Coverage for python/lsst/daf/butler/registry/dimensions/query.py: 81%

66 statements  

« prev     ^ index     » next       coverage.py v6.4.2, created at 2022-07-14 22:48 +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/>. 

21from __future__ import annotations 

22 

23__all__ = ["QueryDimensionRecordStorage"] 

24 

25from typing import Any, Iterable, Mapping, Optional 

26 

27import sqlalchemy 

28 

29from ...core import ( 

30 DatabaseDimension, 

31 DatabaseDimensionElement, 

32 DataCoordinateIterable, 

33 DimensionElement, 

34 DimensionRecord, 

35 GovernorDimension, 

36 NamedKeyDict, 

37 NamedKeyMapping, 

38 SpatialRegionDatabaseRepresentation, 

39 TimespanDatabaseRepresentation, 

40) 

41from ..interfaces import ( 

42 Database, 

43 DatabaseDimensionRecordStorage, 

44 GovernorDimensionRecordStorage, 

45 StaticTablesContext, 

46) 

47from ..queries import QueryBuilder 

48 

49 

50class QueryDimensionRecordStorage(DatabaseDimensionRecordStorage): 

51 """A read-only record storage implementation backed by SELECT query. 

52 

53 At present, the only query this class supports is a SELECT DISTINCT over 

54 the table for some other dimension that has this dimension as an implied 

55 dependency. For example, we can use this class to provide access to the 

56 set of ``band`` names referenced by any ``physical_filter``. 

57 

58 Parameters 

59 ---------- 

60 db : `Database` 

61 Interface to the database engine and namespace that will hold these 

62 dimension records. 

63 element : `DatabaseDimensionElement` 

64 The element whose records this storage will manage. 

65 """ 

66 

67 def __init__(self, db: Database, element: DatabaseDimensionElement, viewOf: str): 

68 assert isinstance( 

69 element, DatabaseDimension 

70 ), "An element cannot be a dependency unless it is a dimension." 

71 self._db = db 

72 self._element = element 

73 self._target = element.universe[viewOf] 

74 self._targetSpec = self._target.RecordClass.fields.makeTableSpec( 

75 RegionReprClass=self._db.getSpatialRegionRepresentation(), 

76 TimespanReprClass=self._db.getTimespanRepresentation(), 

77 ) 

78 self._viewOf = viewOf 

79 self._query = None # Constructed on first use. 

80 if element not in self._target.graph.dimensions: 80 ↛ 81line 80 didn't jump to line 81, because the condition on line 80 was never true

81 raise NotImplementedError("Query-backed dimension must be a dependency of its target.") 

82 if element.metadata: 82 ↛ 83line 82 didn't jump to line 83, because the condition on line 82 was never true

83 raise NotImplementedError("Cannot use query to back dimension with metadata.") 

84 if element.implied: 84 ↛ 85line 84 didn't jump to line 85, because the condition on line 84 was never true

85 raise NotImplementedError("Cannot use query to back dimension with implied dependencies.") 

86 if element.alternateKeys: 86 ↛ 87line 86 didn't jump to line 87, because the condition on line 86 was never true

87 raise NotImplementedError("Cannot use query to back dimension with alternate unique keys.") 

88 if element.spatial is not None: 88 ↛ 89line 88 didn't jump to line 89, because the condition on line 88 was never true

89 raise NotImplementedError("Cannot use query to back spatial dimension.") 

90 if element.temporal is not None: 90 ↛ 91line 90 didn't jump to line 91, because the condition on line 90 was never true

91 raise NotImplementedError("Cannot use query to back temporal dimension.") 

92 

93 @classmethod 

94 def initialize( 

95 cls, 

96 db: Database, 

97 element: DatabaseDimensionElement, 

98 *, 

99 context: Optional[StaticTablesContext] = None, 

100 config: Mapping[str, Any], 

101 governors: NamedKeyMapping[GovernorDimension, GovernorDimensionRecordStorage], 

102 ) -> DatabaseDimensionRecordStorage: 

103 # Docstring inherited from DatabaseDimensionRecordStorage. 

104 viewOf = config["view_of"] 

105 return cls(db, element, viewOf) 

106 

107 @property 

108 def element(self) -> DatabaseDimension: 

109 # Docstring inherited from DimensionRecordStorage.element. 

110 return self._element 

111 

112 def clearCaches(self) -> None: 

113 # Docstring inherited from DimensionRecordStorage.clearCaches. 

114 pass 

115 

116 def _ensureQuery(self) -> None: 

117 if self._query is None: 

118 targetTable = self._db.getExistingTable(self._target.name, self._targetSpec) 

119 assert targetTable is not None 

120 columns = [] 

121 # The only columns for this dimension are ones for its required 

122 # dependencies and its own primary key (guaranteed by the checks in 

123 # the ctor). 

124 for dimension in self.element.required: 

125 if dimension == self.element: 125 ↛ 128line 125 didn't jump to line 128, because the condition on line 125 was never false

126 columns.append(targetTable.columns[dimension.name].label(dimension.primaryKey.name)) 

127 else: 

128 columns.append(targetTable.columns[dimension.name].label(dimension.name)) 

129 # This query doesn't do a SELECT DISTINCT, because that's confusing 

130 # and potentially wasteful if we apply a restrictive WHERE clause, 

131 # as SelectableDimensionRecordStorage.fetch will do. 

132 # Instead, we add DISTINCT in join() only. 

133 self._query = ( 

134 sqlalchemy.sql.select(*columns).distinct().select_from(targetTable).alias(self.element.name) 

135 ) 

136 

137 def join( 

138 self, 

139 builder: QueryBuilder, 

140 *, 

141 regions: Optional[NamedKeyDict[DimensionElement, SpatialRegionDatabaseRepresentation]] = None, 

142 timespans: Optional[NamedKeyDict[DimensionElement, TimespanDatabaseRepresentation]] = None, 

143 ) -> None: 

144 # Docstring inherited from DimensionRecordStorage. 

145 assert regions is None, "Should be guaranteed by constructor checks." 

146 assert timespans is None, "Should be guaranteed by constructor checks." 

147 if self._target in builder.summary.mustHaveKeysJoined: 147 ↛ 152line 147 didn't jump to line 152, because the condition on line 147 was never true

148 # Do nothing; the target dimension is already being included, so 

149 # joining against a subquery referencing it would just produce a 

150 # more complicated query that's guaranteed to return the same 

151 # results. 

152 return 

153 self._ensureQuery() 

154 joinOn = builder.startJoin( 

155 self._query, self.element.required, self.element.RecordClass.fields.required.names 

156 ) 

157 builder.finishJoin(self._query, joinOn) 

158 return self._query 

159 

160 def insert(self, *records: DimensionRecord, replace: bool = False, skip_existing: bool = False) -> None: 

161 # Docstring inherited from DimensionRecordStorage.insert. 

162 raise TypeError( 

163 f"Cannot insert {self.element.name} records, define as part of {self._viewOf} instead." 

164 ) 

165 

166 def sync(self, record: DimensionRecord, update: bool = False) -> bool: 

167 # Docstring inherited from DimensionRecordStorage.sync. 

168 raise TypeError(f"Cannot sync {self.element.name} records, define as part of {self._viewOf} instead.") 

169 

170 def fetch(self, dataIds: DataCoordinateIterable) -> Iterable[DimensionRecord]: 

171 # Docstring inherited from DimensionRecordStorage.fetch. 

172 RecordClass = self.element.RecordClass 

173 for dataId in dataIds: 

174 # Given the restrictions imposed at construction, we know there's 

175 # nothing to actually fetch: everything we need is in the data ID. 

176 yield RecordClass(**dataId.byName()) 

177 

178 def digestTables(self) -> Iterable[sqlalchemy.schema.Table]: 

179 # Docstring inherited from DimensionRecordStorage.digestTables. 

180 return []