Coverage for src / CSET / operators / read.py: 91%

1# © Crown copyright, Met Office (2022-2025) and CSET contributors. 

2# 

3# Licensed under the Apache License, Version 2.0 (the "License"); 

4# you may not use this file except in compliance with the License. 

5# You may obtain a copy of the License at 

6# 

7# http://www.apache.org/licenses/LICENSE-2.0 

8# 

9# Unless required by applicable law or agreed to in writing, software 

10# distributed under the License is distributed on an "AS IS" BASIS, 

11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 

12# See the License for the specific language governing permissions and 

13# limitations under the License. 

14 

15"""Operators for reading various types of files from disk.""" 

16 

17import ast 

18import datetime 

19import functools 

20import glob 

21import itertools 

22import logging 

23from pathlib import Path 

24from typing import Literal 

25 

26import iris 

27import iris.coord_systems 

28import iris.coords 

29import iris.cube 

30import iris.exceptions 

31import iris.util 

32import numpy as np 

33from iris.analysis.cartography import rotate_pole, rotate_winds 

34 

35from CSET._common import iter_maybe 

36from CSET.operators._stash_to_lfric import STASH_TO_LFRIC 

37from CSET.operators._utils import ( 

38 get_cube_coordindex, 

39 get_cube_yxcoordname, 

40 is_spatialdim, 

41) 

42 

43 

44class NoDataError(FileNotFoundError): 

45 """Error that no data has been loaded.""" 

46 

47 

48def read_cube( 

49 file_paths: list[str] | str, 

50 constraint: iris.Constraint = None, 

51 model_names: list[str] | str | None = None, 

52 subarea_type: str = None, 

53 subarea_extent: list[float] = None, 

54 **kwargs, 

55) -> iris.cube.Cube: 

56 """Read a single cube from files. 

57 

58 Read operator that takes a path string (can include shell-style glob 

59 patterns), and loads the cube matching the constraint. If any paths point to 

60 directory, all the files contained within are loaded. 

61 

62 Ensemble data can also be loaded. If it has a realization coordinate 

63 already, it will be directly used. If not, it will have its member number 

64 guessed from the filename, based on one of several common patterns. For 

65 example the pattern *emXX*, where XX is the realization. 

66 

67 Deterministic data will be loaded with a realization of 0, allowing it to be 

68 processed in the same way as ensemble data. 

69 

70 Arguments 

71 --------- 

72 file_paths: str | list[str] 

73 Path or paths to where .pp/.nc files are located 

74 constraint: iris.Constraint | iris.ConstraintCombination, optional 

75 Constraints to filter data by. Defaults to unconstrained. 

76 model_names: str | list[str], optional 

77 Names of the models that correspond to respective paths in file_paths. 

78 subarea_type: "gridcells" | "modelrelative" | "realworld", optional 

79 Whether to constrain data by model relative coordinates or real world 

80 coordinates. 

81 subarea_extent: list, optional 

82 List of coordinates to constraint data by, in order lower latitude, 

83 upper latitude, lower longitude, upper longitude. 

84 

85 Returns 

86 ------- 

87 cubes: iris.cube.Cube 

88 Cube loaded 

89 

90 Raises 

91 ------ 

92 FileNotFoundError 

93 If the provided path does not exist 

94 ValueError 

95 If the constraint doesn't produce a single cube. 

96 """ 

97 cubes = read_cubes( 

98 file_paths=file_paths, 

99 constraint=constraint, 

100 model_names=model_names, 

101 subarea_type=subarea_type, 

102 subarea_extent=subarea_extent, 

103 ) 

104 # Check filtered cubes is a CubeList containing one cube. 

105 if len(cubes) == 1: 

106 return cubes[0] 

107 else: 

108 raise ValueError( 

109 f"Constraint doesn't produce single cube: {constraint}\n{cubes}" 

110 ) 

111 

112 

113def read_cubes( 

114 file_paths: list[str] | str, 

115 constraint: iris.Constraint | None = None, 

116 model_names: str | list[str] | None = None, 

117 subarea_type: str = None, 

118 subarea_extent: list = None, 

119 **kwargs, 

120) -> iris.cube.CubeList: 

121 """Read cubes from files. 

122 

123 Read operator that takes a path string (can include shell-style glob 

124 patterns), and loads the cubes matching the constraint. If any paths point 

125 to directory, all the files contained within are loaded. 

126 

127 Ensemble data can also be loaded. If it has a realization coordinate 

128 already, it will be directly used. If not, it will have its member number 

129 guessed from the filename, based on one of several common patterns. For 

130 example the pattern *emXX*, where XX is the realization. 

131 

132 Deterministic data will be loaded with a realization of 0, allowing it to be 

133 processed in the same way as ensemble data. 

134 

135 Data output by XIOS (such as LFRic) has its per-file metadata removed so 

136 that the cubes merge across files. 

137 

138 Arguments 

139 --------- 

140 file_paths: str | list[str] 

141 Path or paths to where .pp/.nc files are located. Can include globs. 

142 constraint: iris.Constraint | iris.ConstraintCombination, optional 

143 Constraints to filter data by. Defaults to unconstrained. 

144 model_names: str | list[str], optional 

145 Names of the models that correspond to respective paths in file_paths. 

146 subarea_type: str, optional 

147 Whether to constrain data by model relative coordinates or real world 

148 coordinates. 

149 subarea_extent: list[float], optional 

150 List of coordinates to constraint data by, in order lower latitude, 

151 upper latitude, lower longitude, upper longitude. 

152 

153 Returns 

154 ------- 

155 cubes: iris.cube.CubeList 

156 Cubes loaded after being merged and concatenated. 

157 

158 Raises 

159 ------ 

160 FileNotFoundError 

161 If the provided path does not exist 

162 """ 

163 # Get iterable of paths. Each path corresponds to 1 model. 

164 paths = iter_maybe(file_paths) 

165 model_names = iter_maybe(model_names) 

166 

167 # Check we have appropriate number of model names. 

168 if model_names != (None,) and len(model_names) != len(paths): 

169 raise ValueError( 

170 f"The number of model names ({len(model_names)}) should equal " 

171 f"the number of paths given ({len(paths)})." 

172 ) 

173 

174 # Load the data for each model into a CubeList per model. 

175 model_cubes = ( 

176 _load_model(path, name, constraint) 

177 for path, name in itertools.zip_longest(paths, model_names, fillvalue=None) 

178 ) 

179 

180 # Split out first model's cubes and mark it as the base for comparisons. 

181 cubes = next(model_cubes) 

182 for cube in cubes: 

183 # Use 1 to indicate True, as booleans can't be saved in NetCDF attributes. 

184 cube.attributes["cset_comparison_base"] = 1 

185 

186 # Load the rest of the models. 

187 cubes.extend(itertools.chain.from_iterable(model_cubes)) 

188 

189 # Unify time units so different case studies can merge. 

190 iris.util.unify_time_units(cubes) 

191 

192 # Select sub region. 

193 cubes = _cutout_cubes(cubes, subarea_type, subarea_extent) 

194 

195 # Merge and concatenate cubes now metadata has been fixed. 

196 cubes = cubes.merge() 

197 cubes = cubes.concatenate() 

198 

199 # Squeeze single valued coordinates into scalar coordinates. 

200 cubes = iris.cube.CubeList(iris.util.squeeze(cube) for cube in cubes) 

201 

202 # Ensure dimension coordinates are bounded. 

203 for cube in cubes: 

204 for dim_coord in cube.coords(dim_coords=True): 

205 # Iris can't guess the bounds of a scalar coordinate. 

206 if not dim_coord.has_bounds() and dim_coord.shape[0] > 1: 

207 dim_coord.guess_bounds() 

208 

209 logging.info("Loaded cubes: %s", cubes) 

210 if len(cubes) == 0: 

211 raise NoDataError("No cubes loaded, check your constraints!") 

212 return cubes 

213 

214 

215def _load_model( 

216 paths: str | list[str], 

217 model_name: str | None, 

218 constraint: iris.Constraint | None, 

219) -> iris.cube.CubeList: 

220 """Load a single model's data into a CubeList.""" 

221 input_files = _check_input_files(paths) 

222 # If unset, a constraint of None lets everything be loaded. 

223 logging.debug("Constraint: %s", constraint) 

224 cubes = iris.load(input_files, constraint, callback=_loading_callback) 

225 # Make the UM's winds consistent with LFRic. 

226 _fix_um_winds(cubes) 

227 

228 # Add model_name attribute to each cube to make it available at any further 

229 # step without needing to pass it as function parameter. 

230 if model_name is not None: 

231 for cube in cubes: 

232 cube.attributes["model_name"] = model_name 

233 return cubes 

234 

235 

236def _check_input_files(input_paths: str | list[str]) -> list[Path]: 

237 """Get an iterable of files to load, and check that they all exist. 

238 

239 Arguments 

240 --------- 

241 input_paths: list[str] 

242 List of paths to input files or directories. The path may itself contain 

243 glob patterns, but unlike in shells it will match directly first. 

244 

245 Returns 

246 ------- 

247 list[Path] 

248 A list of files to load. 

249 

250 Raises 

251 ------ 

252 FileNotFoundError: 

253 If the provided arguments don't resolve to at least one existing file. 

254 """ 

255 files = [] 

256 for raw_filename in iter_maybe(input_paths): 

257 # Match glob-like files first, if they exist. 

258 raw_path = Path(raw_filename) 

259 if raw_path.is_file(): 

260 files.append(raw_path) 

261 else: 

262 for input_path in glob.glob(raw_filename): 

263 # Convert string paths into Path objects. 

264 input_path = Path(input_path) 

265 # Get the list of files in the directory, or use it directly. 

266 if input_path.is_dir(): 

267 logging.debug("Checking directory '%s' for files", input_path) 

268 files.extend(p for p in input_path.iterdir() if p.is_file()) 

269 else: 

270 files.append(input_path) 

271 

272 files.sort() 

273 logging.info("Loading files:\n%s", "\n".join(str(path) for path in files)) 

274 if len(files) == 0: 

275 raise FileNotFoundError(f"No files found for {input_paths}") 

276 return files 

277 

278 

279def _cutout_cubes( 

280 cubes: iris.cube.CubeList, 

281 subarea_type: Literal["gridcells", "realworld", "modelrelative"] | None, 

282 subarea_extent: list[float, float, float, float], 

283): 

284 """Cut out a subarea from a CubeList.""" 

285 if subarea_type is None: 

286 logging.debug("Subarea selection is disabled.") 

287 return cubes 

288 

289 # If selected, cutout according to number of grid cells to trim from each edge. 

290 cutout_cubes = iris.cube.CubeList() 

291 # Find spatial coordinates 

292 for cube in cubes: 

293 # Find dimension coordinates. 

294 lat_name, lon_name = get_cube_yxcoordname(cube) 

295 

296 # Compute cutout based on number of cells to trim from edges. 

297 if subarea_type == "gridcells": 

298 logging.debug( 

299 "User requested LowerTrim: %s LeftTrim: %s UpperTrim: %s RightTrim: %s", 

300 subarea_extent[0], 

301 subarea_extent[1], 

302 subarea_extent[2], 

303 subarea_extent[3], 

304 ) 

305 lat_points = np.sort(cube.coord(lat_name).points) 

306 lon_points = np.sort(cube.coord(lon_name).points) 

307 # Define cutout region using user provided cell points. 

308 lats = [lat_points[subarea_extent[0]], lat_points[-subarea_extent[2] - 1]] 

309 lons = [lon_points[subarea_extent[1]], lon_points[-subarea_extent[3] - 1]] 

310 

311 # Compute cutout based on specified coordinate values. 

312 elif subarea_type == "realworld" or subarea_type == "modelrelative": 

313 # If not gridcells, cutout by requested geographic area, 

314 logging.debug( 

315 "User requested LLat: %s ULat: %s LLon: %s ULon: %s", 

316 subarea_extent[0], 

317 subarea_extent[1], 

318 subarea_extent[2], 

319 subarea_extent[3], 

320 ) 

321 # Define cutout region using user provided coordinates. 

322 lats = np.array(subarea_extent[0:2]) 

323 lons = np.array(subarea_extent[2:4]) 

324 # Ensure cutout longitudes are within +/- 180.0 bounds. 

325 while lons[0] < -180.0: 

326 lons += 360.0 

327 while lons[1] > 180.0: 

328 lons -= 360.0 

329 # If the coordinate system is rotated we convert coordinates into 

330 # model-relative coordinates to extract the appropriate cutout. 

331 coord_system = cube.coord(lat_name).coord_system 

332 if subarea_type == "realworld" and isinstance( 

333 coord_system, iris.coord_systems.RotatedGeogCS 

334 ): 

335 lons, lats = rotate_pole( 

336 lons, 

337 lats, 

338 pole_lon=coord_system.grid_north_pole_longitude, 

339 pole_lat=coord_system.grid_north_pole_latitude, 

340 ) 

341 else: 

342 raise ValueError("Unknown subarea_type:", subarea_type) 

343 

344 # Do cutout and add to cutout_cubes. 

345 intersection_args = {lat_name: lats, lon_name: lons} 

346 logging.debug("Cutting out coords: %s", intersection_args) 

347 try: 

348 cutout_cubes.append(cube.intersection(**intersection_args)) 

349 except IndexError as err: 

350 raise ValueError( 

351 "Region cutout error. Check and update SUBAREA_EXTENT." 

352 "Cutout region requested should be contained within data area. " 

353 "Also check if cutout region requested is smaller than input grid spacing." 

354 ) from err 

355 

356 return cutout_cubes 

357 

358 

359def _loading_callback(cube: iris.cube.Cube, field, filename: str) -> iris.cube.Cube: 

360 """Compose together the needed callbacks into a single function.""" 

361 # Most callbacks operate in-place, but save the cube when returned! 

362 _realization_callback(cube, field, filename) 

363 _um_normalise_callback(cube, field, filename) 

364 _lfric_normalise_callback(cube, field, filename) 

365 cube = _lfric_time_coord_fix_callback(cube, field, filename) 

366 _normalise_var0_varname(cube) 

367 cube = _fix_no_spatial_coords_callback(cube) 

368 _fix_spatial_coords_callback(cube) 

369 _fix_pressure_coord_callback(cube) 

370 _fix_um_radtime(cube) 

371 _fix_cell_methods(cube) 

372 cube = _convert_cube_units_callback(cube) 

373 cube = _grid_longitude_fix_callback(cube) 

374 _fix_lfric_cloud_base_altitude(cube) 

375 _proleptic_gregorian_fix(cube) 

376 _lfric_time_callback(cube) 

377 _lfric_forecast_period_callback(cube) 

378 _normalise_ML_varname(cube) 

379 return cube 

380 

381 

382def _realization_callback(cube, field, filename): 

383 """Give deterministic cubes a realization of 0. 

384 

385 This means they can be handled in the same way as ensembles through the rest 

386 of the code. 

387 """ 

388 # Only add if realization coordinate does not exist. 

389 if not cube.coords("realization"): 

390 cube.add_aux_coord( 

391 iris.coords.DimCoord(0, standard_name="realization", units="1") 

392 ) 

393 

394 

395@functools.lru_cache(None) 

396def _warn_once(msg): 

397 """Print a warning message, skipping recent duplicates.""" 

398 logging.warning(msg) 

399 

400 

401def _um_normalise_callback(cube: iris.cube.Cube, field, filename): 

402 """Normalise UM STASH variable long names to LFRic variable names. 

403 

404 Note standard names will remain associated with cubes where different. 

405 Long name will be used consistently in output filename and titles. 

406 """ 

407 # Convert STASH to LFRic variable name 

408 if "STASH" in cube.attributes: 

409 stash = cube.attributes["STASH"] 

410 try: 

411 (name, grid) = STASH_TO_LFRIC[str(stash)] 

412 cube.long_name = name 

413 except KeyError: 

414 # Don't change cubes with unknown stash codes. 

415 _warn_once( 

416 f"Unknown STASH code: {stash}. Please check file stash_to_lfric.py to update." 

417 ) 

418 

419 

420def _lfric_normalise_callback(cube: iris.cube.Cube, field, filename): 

421 """Normalise attributes that prevents LFRic cube from merging. 

422 

423 The uuid and timeStamp relate to the output file, as saved by XIOS, and has 

424 no relation to the data contained. These attributes are removed. 

425 

426 The um_stash_source is a list of STASH codes for when an LFRic field maps to 

427 multiple UM fields, however it can be encoded in any order. This attribute 

428 is sorted to prevent this. This attribute is only present in LFRic data that 

429 has been converted to look like UM data. 

430 """ 

431 # Remove unwanted attributes. 

432 cube.attributes.pop("timeStamp", None) 

433 cube.attributes.pop("uuid", None) 

434 cube.attributes.pop("name", None) 

435 cube.attributes.pop("source", None) 

436 cube.attributes.pop("analysis_source", None) 

437 cube.attributes.pop("history", None) 

438 

439 # Sort STASH code list. 

440 stash_list = cube.attributes.get("um_stash_source") 

441 if stash_list: 

442 # Parse the string as a list, sort, then re-encode as a string. 

443 cube.attributes["um_stash_source"] = str(sorted(ast.literal_eval(stash_list))) 

444 

445 

446def _lfric_time_coord_fix_callback( 

447 cube: iris.cube.Cube, field, filename 

448) -> iris.cube.Cube: 

449 """Ensure the time coordinate is a DimCoord rather than an AuxCoord. 

450 

451 The coordinate is converted and replaced if not. SLAMed LFRic data has this 

452 issue, though the coordinate satisfies all the properties for a DimCoord. 

453 Scalar time values are left as AuxCoords. 

454 """ 

455 # This issue seems to come from iris's handling of NetCDF files where time 

456 # always ends up as an AuxCoord. 

457 if cube.coords("time"): 

458 time_coord = cube.coord("time") 

459 if ( 

460 not isinstance(time_coord, iris.coords.DimCoord) 

461 and len(cube.coord_dims(time_coord)) == 1 

462 ): 

463 # Fudge the bounds to foil checking for strict monotonicity. 

464 if time_coord.has_bounds(): 464 ↛ 465line 464 didn't jump to line 465 because the condition on line 464 was never true

465 if (time_coord.bounds[-1][0] - time_coord.bounds[0][0]) < 1.0e-8: 

466 time_coord.bounds = [ 

467 [ 

468 time_coord.bounds[i][0] + 1.0e-8 * float(i), 

469 time_coord.bounds[i][1], 

470 ] 

471 for i in range(len(time_coord.bounds)) 

472 ] 

473 iris.util.promote_aux_coord_to_dim_coord(cube, time_coord) 

474 return cube 

475 

476 

477def _grid_longitude_fix_callback(cube: iris.cube.Cube) -> iris.cube.Cube: 

478 """Check grid_longitude coordinates are in the range -180 deg to 180 deg. 

479 

480 This is necessary if comparing two models with different conventions -- 

481 for example, models where the prime meridian is defined as 0 deg or 

482 360 deg. If not in the range -180 deg to 180 deg, we wrap the grid_longitude 

483 so that it falls in this range. Checks are for near-180 bounds given 

484 model data bounds may not extend exactly to 0. or 360. 

485 Input cubes on non-rotated grid coordinates are not impacted. 

486 """ 

487 try: 

488 y, x = get_cube_yxcoordname(cube) 

489 except ValueError: 

490 # Don't modify non-spatial cubes. 

491 return cube 

492 

493 long_coord = cube.coord(x) 

494 # Wrap longitudes if rotated pole coordinates 

495 coord_system = long_coord.coord_system 

496 if x == "grid_longitude" and isinstance( 

497 coord_system, iris.coord_systems.RotatedGeogCS 

498 ): 

499 long_points = long_coord.points.copy() 

500 long_centre = np.median(long_points) 

501 while long_centre < -175.0: 

502 long_centre += 360.0 

503 long_points += 360.0 

504 while long_centre >= 175.0: 

505 long_centre -= 360.0 

506 long_points -= 360.0 

507 long_coord.points = long_points 

508 

509 # Update coord bounds to be consistent with wrapping. 

510 if long_coord.has_bounds(): 

511 long_coord.bounds = None 

512 long_coord.guess_bounds() 

513 

514 return cube 

515 

516 

517def _fix_no_spatial_coords_callback(cube: iris.cube.Cube): 

518 import CSET.operators._utils as utils 

519 

520 # Don't modify spatial cubes that already have spatial dimensions 

521 if utils.is_spatialdim(cube): 

522 return cube 

523 

524 else: 

525 # attempt to get lat/long from cube attributes 

526 try: 

527 lat_min = cube.attributes.get("geospatial_lat_min") 

528 lat_max = cube.attributes.get("geospatial_lat_max") 

529 lon_min = cube.attributes.get("geospatial_lon_min") 

530 lon_max = cube.attributes.get("geospatial_lon_max") 

531 

532 lon_val = (lon_min + lon_max) / 2.0 

533 lat_val = (lat_min + lat_max) / 2.0 

534 

535 lat_coord = iris.coords.DimCoord( 

536 lat_val, 

537 standard_name="latitude", 

538 units="degrees_north", 

539 var_name="latitude", 

540 coord_system=iris.coord_systems.GeogCS(6371229.0), 

541 circular=True, 

542 ) 

543 

544 lon_coord = iris.coords.DimCoord( 

545 lon_val, 

546 standard_name="longitude", 

547 units="degrees_east", 

548 var_name="longitude", 

549 coord_system=iris.coord_systems.GeogCS(6371229.0), 

550 circular=True, 

551 ) 

552 

553 cube.add_aux_coord(lat_coord) 

554 cube.add_aux_coord(lon_coord) 

555 return cube 

556 

557 # if lat/long are not in attributes, then return cube unchanged: 

558 except TypeError: 

559 return cube 

560 

561 

562def _fix_spatial_coords_callback(cube: iris.cube.Cube): 

563 """Check latitude and longitude coordinates name. 

564 

565 This is necessary as some models define their grid as on rotated 

566 'grid_latitude' and 'grid_longitude' coordinates while others define 

567 the grid on non-rotated 'latitude' and 'longitude'. 

568 Cube dimensions need to be made consistent to avoid recipe failures, 

569 particularly where comparing multiple input models with differing spatial 

570 coordinates. 

571 """ 

572 # Check if cube is spatial. 

573 if not is_spatialdim(cube): 

574 # Don't modify non-spatial cubes. 

575 return 

576 

577 # Get spatial coords and dimension index. 

578 y_name, x_name = get_cube_yxcoordname(cube) 

579 ny = get_cube_coordindex(cube, y_name) 

580 nx = get_cube_coordindex(cube, x_name) 

581 

582 # Remove spatial coords bounds if erroneous values detected. 

583 # Aims to catch some errors in input coord bounds by setting 

584 # invalid threshold of 10000.0 

585 if cube.coord(x_name).has_bounds() and cube.coord(y_name).has_bounds(): 

586 bx_max = np.max(np.abs(cube.coord(x_name).bounds)) 

587 by_max = np.max(np.abs(cube.coord(y_name).bounds)) 

588 if bx_max > 10000.0 or by_max > 10000.0: 

589 cube.coord(x_name).bounds = None 

590 cube.coord(y_name).bounds = None 

591 

592 # Translate [grid_latitude, grid_longitude] to an unrotated 1-d DimCoord 

593 # [latitude, longitude] for instances where rotated_pole=90.0 

594 if "grid_latitude" in [coord.name() for coord in cube.coords(dim_coords=True)]: 

595 coord_system = cube.coord("grid_latitude").coord_system 

596 pole_lat = getattr(coord_system, "grid_north_pole_latitude", None) 

597 if pole_lat == 90.0: 597 ↛ 598line 597 didn't jump to line 598 because the condition on line 597 was never true

598 lats = cube.coord("grid_latitude").points 

599 lons = cube.coord("grid_longitude").points 

600 

601 cube.remove_coord("grid_latitude") 

602 cube.add_dim_coord( 

603 iris.coords.DimCoord( 

604 lats, 

605 standard_name="latitude", 

606 var_name="latitude", 

607 units="degrees", 

608 coord_system=iris.coord_systems.GeogCS(6371229.0), 

609 circular=True, 

610 ), 

611 ny, 

612 ) 

613 y_name = "latitude" 

614 cube.remove_coord("grid_longitude") 

615 cube.add_dim_coord( 

616 iris.coords.DimCoord( 

617 lons, 

618 standard_name="longitude", 

619 var_name="longitude", 

620 units="degrees", 

621 coord_system=iris.coord_systems.GeogCS(6371229.0), 

622 circular=True, 

623 ), 

624 nx, 

625 ) 

626 x_name = "longitude" 

627 

628 # Create additional AuxCoord [grid_latitude, grid_longitude] with 

629 # rotated pole attributes for cases with [lat, lon] inputs 

630 if y_name in ["latitude"] and cube.coord(y_name).units in [ 

631 "degrees", 

632 "degrees_north", 

633 "degrees_south", 

634 ]: 

635 # Add grid_latitude AuxCoord 

636 if "grid_latitude" not in [ 636 ↛ 649line 636 didn't jump to line 649 because the condition on line 636 was always true

637 coord.name() for coord in cube.coords(dim_coords=False) 

638 ]: 

639 cube.add_aux_coord( 

640 iris.coords.AuxCoord( 

641 cube.coord(y_name).points, 

642 var_name="grid_latitude", 

643 units="degrees", 

644 ), 

645 ny, 

646 ) 

647 # Ensure input latitude DimCoord has CoordSystem 

648 # This attribute is sometimes lost on iris.save 

649 if not cube.coord(y_name).coord_system: 

650 cube.coord(y_name).coord_system = iris.coord_systems.GeogCS(6371229.0) 

651 

652 if x_name in ["longitude"] and cube.coord(x_name).units in [ 

653 "degrees", 

654 "degrees_west", 

655 "degrees_east", 

656 ]: 

657 # Add grid_longitude AuxCoord 

658 if "grid_longitude" not in [ 658 ↛ 672line 658 didn't jump to line 672 because the condition on line 658 was always true

659 coord.name() for coord in cube.coords(dim_coords=False) 

660 ]: 

661 cube.add_aux_coord( 

662 iris.coords.AuxCoord( 

663 cube.coord(x_name).points, 

664 var_name="grid_longitude", 

665 units="degrees", 

666 ), 

667 nx, 

668 ) 

669 

670 # Ensure input longitude DimCoord has CoordSystem 

671 # This attribute is sometimes lost on iris.save 

672 if not cube.coord(x_name).coord_system: 

673 cube.coord(x_name).coord_system = iris.coord_systems.GeogCS(6371229.0) 

674 

675 

676def _fix_pressure_coord_callback(cube: iris.cube.Cube): 

677 """Rename pressure coordinate to "pressure" if it exists and ensure hPa units. 

678 

679 This problem was raised because the AIFS model data from ECMWF 

680 defines the pressure coordinate with the name "pressure_level" rather 

681 than compliant CF coordinate names. 

682 

683 Additionally, set the units of pressure to be hPa to be consistent with the UM, 

684 and approach the coordinates in a unified way. 

685 """ 

686 for coord in cube.dim_coords: 

687 if coord.name() in ["pressure_level", "pressure_levels"]: 

688 coord.rename("pressure") 

689 

690 if coord.name() == "pressure": 

691 if str(cube.coord("pressure").units) != "hPa": 

692 cube.coord("pressure").convert_units("hPa") 

693 

694 

695def _fix_um_radtime(cube: iris.cube.Cube): 

696 """Move radiation diagnostics from timestamps which are output N minutes or seconds past every hour. 

697 

698 This callback does not have any effect for output diagnostics with 

699 timestamps exactly 00 or 30 minutes past the hour. Only radiation 

700 diagnostics are checked. 

701 Note this callback does not interpolate the data in time, only adjust 

702 timestamps to sit on the hour to enable time-to-time difference plotting 

703 with models which may output radiation data on the hour. 

704 """ 

705 try: 

706 if cube.attributes["STASH"] in [ 

707 "m01s01i207", 

708 "m01s01i208", 

709 "m01s02i205", 

710 "m01s02i201", 

711 "m01s01i207", 

712 "m01s02i207", 

713 "m01s01i235", 

714 ]: 

715 time_coord = cube.coord("time") 

716 

717 # Convert time points to datetime objects 

718 time_unit = time_coord.units 

719 time_points = time_unit.num2date(time_coord.points) 

720 # Skip if times don't need fixing. 

721 if time_points[0].minute == 0 and time_points[0].second == 0: 

722 return 

723 if time_points[0].minute == 30 and time_points[0].second == 0: 723 ↛ 724line 723 didn't jump to line 724 because the condition on line 723 was never true

724 return 

725 

726 # Subtract time difference from the hour from each time point 

727 n_minute = time_points[0].minute 

728 n_second = time_points[0].second 

729 # If times closer to next hour, compute difference to add on to following hour 

730 if n_minute > 30: 

731 n_minute = n_minute - 60 

732 # Compute new diagnostic time stamp 

733 new_time_points = ( 

734 time_points 

735 - datetime.timedelta(minutes=n_minute) 

736 - datetime.timedelta(seconds=n_second) 

737 ) 

738 

739 # Convert back to numeric values using the original time unit. 

740 new_time_values = time_unit.date2num(new_time_points) 

741 

742 # Replace the time coordinate with updated values. 

743 time_coord.points = new_time_values 

744 

745 # Recompute forecast_period with corrected values. 

746 if cube.coord("forecast_period"): 746 ↛ exitline 746 didn't return from function '_fix_um_radtime' because the condition on line 746 was always true

747 fcst_prd_points = cube.coord("forecast_period").points 

748 new_fcst_points = ( 

749 time_unit.num2date(fcst_prd_points) 

750 - datetime.timedelta(minutes=n_minute) 

751 - datetime.timedelta(seconds=n_second) 

752 ) 

753 cube.coord("forecast_period").points = time_unit.date2num( 

754 new_fcst_points 

755 ) 

756 except KeyError: 

757 pass 

758 

759 

760def _fix_cell_methods(cube: iris.cube.Cube): 

761 """To fix the assumed cell_methods in accumulation STASH from UM. 

762 

763 Lightning (m01s21i104), rainfall amount (m01s04i201, m01s05i201) and snowfall amount 

764 (m01s04i202, m01s05i202) in UM is being output as a time accumulation, 

765 over each hour (TAcc1hr), but input cubes show cell_methods as "mean". 

766 For UM and LFRic inputs to be compatible, we assume accumulated cell_methods are 

767 "sum". This callback changes "mean" cube attribute cell_method to "sum", 

768 enabling the cell_method constraint on reading to select correct input. 

769 """ 

770 # Shift "mean" cell_method to "sum" for selected UM inputs. 

771 if cube.attributes.get("STASH") in [ 

772 "m01s21i104", 

773 "m01s04i201", 

774 "m01s04i202", 

775 "m01s05i201", 

776 "m01s05i202", 

777 ]: 

778 # Check if input cell_method contains "mean" time-processing. 

779 if set(cm.method for cm in cube.cell_methods) == {"mean"}: 779 ↛ exitline 779 didn't return from function '_fix_cell_methods' because the condition on line 779 was always true

780 # Retrieve interval and any comment information. 

781 for cell_method in cube.cell_methods: 

782 interval_str = cell_method.intervals 

783 comment_str = cell_method.comments 

784 

785 # Remove input aggregation method. 

786 cube.cell_methods = () 

787 

788 # Replace "mean" with "sum" cell_method to indicate aggregation. 

789 cube.add_cell_method( 

790 iris.coords.CellMethod( 

791 method="sum", 

792 coords="time", 

793 intervals=interval_str, 

794 comments=comment_str, 

795 ) 

796 ) 

797 

798 

799def _convert_cube_units_callback(cube: iris.cube.Cube): 

800 """Adjust diagnostic units for specific variables. 

801 

802 Some precipitation diagnostics are output with unit kg m-2 s-1 and are 

803 converted here to mm hr-1. 

804 

805 Visibility diagnostics are converted here from m to km to improve output 

806 formatting. 

807 """ 

808 # Convert precipitation diagnostic units if required. 

809 varnames = filter(None, [cube.long_name, cube.standard_name, cube.var_name]) 

810 if any("surface_microphysical" in name for name in varnames): 

811 if cube.units == "kg m-2 s-1": 

812 logging.debug( 

813 "Converting precipitation rate units from kg m-2 s-1 to mm hr-1" 

814 ) 

815 # Convert from kg m-2 s-1 to mm s-1 assuming 1kg water = 1l water = 1dm^3 water. 

816 # This is a 1:1 conversion, so we just change the units. 

817 cube.units = "mm s-1" 

818 # Convert the units to per hour. 

819 cube.convert_units("mm hr-1") 

820 elif cube.units == "kg m-2": 820 ↛ 827line 820 didn't jump to line 827 because the condition on line 820 was always true

821 logging.debug("Converting precipitation amount units from kg m-2 to mm") 

822 # Convert from kg m-2 to mm assuming 1kg water = 1l water = 1dm^3 water. 

823 # This is a 1:1 conversion, so we just change the units. 

824 cube.units = "mm" 

825 

826 # Convert visibility diagnostic units if required. 

827 varnames = filter(None, [cube.long_name, cube.standard_name, cube.var_name]) 

828 if any("visibility" in name for name in varnames): 

829 if cube.units == "m": 829 ↛ 834line 829 didn't jump to line 834 because the condition on line 829 was always true

830 logging.debug("Converting visibility units m to km.") 

831 # Convert the units to km. 

832 cube.convert_units("km") 

833 

834 return cube 

835 

836 

837def _fix_lfric_cloud_base_altitude(cube: iris.cube.Cube): 

838 """Mask cloud_base_altitude diagnostic in regions with no cloud.""" 

839 varnames = filter(None, [cube.long_name, cube.standard_name, cube.var_name]) 

840 if any("cloud_base_altitude" in name for name in varnames): 

841 # Mask cube where set > 144kft to catch default 144.35695538058164 

842 cube.data = np.ma.masked_array(cube.data) 

843 cube.data[cube.data > 144.0] = np.ma.masked 

844 

845 

846def _fix_um_winds(cubes: iris.cube.CubeList): 

847 """To make winds from the UM consistent with those from LFRic. 

848 

849 Diagnostics of wind are not always consistent between the UM 

850 and LFric. Here, winds from the UM are adjusted to make them i 

851 consistent with LFRic. 

852 """ 

853 # Check whether we have components of the wind identified by STASH, 

854 # (so this will apply only to cubes from the UM), but not the 

855 # wind speed and calculate it if it is missing. Note that 

856 # this will be biased low in general because the components will mostly 

857 # be time averages. For simplicity, we do this only if there is just one 

858 # cube of a component. A more complicated approach would be to consider 

859 # the cell methods, but it may not be warranted. 

860 u_constr = iris.AttributeConstraint(STASH="m01s03i225") 

861 v_constr = iris.AttributeConstraint(STASH="m01s03i226") 

862 speed_constr = iris.AttributeConstraint(STASH="m01s03i227") 

863 try: 

864 if cubes.extract(u_constr) and cubes.extract(v_constr): 864 ↛ 865line 864 didn't jump to line 865 because the condition on line 864 was never true

865 if len(cubes.extract(u_constr)) == 1 and not cubes.extract(speed_constr): 

866 _add_wind_speed_um(cubes) 

867 # Convert winds in the UM to be relative to true east and true north. 

868 _convert_wind_true_dirn_um(cubes) 

869 except (KeyError, AttributeError): 

870 pass 

871 

872 

873def _add_wind_speed_um(cubes: iris.cube.CubeList): 

874 """Add windspeeds to cubes from the UM.""" 

875 wspd10 = ( 

876 cubes.extract_cube(iris.AttributeConstraint(STASH="m01s03i225"))[0] ** 2 

877 + cubes.extract_cube(iris.AttributeConstraint(STASH="m01s03i226"))[0] ** 2 

878 ) ** 0.5 

879 wspd10.attributes["STASH"] = "m01s03i227" 

880 wspd10.standard_name = "wind_speed" 

881 wspd10.long_name = "wind_speed_at_10m" 

882 cubes.append(wspd10) 

883 

884 

885def _convert_wind_true_dirn_um(cubes: iris.cube.CubeList): 

886 """To convert winds to true directions. 

887 

888 Convert from the components relative to the grid to true directions. 

889 This functionality only handles the simplest case. 

890 """