Coverage for python/felis/datamodel.py: 62%

240 statements  

« prev     ^ index     » next       coverage.py v7.4.1, created at 2024-02-17 10:27 +0000

1# This file is part of felis. 

2# 

3# Developed for the LSST Data Management System. 

4# This product includes software developed by the LSST Project 

5# (https://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 <https://www.gnu.org/licenses/>. 

21 

22from __future__ import annotations 

23 

24import logging 

25from collections.abc import Mapping, Sequence 

26from enum import Enum 

27from typing import Any, Literal 

28 

29from astropy import units as units # type: ignore 

30from astropy.io.votable import ucd # type: ignore 

31from pydantic import BaseModel, ConfigDict, Field, field_validator, model_validator 

32 

33logger = logging.getLogger(__name__) 

34# logger.setLevel(logging.DEBUG) 

35 

36__all__ = ( 

37 "BaseObject", 

38 "Column", 

39 "CheckConstraint", 

40 "Constraint", 

41 "ForeignKeyConstraint", 

42 "Index", 

43 "Schema", 

44 "SchemaVersion", 

45 "Table", 

46 "UniqueConstraint", 

47) 

48 

49CONFIG = ConfigDict( 

50 populate_by_name=True, # Populate attributes by name. 

51 extra="forbid", # Do not allow extra fields. 

52 use_enum_values=True, # Use enum values instead of names. 

53 validate_assignment=True, # Validate assignments after model is created. 

54 str_strip_whitespace=True, # Strip whitespace from string fields. 

55) 

56"""Pydantic model configuration as described in: 

57https://docs.pydantic.dev/2.0/api/config/#pydantic.config.ConfigDict 

58""" 

59 

60DESCR_MIN_LENGTH = 3 

61"""Minimum length for a description field.""" 

62 

63 

64class BaseObject(BaseModel): 

65 """Base class for all Felis objects.""" 

66 

67 model_config = CONFIG 

68 """Pydantic model configuration.""" 

69 

70 name: str 

71 """The name of the database object. 

72 

73 All Felis database objects must have a name. 

74 """ 

75 

76 id: str = Field(alias="@id") 

77 """The unique identifier of the database object. 

78 

79 All Felis database objects must have a unique identifier. 

80 """ 

81 

82 description: str | None = Field(None, min_length=DESCR_MIN_LENGTH) 

83 """A description of the database object. 

84 

85 By default, the description is optional but will be required if 

86 `BaseObject.Config.require_description` is set to `True` by the user. 

87 """ 

88 

89 @model_validator(mode="before") 

90 @classmethod 

91 def check_description(cls, values: dict[str, Any]) -> dict[str, Any]: 

92 """Check that the description is present if required.""" 

93 if Schema.is_description_required(): 

94 if "description" not in values or not values["description"]: 

95 raise ValueError("Description is required and must be non-empty") 

96 if len(values["description"].strip()) < DESCR_MIN_LENGTH: 

97 raise ValueError(f"Description must be at least {DESCR_MIN_LENGTH} characters long") 

98 return values 

99 

100 

101class DataType(Enum): 

102 """`Enum` representing the data types supported by Felis.""" 

103 

104 BOOLEAN = "boolean" 

105 BYTE = "byte" 

106 SHORT = "short" 

107 INT = "int" 

108 LONG = "long" 

109 FLOAT = "float" 

110 DOUBLE = "double" 

111 CHAR = "char" 

112 STRING = "string" 

113 UNICODE = "unicode" 

114 TEXT = "text" 

115 BINARY = "binary" 

116 TIMESTAMP = "timestamp" 

117 

118 

119class Column(BaseObject): 

120 """A column in a table.""" 

121 

122 datatype: DataType 

123 """The datatype of the column.""" 

124 

125 length: int | None = None 

126 """The length of the column.""" 

127 

128 nullable: bool = True 

129 """Whether the column can be `NULL`.""" 

130 

131 value: Any = None 

132 """The default value of the column.""" 

133 

134 autoincrement: bool | None = None 

135 """Whether the column is autoincremented.""" 

136 

137 mysql_datatype: str | None = Field(None, alias="mysql:datatype") 

138 """The MySQL datatype of the column.""" 

139 

140 ivoa_ucd: str | None = Field(None, alias="ivoa:ucd") 

141 """The IVOA UCD of the column.""" 

142 

143 fits_tunit: str | None = Field(None, alias="fits:tunit") 

144 """The FITS TUNIT of the column.""" 

145 

146 ivoa_unit: str | None = Field(None, alias="ivoa:unit") 

147 """The IVOA unit of the column.""" 

148 

149 tap_column_index: int | None = Field(None, alias="tap:column_index") 

150 """The TAP_SCHEMA column index of the column.""" 

151 

152 tap_principal: int | None = Field(0, alias="tap:principal", ge=0, le=1) 

153 """Whether this is a TAP_SCHEMA principal column; can be either 0 or 1. 

154 """ 

155 

156 votable_arraysize: int | Literal["*"] | None = Field(None, alias="votable:arraysize") 

157 """The VOTable arraysize of the column.""" 

158 

159 tap_std: int | None = Field(0, alias="tap:std", ge=0, le=1) 

160 """TAP_SCHEMA indication that this column is defined by an IVOA standard. 

161 """ 

162 

163 votable_utype: str | None = Field(None, alias="votable:utype") 

164 """The VOTable utype (usage-specific or unique type) of the column.""" 

165 

166 votable_xtype: str | None = Field(None, alias="votable:xtype") 

167 """The VOTable xtype (extended type) of the column.""" 

168 

169 @field_validator("ivoa_ucd") 

170 @classmethod 

171 def check_ivoa_ucd(cls, ivoa_ucd: str) -> str: 

172 """Check that IVOA UCD values are valid.""" 

173 if ivoa_ucd is not None: 

174 try: 

175 ucd.parse_ucd(ivoa_ucd, check_controlled_vocabulary=True, has_colon=";" in ivoa_ucd) 

176 except ValueError as e: 

177 raise ValueError(f"Invalid IVOA UCD: {e}") 

178 return ivoa_ucd 

179 

180 @model_validator(mode="before") 

181 @classmethod 

182 def check_units(cls, values: dict[str, Any]) -> dict[str, Any]: 

183 """Check that units are valid.""" 

184 fits_unit = values.get("fits:tunit") 

185 ivoa_unit = values.get("ivoa:unit") 

186 

187 if fits_unit and ivoa_unit: 

188 raise ValueError("Column cannot have both FITS and IVOA units") 

189 unit = fits_unit or ivoa_unit 

190 

191 if unit is not None: 

192 try: 

193 units.Unit(unit) 

194 except ValueError as e: 

195 raise ValueError(f"Invalid unit: {e}") 

196 

197 return values 

198 

199 

200class Constraint(BaseObject): 

201 """A database table constraint.""" 

202 

203 deferrable: bool = False 

204 """If `True` then this constraint will be declared as deferrable.""" 

205 

206 initially: str | None = None 

207 """Value for ``INITIALLY`` clause, only used if ``deferrable`` is True.""" 

208 

209 annotations: Mapping[str, Any] = Field(default_factory=dict) 

210 """Additional annotations for this constraint.""" 

211 

212 type: str | None = Field(None, alias="@type") 

213 """The type of the constraint.""" 

214 

215 

216class CheckConstraint(Constraint): 

217 """A check constraint on a table.""" 

218 

219 expression: str 

220 """The expression for the check constraint.""" 

221 

222 

223class UniqueConstraint(Constraint): 

224 """A unique constraint on a table.""" 

225 

226 columns: list[str] 

227 """The columns in the unique constraint.""" 

228 

229 

230class Index(BaseObject): 

231 """A database table index. 

232 

233 An index can be defined on either columns or expressions, but not both. 

234 """ 

235 

236 columns: list[str] | None = None 

237 """The columns in the index.""" 

238 

239 expressions: list[str] | None = None 

240 """The expressions in the index.""" 

241 

242 @model_validator(mode="before") 

243 @classmethod 

244 def check_columns_or_expressions(cls, values: dict[str, Any]) -> dict[str, Any]: 

245 """Check that columns or expressions are specified, but not both.""" 

246 if "columns" in values and "expressions" in values: 

247 raise ValueError("Defining columns and expressions is not valid") 

248 elif "columns" not in values and "expressions" not in values: 

249 raise ValueError("Must define columns or expressions") 

250 return values 

251 

252 

253class ForeignKeyConstraint(Constraint): 

254 """A foreign key constraint on a table. 

255 

256 These will be reflected in the TAP_SCHEMA keys and key_columns data. 

257 """ 

258 

259 columns: list[str] 

260 """The columns comprising the foreign key.""" 

261 

262 referenced_columns: list[str] = Field(alias="referencedColumns") 

263 """The columns referenced by the foreign key.""" 

264 

265 

266class Table(BaseObject): 

267 """A database table.""" 

268 

269 columns: Sequence[Column] 

270 """The columns in the table.""" 

271 

272 constraints: list[Constraint] = Field(default_factory=list) 

273 """The constraints on the table.""" 

274 

275 indexes: list[Index] = Field(default_factory=list) 

276 """The indexes on the table.""" 

277 

278 primaryKey: str | list[str] | None = None 

279 """The primary key of the table.""" 

280 

281 tap_table_index: int | None = Field(None, alias="tap:table_index") 

282 """The IVOA TAP_SCHEMA table index of the table.""" 

283 

284 mysql_engine: str | None = Field(None, alias="mysql:engine") 

285 """The mysql engine to use for the table. 

286 

287 For now this is a freeform string but it could be constrained to a list of 

288 known engines in the future. 

289 """ 

290 

291 mysql_charset: str | None = Field(None, alias="mysql:charset") 

292 """The mysql charset to use for the table. 

293 

294 For now this is a freeform string but it could be constrained to a list of 

295 known charsets in the future. 

296 """ 

297 

298 @model_validator(mode="before") 

299 @classmethod 

300 def create_constraints(cls, values: dict[str, Any]) -> dict[str, Any]: 

301 """Create constraints from the ``constraints`` field.""" 

302 if "constraints" in values: 

303 new_constraints: list[Constraint] = [] 

304 for item in values["constraints"]: 

305 if item["@type"] == "ForeignKey": 

306 new_constraints.append(ForeignKeyConstraint(**item)) 

307 elif item["@type"] == "Unique": 

308 new_constraints.append(UniqueConstraint(**item)) 

309 elif item["@type"] == "Check": 

310 new_constraints.append(CheckConstraint(**item)) 

311 else: 

312 raise ValueError(f"Unknown constraint type: {item['@type']}") 

313 values["constraints"] = new_constraints 

314 return values 

315 

316 @field_validator("columns", mode="after") 

317 @classmethod 

318 def check_unique_column_names(cls, columns: list[Column]) -> list[Column]: 

319 """Check that column names are unique.""" 

320 if len(columns) != len(set(column.name for column in columns)): 

321 raise ValueError("Column names must be unique") 

322 return columns 

323 

324 

325class SchemaVersion(BaseModel): 

326 """The version of the schema.""" 

327 

328 current: str 

329 """The current version of the schema.""" 

330 

331 compatible: list[str] = Field(default_factory=list) 

332 """The compatible versions of the schema.""" 

333 

334 read_compatible: list[str] = Field(default_factory=list) 

335 """The read compatible versions of the schema.""" 

336 

337 

338class SchemaIdVisitor: 

339 """Visitor to build a Schema object's map of IDs to objects. 

340 

341 Duplicates are added to a set when they are encountered, which can be 

342 accessed via the `duplicates` attribute. The presence of duplicates will 

343 not throw an error. Only the first object with a given ID will be added to 

344 the map, but this should not matter, since a ValidationError will be thrown 

345 by the `model_validator` method if any duplicates are found in the schema. 

346 

347 This class is intended for internal use only. 

348 """ 

349 

350 def __init__(self) -> None: 

351 """Create a new SchemaVisitor.""" 

352 self.schema: "Schema" | None = None 

353 self.duplicates: set[str] = set() 

354 

355 def add(self, obj: BaseObject) -> None: 

356 """Add an object to the ID map.""" 

357 if hasattr(obj, "id"): 

358 obj_id = getattr(obj, "id") 

359 if self.schema is not None: 

360 if obj_id in self.schema.id_map: 

361 self.duplicates.add(obj_id) 

362 else: 

363 self.schema.id_map[obj_id] = obj 

364 

365 def visit_schema(self, schema: "Schema") -> None: 

366 """Visit the schema object that was added during initialization. 

367 

368 This will set an internal variable pointing to the schema object. 

369 """ 

370 self.schema = schema 

371 self.duplicates.clear() 

372 self.add(self.schema) 

373 for table in self.schema.tables: 

374 self.visit_table(table) 

375 

376 def visit_table(self, table: Table) -> None: 

377 """Visit a table object.""" 

378 self.add(table) 

379 for column in table.columns: 

380 self.visit_column(column) 

381 for constraint in table.constraints: 

382 self.visit_constraint(constraint) 

383 

384 def visit_column(self, column: Column) -> None: 

385 """Visit a column object.""" 

386 self.add(column) 

387 

388 def visit_constraint(self, constraint: Constraint) -> None: 

389 """Visit a constraint object.""" 

390 self.add(constraint) 

391 

392 

393class Schema(BaseObject): 

394 """The database schema containing the tables.""" 

395 

396 class ValidationConfig: 

397 """Validation configuration which is specific to Felis.""" 

398 

399 _require_description = False 

400 """Flag to require a description for all objects. 

401 

402 This is set by the `require_description` class method. 

403 """ 

404 

405 version: SchemaVersion | str | None = None 

406 """The version of the schema.""" 

407 

408 tables: Sequence[Table] 

409 """The tables in the schema.""" 

410 

411 id_map: dict[str, Any] = Field(default_factory=dict, exclude=True) 

412 """Map of IDs to objects.""" 

413 

414 @field_validator("tables", mode="after") 

415 @classmethod 

416 def check_unique_table_names(cls, tables: list[Table]) -> list[Table]: 

417 """Check that table names are unique.""" 

418 if len(tables) != len(set(table.name for table in tables)): 

419 raise ValueError("Table names must be unique") 

420 return tables 

421 

422 @model_validator(mode="after") 

423 def create_id_map(self: Schema) -> Schema: 

424 """Create a map of IDs to objects.""" 

425 visitor: SchemaIdVisitor = SchemaIdVisitor() 

426 visitor.visit_schema(self) 

427 logger.debug(f"ID map contains {len(self.id_map.keys())} objects") 

428 if len(visitor.duplicates): 

429 raise ValueError( 

430 "Duplicate IDs found in schema:\n " + "\n ".join(visitor.duplicates) + "\n" 

431 ) 

432 return self 

433 

434 def __getitem__(self, id: str) -> BaseObject: 

435 """Get an object by its ID.""" 

436 if id not in self: 

437 raise KeyError(f"Object with ID '{id}' not found in schema") 

438 return self.id_map[id] 

439 

440 def __contains__(self, id: str) -> bool: 

441 """Check if an object with the given ID is in the schema.""" 

442 return id in self.id_map 

443 

444 @classmethod 

445 def require_description(cls, rd: bool = True) -> None: 

446 """Set whether a description is required for all objects. 

447 

448 This includes the schema, tables, columns, and constraints. 

449 

450 Users should call this method to set the requirement for a description 

451 when validating schemas, rather than change the flag value directly. 

452 """ 

453 logger.debug(f"Setting description requirement to '{rd}'") 

454 cls.ValidationConfig._require_description = rd 

455 

456 @classmethod 

457 def is_description_required(cls) -> bool: 

458 """Return whether a description is required for all objects.""" 

459 return cls.ValidationConfig._require_description