Coverage for python / lsst / pipe / base / caching_limited_butler.py: 26%

1# This file is part of pipe_base. 

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/>. 

27 

28from __future__ import annotations 

29 

30__all__ = ["CachingLimitedButler"] 

31 

32import logging 

33from collections.abc import Iterable, Set 

34from typing import Any 

35 

36from lsst.daf.butler import ( 

37 ButlerMetrics, 

38 DatasetId, 

39 DatasetProvenance, 

40 DatasetRef, 

41 DeferredDatasetHandle, 

42 DimensionUniverse, 

43 LimitedButler, 

44 StorageClass, 

45) 

46 

47from ._dataset_handle import InMemoryDatasetHandle 

48 

49_LOG = logging.getLogger(__name__) 

50 

51 

52class CachingLimitedButler(LimitedButler): 

53 """A `LimitedButler` that caches datasets. 

54 

55 A `CachingLimitedButler` caches on both `.put()` and `.get()`, and holds a 

56 single instance of the most recently used dataset type for that put/get. 

57 

58 The dataset types which will be cached on put/get are controlled via the 

59 `cache_on_put` and `cache_on_get` attributes, respectively. 

60 

61 By default, copies of the cached items are returned on `get`, so that code 

62 is free to operate on data in-place. A `no_copy_on_cache` attribute also 

63 exists to tell the `CachingLimitedButler` not to return copies when it is 

64 known that the calling code can be trusted not to change values, e.g. when 

65 passing calibs to `isrTask`. 

66 

67 Parameters 

68 ---------- 

69 wrapped : `LimitedButler` 

70 The butler to wrap. 

71 cache_on_put : `set` [`str`], optional 

72 The dataset types to cache on put. 

73 cache_on_get : `set` [`str`], optional 

74 The dataset types to cache on get. 

75 no_copy_on_cache : `set` [`str`], optional 

76 The dataset types for which to not return copies when cached. 

77 """ 

78 

79 def __init__( 

80 self, 

81 wrapped: LimitedButler, 

82 cache_on_put: Set[str] = frozenset(), 

83 cache_on_get: Set[str] = frozenset(), 

84 no_copy_on_cache: Set[str] = frozenset(), 

85 ): 

86 self._wrapped = wrapped 

87 self.storageClasses = self._wrapped.storageClasses 

88 self._cache_on_put = cache_on_put 

89 self._cache_on_get = cache_on_get 

90 self._cache: dict[str, tuple[DatasetId, InMemoryDatasetHandle]] = {} 

91 self._no_copy_on_cache = no_copy_on_cache 

92 

93 def close(self) -> None: 

94 self._wrapped.close() 

95 

96 @property 

97 def _metrics(self) -> ButlerMetrics: 

98 # Need to always forward from the wrapped metrics object. 

99 return self._wrapped._metrics 

100 

101 @_metrics.setter 

102 def _metrics(self, metrics: ButlerMetrics) -> None: 

103 # Allow record_metrics() context manager to override the wrapped 

104 # butler. 

105 self._wrapped._metrics = metrics 

106 

107 def get( 

108 self, 

109 ref: DatasetRef, 

110 /, 

111 *, 

112 parameters: dict[str, Any] | None = None, 

113 storageClass: StorageClass | str | None = None, 

114 ) -> Any: 

115 if storageClass is None: 

116 storageClass = ref.datasetType.storageClass 

117 elif isinstance(storageClass, str): 

118 storageClass = self.storageClasses.getStorageClass(storageClass) 

119 

120 # check if we have this dataset type in the cache 

121 if cached := self._cache.get(ref.datasetType.name): 

122 dataset_id, handle = cached 

123 if dataset_id == ref.id: # if we do, check it's the right object 

124 _LOG.debug("Returning cached dataset %s", ref) 

125 return handle.get(parameters=parameters, storageClass=storageClass) 

126 

127 obj = self._wrapped.get(ref, parameters=parameters, storageClass=storageClass) 

128 if ref.datasetType.name in self._cache_on_get and not parameters: 

129 handle = InMemoryDatasetHandle( 

130 obj, 

131 storageClass=storageClass, 

132 dataId=ref.dataId, 

133 copy=ref.datasetType.name not in self._no_copy_on_cache, 

134 ) 

135 # and not parameters is to make sure we don't cache sub-images etc 

136 self._cache[ref.datasetType.name] = (ref.id, handle) 

137 _LOG.debug("Cached dataset %s", ref) 

138 # make sure copy fires if needed 

139 return handle.get() 

140 return obj 

141 

142 def getDeferred( 

143 self, 

144 ref: DatasetRef, 

145 /, 

146 *, 

147 parameters: dict[str, Any] | None = None, 

148 storageClass: str | StorageClass | None = None, 

149 ) -> DeferredDatasetHandle: 

150 # note that this does not use the cache at all 

151 return self._wrapped.getDeferred(ref, parameters=parameters, storageClass=storageClass) 

152 

153 def stored_many(self, refs: Iterable[DatasetRef]) -> dict[DatasetRef, bool]: 

154 result = {} 

155 unknown_refs = [] 

156 for ref in refs: 

157 if cached := self._cache.get(ref.datasetType.name): 

158 dataset_id, _ = cached 

159 if dataset_id == ref.id: 

160 result[ref] = True 

161 continue 

162 unknown_refs.append(ref) 

163 

164 result.update(self._wrapped.stored_many(unknown_refs)) 

165 return result 

166 

167 def isWriteable(self) -> bool: 

168 return self._wrapped.isWriteable() 

169 

170 def put(self, obj: Any, ref: DatasetRef, /, *, provenance: DatasetProvenance | None = None) -> DatasetRef: 

171 if ref.datasetType.name in self._cache_on_put: 

172 self._cache[ref.datasetType.name] = ( 

173 ref.id, 

174 InMemoryDatasetHandle( 

175 obj, 

176 storageClass=ref.datasetType.storageClass, 

177 dataId=ref.dataId, 

178 copy=ref.datasetType.name not in self._no_copy_on_cache, 

179 ), 

180 ) 

181 _LOG.debug("Cached dataset %s on put", ref) 

182 return self._wrapped.put(obj, ref, provenance=provenance) 

183 

184 def pruneDatasets( 

185 self, 

186 refs: Iterable[DatasetRef], 

187 *, 

188 disassociate: bool = True, 

189 unstore: bool = False, 

190 tags: Iterable[str] = (), 

191 purge: bool = False, 

192 ) -> None: 

193 refs = list(refs) 

194 for ref in refs: 

195 if cached := self._cache.get(ref.datasetType.name): 

196 dataset_id, _ = cached 

197 if dataset_id == ref.id: 

198 del self._cache[ref.datasetType.name] 

199 

200 return self._wrapped.pruneDatasets( 

201 refs, disassociate=disassociate, unstore=unstore, tags=tags, purge=purge 

202 ) 

203 

204 @property 

205 def dimensions(self) -> DimensionUniverse: 

206 return self._wrapped.dimensions 

207 

208 @property 

209 def _datastore(self) -> Any: 

210 return self._wrapped._datastore 

211 

212 @_datastore.setter # demanded by MyPy since we declare it to be an instance attribute in LimitedButler. 

213 def _datastore(self, value: Any) -> None: 

214 self._wrapped._datastore = value