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

242 statements  

« prev     ^ index     » next       coverage.py v7.4.3, created at 2024-02-27 11:44 +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 Annotated, Any, Literal, TypeAlias 

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 "DescriptionStr", 

42 "ForeignKeyConstraint", 

43 "Index", 

44 "Schema", 

45 "SchemaVersion", 

46 "Table", 

47 "UniqueConstraint", 

48) 

49 

50CONFIG = ConfigDict( 

51 populate_by_name=True, # Populate attributes by name. 

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

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

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

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

56) 

57"""Pydantic model configuration as described in: 

58https://docs.pydantic.dev/2.0/api/config/#pydantic.config.ConfigDict 

59""" 

60 

61DESCR_MIN_LENGTH = 3 

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

63 

64DescriptionStr: TypeAlias = Annotated[str, Field(min_length=DESCR_MIN_LENGTH)] 

65"""Define a type for a description string, which must be three or more 

66characters long. Stripping of whitespace is done globally on all str fields.""" 

67 

68 

69class BaseObject(BaseModel): 

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

71 

72 model_config = CONFIG 

73 """Pydantic model configuration.""" 

74 

75 name: str 

76 """The name of the database object. 

77 

78 All Felis database objects must have a name. 

79 """ 

80 

81 id: str = Field(alias="@id") 

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

83 

84 All Felis database objects must have a unique identifier. 

85 """ 

86 

87 description: DescriptionStr | None = None 

88 """A description of the database object. 

89 

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

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

92 """ 

93 

94 @model_validator(mode="before") 

95 @classmethod 

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

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

98 if Schema.is_description_required(): 

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

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

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

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

103 return values 

104 

105 

106class DataType(Enum): 

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

108 

109 BOOLEAN = "boolean" 

110 BYTE = "byte" 

111 SHORT = "short" 

112 INT = "int" 

113 LONG = "long" 

114 FLOAT = "float" 

115 DOUBLE = "double" 

116 CHAR = "char" 

117 STRING = "string" 

118 UNICODE = "unicode" 

119 TEXT = "text" 

120 BINARY = "binary" 

121 TIMESTAMP = "timestamp" 

122 

123 

124class Column(BaseObject): 

125 """A column in a table.""" 

126 

127 datatype: DataType 

128 """The datatype of the column.""" 

129 

130 length: int | None = None 

131 """The length of the column.""" 

132 

133 nullable: bool = True 

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

135 

136 value: Any = None 

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

138 

139 autoincrement: bool | None = None 

140 """Whether the column is autoincremented.""" 

141 

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

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

144 

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

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

147 

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

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

150 

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

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

153 

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

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

156 

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

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

159 """ 

160 

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

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

163 

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

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

166 """ 

167 

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

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

170 

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

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

173 

174 @field_validator("ivoa_ucd") 

175 @classmethod 

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

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

178 if ivoa_ucd is not None: 

179 try: 

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

181 except ValueError as e: 

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

183 return ivoa_ucd 

184 

185 @model_validator(mode="before") 

186 @classmethod 

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

188 """Check that units are valid.""" 

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

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

191 

192 if fits_unit and ivoa_unit: 

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

194 unit = fits_unit or ivoa_unit 

195 

196 if unit is not None: 

197 try: 

198 units.Unit(unit) 

199 except ValueError as e: 

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

201 

202 return values 

203 

204 

205class Constraint(BaseObject): 

206 """A database table constraint.""" 

207 

208 deferrable: bool = False 

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

210 

211 initially: str | None = None 

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

213 

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

215 """Additional annotations for this constraint.""" 

216 

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

218 """The type of the constraint.""" 

219 

220 

221class CheckConstraint(Constraint): 

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

223 

224 expression: str 

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

226 

227 

228class UniqueConstraint(Constraint): 

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

230 

231 columns: list[str] 

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

233 

234 

235class Index(BaseObject): 

236 """A database table index. 

237 

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

239 """ 

240 

241 columns: list[str] | None = None 

242 """The columns in the index.""" 

243 

244 expressions: list[str] | None = None 

245 """The expressions in the index.""" 

246 

247 @model_validator(mode="before") 

248 @classmethod 

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

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

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

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

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

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

255 return values 

256 

257 

258class ForeignKeyConstraint(Constraint): 

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

260 

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

262 """ 

263 

264 columns: list[str] 

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

266 

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

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

269 

270 

271class Table(BaseObject): 

272 """A database table.""" 

273 

274 columns: Sequence[Column] 

275 """The columns in the table.""" 

276 

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

278 """The constraints on the table.""" 

279 

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

281 """The indexes on the table.""" 

282 

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

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

285 

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

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

288 

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

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

291 

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

293 known engines in the future. 

294 """ 

295 

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

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

298 

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

300 known charsets in the future. 

301 """ 

302 

303 @model_validator(mode="before") 

304 @classmethod 

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

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

307 if "constraints" in values: 

308 new_constraints: list[Constraint] = [] 

309 for item in values["constraints"]: 

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

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

312 elif item["@type"] == "Unique": 

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

314 elif item["@type"] == "Check": 

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

316 else: 

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

318 values["constraints"] = new_constraints 

319 return values 

320 

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

322 @classmethod 

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

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

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

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

327 return columns 

328 

329 

330class SchemaVersion(BaseModel): 

331 """The version of the schema.""" 

332 

333 current: str 

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

335 

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

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

338 

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

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

341 

342 

343class SchemaIdVisitor: 

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

345 

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

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

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

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

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

351 

352 This class is intended for internal use only. 

353 """ 

354 

355 def __init__(self) -> None: 

356 """Create a new SchemaVisitor.""" 

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

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

359 

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

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

362 if hasattr(obj, "id"): 

363 obj_id = getattr(obj, "id") 

364 if self.schema is not None: 

365 if obj_id in self.schema.id_map: 

366 self.duplicates.add(obj_id) 

367 else: 

368 self.schema.id_map[obj_id] = obj 

369 

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

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

372 

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

374 """ 

375 self.schema = schema 

376 self.duplicates.clear() 

377 self.add(self.schema) 

378 for table in self.schema.tables: 

379 self.visit_table(table) 

380 

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

382 """Visit a table object.""" 

383 self.add(table) 

384 for column in table.columns: 

385 self.visit_column(column) 

386 for constraint in table.constraints: 

387 self.visit_constraint(constraint) 

388 

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

390 """Visit a column object.""" 

391 self.add(column) 

392 

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

394 """Visit a constraint object.""" 

395 self.add(constraint) 

396 

397 

398class Schema(BaseObject): 

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

400 

401 class ValidationConfig: 

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

403 

404 _require_description = False 

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

406 

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

408 """ 

409 

410 version: SchemaVersion | str | None = None 

411 """The version of the schema.""" 

412 

413 tables: Sequence[Table] 

414 """The tables in the schema.""" 

415 

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

417 """Map of IDs to objects.""" 

418 

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

420 @classmethod 

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

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

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

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

425 return tables 

426 

427 @model_validator(mode="after") 

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

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

430 visitor: SchemaIdVisitor = SchemaIdVisitor() 

431 visitor.visit_schema(self) 

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

433 if len(visitor.duplicates): 

434 raise ValueError( 

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

436 ) 

437 return self 

438 

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

440 """Get an object by its ID.""" 

441 if id not in self: 

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

443 return self.id_map[id] 

444 

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

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

447 return id in self.id_map 

448 

449 @classmethod 

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

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

452 

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

454 

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

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

457 """ 

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

459 cls.ValidationConfig._require_description = rd 

460 

461 @classmethod 

462 def is_description_required(cls) -> bool: 

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

464 return cls.ValidationConfig._require_description