : `str`
124 Name of component.
126 Returns
127 -------
128 compTypeName : `str`
129 Name to use for component DatasetType.
130 """
131 return "{}.{}".format(datasetTypeName, componentName)
133 def __init__(self, name: str, dimensions: Union[DimensionGraph, Iterable[Dimension]],
134 storageClass: Union[StorageClass, str],
135 parentStorageClass: Optional[Union[StorageClass, str]] = None, *,
136 universe: Optional[DimensionUniverse] = None,
137 isCalibration: bool = False):
138 if self.VALID_NAME_REGEX.match(name) is None:
139 raise ValueError(f"DatasetType name '{name}' is invalid.")
140 self._name = name
141 if not isinstance(dimensions, DimensionGraph):
142 if universe is None:
143 raise ValueError("If dimensions is not a normalized DimensionGraph, "
144 "a universe must be provided.")
145 dimensions = universe.extract(dimensions)
146 self._dimensions = dimensions
147 if not isinstance(storageClass, (StorageClass, str)):
148 raise ValueError("StorageClass argument must be StorageClass or str. "
149 f"Got {storageClass}")
150 self._storageClass: Optional[StorageClass]
151 if isinstance(storageClass, StorageClass):
152 self._storageClass = storageClass
153 self._storageClassName = storageClass.name
154 else:
155 self._storageClass = None
156 self._storageClassName = storageClass
158 self._parentStorageClass: Optional[StorageClass] = None
159 self._parentStorageClassName: Optional[str] = None
160 if parentStorageClass is not None:
161 if not isinstance(storageClass, (StorageClass, str)):
162 raise ValueError("Parent StorageClass argument must be StorageClass or str. "
163 f"Got {parentStorageClass}")
165 # Only allowed for a component dataset type
166 _, componentName = self.splitDatasetTypeName(self._name)
167 if componentName is None:
168 raise ValueError("Can not specify a parent storage class if this is not a component"
169 f" ({self._name})")
170 if isinstance(parentStorageClass, StorageClass):
171 self._parentStorageClass = parentStorageClass
172 self._parentStorageClassName = parentStorageClass.name
173 else:
174 self._parentStorageClassName = parentStorageClass
176 # Ensure that parent storage class is specified when we have
177 # a component and is not specified when we don't
178 _, componentName = self.splitDatasetTypeName(self._name)
179 if parentStorageClass is None and componentName is not None:
180 raise ValueError(f"Component dataset type '{self._name}' constructed without parent"
181 " storage class")
182 if parentStorageClass is not None and componentName is None:
183 raise ValueError(f"Parent storage class specified by {self._name} is not a composite")
184 self._isCalibration = isCalibration
186 def __repr__(self) -> str:
187 extra = ""
188 if self._parentStorageClassName:
189 extra = f", parentStorageClass={self._parentStorageClassName}"
190 if self._isCalibration:
191 extra += ", isCalibration=True"
192 return f"DatasetType({self.name!r}, {self.dimensions}, {self._storageClassName}{extra})"
194 def __eq__(self, other: Any) -> bool:
195 if not isinstance(other, type(self)):
196 return False
197 if self._name != other._name:
198 return False
199 if self._dimensions != other._dimensions:
200 return False
201 if self._storageClass is not None and other._storageClass is not None:
202 if self._storageClass != other._storageClass:
203 return False
204 else:
205 if self._storageClassName != other._storageClassName:
206 return False
207 if self._isCalibration != other._isCalibration:
208 return False
209 if self._parentStorageClass is not None and other._parentStorageClass is not None:
210 return self._parentStorageClass == other._parentStorageClass
211 else:
212 return self._parentStorageClassName == other._parentStorageClassName
214 def __hash__(self) -> int:
215 """Hash DatasetType instance.
217 This only uses StorageClass name which is it consistent with the
218 implementation of StorageClass hash method.
219 """
220 return hash((self._name, self._dimensions, self._storageClassName,
221 self._parentStorageClassName))
223 @property
224 def name(self) -> str:
225 """A string name for the Dataset; must correspond to the same
226 `DatasetType` across all Registries.
227 """
228 return self._name
230 @property
231 def dimensions(self) -> DimensionGraph:
232 r"""The `Dimension`\ s that label and relate instances of this
233 `DatasetType` (`DimensionGraph`).
234 """
235 return self._dimensions
237 @property
238 def storageClass(self) -> StorageClass:
239 """`StorageClass` instance that defines how this `DatasetType`
240 is persisted. Note that if DatasetType was constructed with a name
241 of a StorageClass then Butler has to be initialized before using
242 this property.
243 """
244 if self._storageClass is None:
245 self._storageClass = StorageClassFactory().getStorageClass(self._storageClassName)
246 return self._storageClass
248 @property
249 def parentStorageClass(self) -> Optional[StorageClass]:
250 """`StorageClass` instance that defines how the composite associated
251 with this `DatasetType` is persisted.
253 Note that if DatasetType was constructed with a name of a
254 StorageClass then Butler has to be initialized before using this
255 property. Can be `None` if this is not a component of a composite.
256 Must be defined if this is a component.
257 """
258 if self._parentStorageClass is None and self._parentStorageClassName is None:
259 return None
260 if self._parentStorageClass is None and self._parentStorageClassName is not None:
261 self._parentStorageClass = StorageClassFactory().getStorageClass(self._parentStorageClassName)
262 return self._parentStorageClass
264 def isCalibration(self) -> bool:
265 """Return whether datasets of this type may be included in calibration
266 collections.
268 Returns
269 -------
270 flag : `bool`
271 `True` if datasets of this type may be included in calibration
272 collections.
273 """
274 return self._isCalibration
276 def finalizeParentStorageClass(self, newParent: StorageClass) -> None:
277 """Replace the current placeholder parent storage class with
278 the real parent.
280 Parameters
281 ----------
282 newParent : `StorageClass`
283 The new parent to be associated with this composite dataset
284 type. This replaces the temporary placeholder parent that
285 was specified during construction.
287 Raises
288 ------
289 ValueError
290 Raised if this dataset type is not a component of a composite.
291 Raised if a StorageClass is not given.
292 Raised if the parent currently associated with the dataset
293 type is not a placeholder.
294 """
295 if not self.isComponent():
296 raise ValueError("Can not set a parent storage class if this is not a component"
297 f" ({self.name})")
298 if self._parentStorageClass != self.PlaceholderParentStorageClass:
299 raise ValueError(f"This DatasetType has a parent of {self._parentStorageClassName} and"
300 " is not a placeholder.")
301 if not isinstance(newParent, StorageClass):
302 raise ValueError(f"Supplied parent must be a StorageClass. Got {newParent!r}")
303 self._parentStorageClass = newParent
304 self._parentStorageClassName = newParent.name
306 @staticmethod
307 def splitDatasetTypeName(datasetTypeName: str) -> Tuple[str, Optional[str]]:
308 """Given a dataset type name, return the root name and the component
309 name.
311 Parameters
312 ----------
313 datasetTypeName : `str`
314 The name of the dataset type, can include a component using
315 a "."-separator.
317 Returns
318 -------
319 rootName : `str`
320 Root name without any components.
321 componentName : `str`
322 The component if it has been specified, else `None`.
324 Notes
325 -----
326 If the dataset type name is ``a.b.c`` this method will return a
327 root name of ``a`` and a component name of ``b.c``.
328 """
329 comp = None
330 root = datasetTypeName
331 if "." in root:
332 # If there is doubt, the component is after the first "."
333 root, comp = root.split(".", maxsplit=1)
334 return root, comp
336 def nameAndComponent(self) -> Tuple[str, Optional[str]]:
337 """Return the root name of this dataset type and the component
338 name (if defined).
340 Returns
341 -------
342 rootName : `str`
343 Root name for this `DatasetType` without any components.
344 componentName : `str`
345 The component if it has been specified, else `None`.
346 """
347 return self.splitDatasetTypeName(self.name)
349 def component(self) -> Optional[str]:
350 """Component name (if defined)
352 Returns
353 -------
354 comp : `str`
355 Name of component part of DatasetType name. `None` if this
356 `DatasetType` is not associated with a component.
357 """
358 _, comp = self.nameAndComponent()
359 return comp
361 def componentTypeName(self, component: str) -> str:
362 """Given a component name, derive the datasetTypeName of that component
364 Parameters
365 ----------
366 component : `str`
367 Name of component
369 Returns
370 -------
371 derived : `str`
372 Compound name of this `DatasetType` and the component.
374 Raises
375 ------
376 KeyError
377 Requested component is not supported by this `DatasetType`.
378 """
379 if component in self.storageClass.allComponents():
380 return self.nameWithComponent(self.name, component)
381 raise KeyError("Requested component ({}) not understood by this DatasetType".format(component))
383 def makeComponentDatasetType(self, component: str) -> DatasetType:
384 """Return a DatasetType suitable for the given component, assuming the
385 same dimensions as the parent.
387 Parameters
388 ----------
389 component : `str`
390 Name of component
392 Returns
393 -------
394 datasetType : `DatasetType`
395 A new DatasetType instance.
396 """
397 # The component could be a read/write or read component
398 return DatasetType(self.componentTypeName(component), dimensions=self.dimensions,
399 storageClass=self.storageClass.allComponents()[component],
400 parentStorageClass=self.storageClass)
402 def makeAllComponentDatasetTypes(self) -> List[DatasetType]:
403 """Return all the component dataset types assocaited with this
404 dataset type.
406 Returns
407 -------
408 all : `list` of `DatasetType`
409 All the component dataset types. If this is not a composite
410 then returns an empty list.
411 """
412 return [self.makeComponentDatasetType(componentName)
413 for componentName in self.storageClass.allComponents()]
415 def isComponent(self) -> bool:
416 """Boolean indicating whether this `DatasetType` refers to a
417 component of a composite.
419 Returns
420 -------
421 isComponent : `bool`
422 `True` if this `DatasetType` is a component, `False` otherwise.
423 """
424 if self.component():
425 return True
426 return False
428 def isComposite(self) -> bool:
429 """Boolean indicating whether this `DatasetType` is a composite type.
431 Returns
432 -------
433 isComposite : `bool`
434 `True` if this `DatasetType` is a composite type, `False`
435 otherwise.
436 """
437 return self.storageClass.isComposite()
439 def _lookupNames(self) -> Tuple[LookupKey, ...]:
440 """Name keys to use when looking up this datasetType in a
441 configuration.
443 The names are returned in order of priority.
445 Returns
446 -------
447 names : `tuple` of `LookupKey`
448 Tuple of the `DatasetType` name and the `StorageClass` name.
449 If the name includes a component the name with the component
450 is first, then the name without the component and finally
451 the storage class name.
452 """
453 rootName, componentName = self.nameAndComponent()
454 lookups: Tuple[LookupKey, ...] = (LookupKey(name=self.name),)
455 if componentName is not None:
456 lookups = lookups + (LookupKey(name=rootName),)
458 if self.dimensions:
459 # Dimensions are a lower priority than dataset type name
460 lookups = lookups + (LookupKey(dimensions=self.dimensions),)
462 return lookups + self.storageClass._lookupNames()
464 def __reduce__(self) -> Tuple[Type[DatasetType], Tuple[str, DimensionGraph, str, Optional[str]]]:
465 """Support pickling.
467 StorageClass instances can not normally be pickled, so we pickle
468 StorageClass name instead of instance.
469 """
470 return (DatasetType, (self.name, self.dimensions, self._storageClassName,
471 self._parentStorageClassName))
473 def __deepcopy__(self, memo: Any) -> DatasetType:
474 """Support for deep copy method.
476 Normally ``deepcopy`` will use pickle mechanism to make copies.
477 We want to avoid that to support (possibly degenerate) use case when
478 DatasetType is constructed with StorageClass instance which is not
479 registered with StorageClassFactory (this happens in unit tests).
480 Instead we re-implement ``__deepcopy__`` method.
481 """
482 return DatasetType(name=deepcopy(self.name, memo),
483 dimensions=deepcopy(self.dimensions, memo),
484 storageClass=deepcopy(self._storageClass or self._storageClassName, memo),
485 parentStorageClass=deepcopy(self._parentStorageClass
486 or self._parentStorageClassName, memo))