Coverage for python/lsst/daf/butler/registry/queries/expressions/categorize.py: 9%

88 statements  

« prev     ^ index     » next       coverage.py v6.4.2, created at 2022-07-16 02:09 -0700

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__ = () # all symbols intentionally private; for internal package use. 

24 

25import enum 

26from typing import Optional, Tuple 

27 

28from ....core import Dimension, DimensionElement, DimensionGraph, DimensionUniverse 

29 

30 

31class ExpressionConstant(enum.Enum): 

32 """Enumeration for constants recognized in all expressions.""" 

33 

34 NULL = "null" 

35 INGEST_DATE = "ingest_date" 

36 

37 

38def categorizeConstant(name: str) -> Optional[ExpressionConstant]: 

39 """Categorize an identifier in a parsed expression as one of a few global 

40 constants. 

41 

42 Parameters 

43 ---------- 

44 name : `str` 

45 Identifier to categorize. Case-insensitive. 

46 

47 Returns 

48 ------- 

49 categorized : `ExpressionConstant` or `None` 

50 Enumeration value if the string represents a constant, `None` 

51 otherwise. 

52 """ 

53 try: 

54 return ExpressionConstant(name.lower()) 

55 except ValueError: 

56 return None 

57 

58 

59def categorizeElementId(universe: DimensionUniverse, name: str) -> Tuple[DimensionElement, Optional[str]]: 

60 """Categorize an identifier in a parsed expression as either a `Dimension` 

61 name (indicating the primary key for that dimension) or a non-primary-key 

62 column in a `DimensionElement` table. 

63 

64 Parameters 

65 ---------- 

66 universe : `DimensionUniverse` 

67 All known dimensions. 

68 name : `str` 

69 Identifier to categorize. 

70 

71 Returns 

72 ------- 

73 element : `DimensionElement` 

74 The `DimensionElement` the identifier refers to. 

75 column : `str` or `None` 

76 The name of a column in the table for ``element``, or `None` if 

77 ``element`` is a `Dimension` and the requested column is its primary 

78 key. 

79 

80 Raises 

81 ------ 

82 LookupError 

83 Raised if the identifier refers to a nonexistent `DimensionElement` 

84 or column. 

85 RuntimeError 

86 Raised if the expression refers to a primary key in an illegal way. 

87 This exception includes a suggestion for how to rewrite the expression, 

88 so at least its message should generally be propagated up to a context 

89 where the error can be interpreted by a human. 

90 """ 

91 table, sep, column = name.partition(".") 

92 if column: 

93 try: 

94 element = universe[table] 

95 except KeyError as err: 

96 raise LookupError(f"No dimension element with name '{table}'.") from err 

97 if isinstance(element, Dimension) and column == element.primaryKey.name: 

98 # Allow e.g. "visit.id = x" instead of just "visit = x"; this 

99 # can be clearer. 

100 return element, None 

101 elif column in element.graph.names: 

102 # User said something like "patch.tract = x" or 

103 # "tract.tract = x" instead of just "tract = x" or 

104 # "tract.id = x", which is at least needlessly confusing and 

105 # possibly not actually a column name, though we can guess 

106 # what they were trying to do. 

107 # Encourage them to clean that up and try again. 

108 name = universe[column].primaryKey.name # type: ignore 

109 raise RuntimeError( 

110 f"Invalid reference to '{table}.{column}' " 

111 f"in expression; please use '{column}' or " 

112 f"'{column}.{name}' instead." 

113 ) 

114 else: 

115 return element, column 

116 else: 

117 try: 

118 dimension = universe[table] 

119 except KeyError as err: 

120 raise LookupError(f"No dimension with name '{table}'.") from err 

121 return dimension, None 

122 

123 

124def categorizeOrderByName(graph: DimensionGraph, name: str) -> Tuple[DimensionElement, Optional[str]]: 

125 """Categorize an identifier in an ORDER BY clause. 

126 

127 Parameters 

128 ---------- 

129 graph : `DimensionGraph` 

130 All known dimensions. 

131 name : `str` 

132 Identifier to categorize. 

133 

134 Returns 

135 ------- 

136 element : `DimensionElement` 

137 The `DimensionElement` the identifier refers to. 

138 column : `str` or `None` 

139 The name of a column in the table for ``element``, or `None` if 

140 ``element`` is a `Dimension` and the requested column is its primary 

141 key. 

142 

143 Raises 

144 ------ 

145 ValueError 

146 Raised if element name is not found in a graph, metadata name is not 

147 recognized, or if there is more than one element has specified 

148 metadata. 

149 

150 Notes 

151 ----- 

152 For ORDER BY identifiers we use slightly different set of rules compared to 

153 the rules in `categorizeElementId`: 

154 

155 - Name can be a dimension element name. e.g. ``visit``. 

156 - Name can be an element name and a metadata name (or key name) separated 

157 by dot, e.g. ``detector.full_name``. 

158 - Name can be a metadata name without element name prefix, e.g. 

159 ``day_obs``; in that case metadata (or key) is searched in all elements 

160 present in a graph. Exception is raised if name appears in more than one 

161 element. 

162 - Two special identifiers ``timespan.begin`` and ``timespan.end`` can be 

163 used with temporal elements, if element name is not given then a temporal 

164 element from a graph is used. 

165 """ 

166 element: DimensionElement 

167 field_name: Optional[str] = None 

168 if name in ("timespan.begin", "timespan.end"): 

169 matches = [element for element in graph.elements if element.temporal] 

170 if len(matches) == 1: 

171 element = matches[0] 

172 field_name = name 

173 elif len(matches) > 1: 

174 raise ValueError( 

175 f"Timespan exists in more than one dimesion element: {matches}," 

176 " qualify timespan with specific dimension name." 

177 ) 

178 else: 

179 raise ValueError(f"Cannot find any temporal dimension element for '{name}'.") 

180 elif "." not in name: 

181 # No dot, can be either a dimension name or a field name (in any of 

182 # the known elements) 

183 if name in graph.elements.names: 

184 element = graph.elements[name] 

185 else: 

186 # Can be a metadata name or any of unique keys 

187 matches = [elem for elem in graph.elements if name in elem.metadata.names] 

188 matches += [dim for dim in graph if name in dim.uniqueKeys.names] 

189 if len(matches) == 1: 

190 element = matches[0] 

191 field_name = name 

192 elif len(matches) > 1: 

193 raise ValueError( 

194 f"Metadata '{name}' exists in more than one dimension element: {matches}," 

195 " qualify metadata name with dimension name." 

196 ) 

197 else: 

198 raise ValueError(f"Metadata '{name}' cannot be found in any dimension.") 

199 else: 

200 # qualified name, must be a dimension element and a field 

201 elem_name, _, field_name = name.partition(".") 

202 if elem_name not in graph.elements.names: 

203 raise ValueError(f"Unknown dimension element name '{elem_name}'") 

204 element = graph.elements[elem_name] 

205 if field_name in ("timespan.begin", "timespan.end"): 

206 if not element.temporal: 

207 raise ValueError(f"Cannot use '{field_name}' with non-temporal element '{element}'.") 

208 elif isinstance(element, Dimension) and field_name == element.primaryKey.name: 

209 # Primary key is optional 

210 field_name = None 

211 else: 

212 if not ( 

213 field_name in element.metadata.names 

214 or (isinstance(element, Dimension) and field_name in element.alternateKeys.names) 

215 ): 

216 raise ValueError(f"Field '{field_name}' does not exist in '{element}'.") 

217 

218 return element, field_name 

219 

220 

221def categorizeElementOrderByName(element: DimensionElement, name: str) -> Optional[str]: 

222 """Categorize an identifier in an ORDER BY clause for a single element. 

223 

224 Parameters 

225 ---------- 

226 element : `DimensionElement` 

227 Dimension element. 

228 name : `str` 

229 Identifier to categorize. 

230 

231 Returns 

232 ------- 

233 column : `str` or `None` 

234 The name of a column in the table for ``element``, or `None` if 

235 ``element`` is a `Dimension` and the requested column is its primary 

236 key. 

237 

238 Raises 

239 ------ 

240 ValueError 

241 Raised if name is not recognized. 

242 

243 Notes 

244 ----- 

245 For ORDER BY identifiers we use slightly different set of rules compared to 

246 the rules in `categorizeElementId`: 

247 

248 - Name can be a dimension element name. e.g. ``visit``. 

249 - Name can be an element name and a metadata name (or key name) separated 

250 by dot, e.g. ``detector.full_name``, element name must correspond to 

251 ``element`` argument 

252 - Name can be a metadata name without element name prefix, e.g. 

253 ``day_obs``. 

254 - Two special identifiers ``timespan.begin`` and ``timespan.end`` can be 

255 used with temporal elements. 

256 """ 

257 field_name: Optional[str] = None 

258 if name in ("timespan.begin", "timespan.end"): 

259 if element.temporal: 

260 field_name = name 

261 else: 

262 raise ValueError(f"Cannot use '{field_name}' with non-temporal element '{element}'.") 

263 elif "." not in name: 

264 # No dot, can be either a dimension name or a field name (in any of 

265 # the known elements) 

266 if name == element.name: 

267 # Must be a dimension element 

268 if not isinstance(element, Dimension): 

269 raise ValueError(f"Element '{element}' is not a dimension.") 

270 else: 

271 # Can be a metadata name or any of the keys 

272 if name in element.metadata.names or ( 

273 isinstance(element, Dimension) and name in element.uniqueKeys.names 

274 ): 

275 field_name = name 

276 else: 

277 raise ValueError(f"Field '{name}' does not exist in '{element}'.") 

278 else: 

279 # qualified name, must be a dimension element and a field 

280 elem_name, _, field_name = name.partition(".") 

281 if elem_name != element.name: 

282 raise ValueError(f"Element name mismatch: '{elem_name}' instead of '{element}'") 

283 if field_name in ("timespan.begin", "timespan.end"): 

284 if not element.temporal: 

285 raise ValueError(f"Cannot use '{field_name}' with non-temporal element '{element}'.") 

286 elif isinstance(element, Dimension) and field_name == element.primaryKey.name: 

287 # Primary key is optional 

288 field_name = None 

289 else: 

290 if not ( 

291 field_name in element.metadata.names 

292 or (isinstance(element, Dimension) and field_name in element.alternateKeys.names) 

293 ): 

294 raise ValueError(f"Field '{field_name}' does not exist in '{element}'.") 

295 

296 return field_name