Coverage for python/lsst/daf/butler/script/ingest_files.py: 22%

53 statements  

« prev     ^ index     » next       coverage.py v7.2.7, created at 2023-06-15 09:13 +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__ = ("ingest_files",) 

24 

25import logging 

26from collections import defaultdict 

27from typing import TYPE_CHECKING, Any 

28 

29from astropy.table import Table 

30from lsst.resources import ResourcePath 

31from lsst.utils import doImport 

32 

33from .._butler import Butler 

34from ..core import DatasetIdGenEnum, DatasetRef, FileDataset 

35 

36if TYPE_CHECKING: 

37 from ..core import DatasetType, DimensionUniverse 

38 

39log = logging.getLogger(__name__) 

40 

41 

42def ingest_files( 

43 repo: str, 

44 dataset_type: str, 

45 run: str, 

46 table_file: str, 

47 data_id: tuple[str, ...] = (), 

48 formatter: str | None = None, 

49 id_generation_mode: str = "UNIQUE", 

50 prefix: str | None = None, 

51 transfer: str = "auto", 

52) -> None: 

53 """Ingest files from a table. 

54 

55 Parameters 

56 ---------- 

57 repo : `str` 

58 URI string of the Butler repo to use. 

59 dataset_type : `str` 

60 The name of the dataset type for the files to be ingested. This 

61 dataset type must exist. 

62 run : `str` 

63 The run in which the files should be ingested. 

64 table_file : `str` 

65 Path to a table file to read. This file can be in any format that 

66 can be read by Astropy so long as Astropy can determine the format 

67 itself. 

68 data_id : `tuple` of `str` 

69 Tuple of strings of the form ``keyword=value`` that can be used 

70 to define dataId elements that are fixed for all ingested files 

71 found in the table file. This allows those columns to be missing 

72 from the table file. Dimensions given here override table columns. 

73 formatter : `str`, optional 

74 Fully-qualified python class name for the `Formatter` to use 

75 to read the ingested files. If `None` the formatter is read from 

76 datastore configuration based on the dataset type. 

77 id_generation_mode : `str`, optional 

78 Mode to use for generating IDs. Should map to `DatasetGenIdEnum`. 

79 prefix : `str`, optional 

80 Prefix to use when resolving relative paths in table files. The default 

81 is to use the current working directory. 

82 transfer : `str`, optional 

83 Transfer mode to use for ingest. 

84 """ 

85 

86 # Check that the formatter can be imported -- validate this as soon 

87 # as possible before we read a potentially large table file. 

88 if formatter: 

89 doImport(formatter) 

90 else: 

91 formatter = None 

92 

93 # Force empty string prefix (from click) to None for API compatibility. 

94 if not prefix: 

95 prefix = None 

96 

97 # Convert the dataset ID gen mode string to enum. 

98 id_gen_mode = DatasetIdGenEnum.__members__[id_generation_mode] 

99 

100 # Create the butler with the relevant run attached. 

101 butler = Butler(repo, run=run) 

102 

103 datasetType = butler.registry.getDatasetType(dataset_type) 

104 

105 # Convert the k=v strings into a dataId dict. 

106 universe = butler.dimensions 

107 common_data_id = parse_data_id_tuple(data_id, universe) 

108 

109 # Read the table assuming that Astropy can work out the format. 

110 uri = ResourcePath(table_file, forceAbsolute=False) 

111 with uri.as_local() as local_file: 

112 table = Table.read(local_file.ospath) 

113 

114 datasets = extract_datasets_from_table( 

115 table, common_data_id, datasetType, run, formatter, prefix, id_gen_mode 

116 ) 

117 

118 butler.ingest(*datasets, transfer=transfer) 

119 

120 

121def extract_datasets_from_table( 

122 table: Table, 

123 common_data_id: dict, 

124 datasetType: DatasetType, 

125 run: str, 

126 formatter: str | None = None, 

127 prefix: str | None = None, 

128 id_generation_mode: DatasetIdGenEnum = DatasetIdGenEnum.UNIQUE, 

129) -> list[FileDataset]: 

130 """Extract datasets from the supplied table. 

131 

132 Parameters 

133 ---------- 

134 table : `astropy.table.Table` 

135 Table containing the datasets. The first column is assumed to be 

136 the file URI and the remaining columns are dimensions. 

137 common_data_id : `dict` 

138 Data ID values that are common to every row in the table. These 

139 take priority if a dimension in this dataId is also present as 

140 a column in the table. 

141 datasetType : `DatasetType` 

142 The dataset type to be associated with the ingested data. 

143 run : `str` 

144 The name of the run that will be receiving these datasets. 

145 formatter : `str`, optional 

146 Fully-qualified python class name for the `Formatter` to use 

147 to read the ingested files. If `None` the formatter is read from 

148 datastore configuration based on the dataset type. 

149 prefix : `str`, optional 

150 Prefix to be used for relative paths. Can be `None` for current 

151 working directory. 

152 id_generation_mode: `DatasetIdGenEnum`, optional 

153 The mode to use when creating the dataset IDs. 

154 

155 Returns 

156 ------- 

157 datasets : `list` of `FileDataset` 

158 The `FileDataset` objects corresponding to the rows in the table. 

159 The number of elements in this list can be smaller than the number 

160 of rows in the file because one file can appear in multiple rows 

161 with different dataIds. 

162 """ 

163 # The file is the first column and everything else is assumed to 

164 # be dimensions so we need to know the name of that column. 

165 file_column = table.colnames[0] 

166 

167 # Handle multiple dataIds per file by grouping by file. 

168 refs_by_file = defaultdict(list) 

169 n_dataset_refs = 0 

170 for row in table: 

171 # Convert the row to a dataId, remembering to extract the 

172 # path column. 

173 dataId = dict(row) 

174 path = dataId.pop(file_column) 

175 

176 # The command line can override a column. 

177 dataId.update(common_data_id) 

178 

179 # Create the dataset ref that is to be ingested. 

180 ref = DatasetRef(datasetType, dataId, run=run, id_generation_mode=id_generation_mode) # type: ignore 

181 

182 # Convert path to absolute (because otherwise system will 

183 # assume relative to datastore root and that is almost certainly 

184 # never the right default here). 

185 path_uri = ResourcePath(path, root=prefix, forceAbsolute=True) 

186 

187 refs_by_file[path_uri].append(ref) 

188 n_dataset_refs += 1 

189 

190 datasets = [ 

191 FileDataset( 

192 path=file_uri, 

193 refs=refs, 

194 formatter=formatter, 

195 ) 

196 for file_uri, refs in refs_by_file.items() 

197 ] 

198 

199 log.info("Ingesting %d dataset ref(s) from %d file(s)", n_dataset_refs, len(datasets)) 

200 

201 return datasets 

202 

203 

204def parse_data_id_tuple(data_ids: tuple[str, ...], universe: DimensionUniverse) -> dict[str, Any]: 

205 # Convert any additional k=v strings in the dataId tuple to dict 

206 # form. 

207 data_id: dict[str, Any] = {} 

208 for id_str in data_ids: 

209 dimension_str, value = id_str.split("=") 

210 

211 try: 

212 dimension = universe.getStaticDimensions()[dimension_str] 

213 except KeyError: 

214 raise ValueError(f"DataID dimension '{dimension_str}' is not known to this universe.") from None 

215 

216 # Cast the value to the right python type (since they will be 

217 # strings at this point). 

218 value = dimension.primaryKey.getPythonType()(value) 

219 

220 data_id[dimension_str] = value 

221 return data_id