Coverage for src/pydal2sql_core/cli_support.py: 100%

314 statements  

« prev     ^ index     » next       coverage.py v7.2.7, created at 2023-11-20 12:05 +0100

1""" 

2CLI-Agnostic support. 

3""" 

4import contextlib 

5import io 

6import os 

7import re 

8import select 

9import string 

10import sys 

11import textwrap 

12import traceback 

13import typing 

14from datetime import datetime 

15from pathlib import Path 

16from typing import Any, Optional 

17 

18import git 

19import gitdb.exc 

20import rich 

21from black.files import find_project_root 

22from git.objects.blob import Blob 

23from git.objects.commit import Commit 

24from git.repo import Repo 

25from witchery import ( 

26 add_function_call, 

27 find_defined_variables, 

28 find_function_to_call, 

29 find_missing_variables, 

30 generate_magic_code, 

31 has_local_imports, 

32 remove_if_falsey_blocks, 

33 remove_import, 

34 remove_local_imports, 

35 remove_specific_variables, 

36) 

37 

38from .helpers import flatten 

39from .types import ( 

40 _SUPPORTED_OUTPUT_FORMATS, 

41 DEFAULT_OUTPUT_FORMAT, 

42 SUPPORTED_DATABASE_TYPES_WITH_ALIASES, 

43 SUPPORTED_OUTPUT_FORMATS, 

44 DummyDAL, 

45 DummyTypeDAL, 

46) 

47 

48 

49def has_stdin_data() -> bool: # pragma: no cover 

50 """ 

51 Check if the program starts with cli data (pipe | or redirect <). 

52 

53 Returns: 

54 bool: True if the program starts with cli data, False otherwise. 

55 

56 See Also: 

57 https://stackoverflow.com/questions/3762881/how-do-i-check-if-stdin-has-some-data 

58 """ 

59 return any( 

60 select.select( 

61 [ 

62 sys.stdin, 

63 ], 

64 [], 

65 [], 

66 0.0, 

67 )[0] 

68 ) 

69 

70 

71AnyCallable = typing.Callable[..., Any] 

72 

73 

74def print_if_interactive(*args: Any, pretty: bool = True, **kwargs: Any) -> None: # pragma: no cover 

75 """ 

76 Print the given arguments if running in an interactive session. 

77 

78 Args: 

79 *args: Variable length list of arguments to be printed. 

80 pretty (bool): If True, print using rich library's rich.print, otherwise use the built-in print function. 

81 **kwargs: Optional keyword arguments to be passed to the print function. 

82 

83 Returns: 

84 None 

85 """ 

86 is_interactive = not has_stdin_data() 

87 _print = typing.cast(AnyCallable, rich.print if pretty else print) # make mypy happy 

88 if is_interactive: 

89 kwargs["file"] = sys.stderr 

90 _print( 

91 *args, 

92 **kwargs, 

93 ) 

94 

95 

96def find_git_root(at: str = None) -> Optional[Path]: 

97 """ 

98 Find the root directory of the Git repository. 

99 

100 Args: 

101 at (str, optional): The directory path to start the search. Defaults to the current working directory. 

102 

103 Returns: 

104 Optional[Path]: The root directory of the Git repository if found, otherwise None. 

105 """ 

106 folder, reason = find_project_root((at or os.getcwd(),)) 

107 if reason != ".git directory": 

108 return None 

109 return folder 

110 

111 

112def find_git_repo(repo: Repo = None, at: str = None) -> Repo: 

113 """ 

114 Find the Git repository instance. 

115 

116 Args: 

117 repo (Repo, optional): An existing Git repository instance. If provided, returns the same instance. 

118 at (str, optional): The directory path to start the search. Defaults to the current working directory. 

119 

120 Returns: 

121 Repo: The Git repository instance. 

122 """ 

123 if repo: 

124 return repo 

125 

126 root = find_git_root(at) 

127 return Repo(str(root)) 

128 

129 

130def latest_commit(repo: Repo = None) -> Commit: 

131 """ 

132 Get the latest commit in the Git repository. 

133 

134 Args: 

135 repo (Repo, optional): An existing Git repository instance. If provided, uses the given instance. 

136 

137 Returns: 

138 Commit: The latest commit in the Git repository. 

139 """ 

140 repo = find_git_repo(repo) 

141 return repo.head.commit 

142 

143 

144def commit_by_id(commit_hash: str, repo: Repo = None) -> Commit: 

145 """ 

146 Get a specific commit in the Git repository by its hash or name. 

147 

148 Args: 

149 commit_hash (str): The hash of the commit to retrieve. Can also be e.g. a branch name. 

150 repo (Repo, optional): An existing Git repository instance. If provided, uses the given instance. 

151 

152 Returns: 

153 Commit: The commit object corresponding to the given commit hash. 

154 """ 

155 repo = find_git_repo(repo) 

156 return repo.commit(commit_hash) 

157 

158 

159@contextlib.contextmanager 

160def open_blob(file: Blob) -> typing.Generator[io.BytesIO, None, None]: 

161 """ 

162 Open a Git Blob object as a context manager, providing access to its data. 

163 

164 Args: 

165 file (Blob): The Git Blob object to open. 

166 

167 Yields: 

168 io.BytesIO: A BytesIO object providing access to the Blob data. 

169 """ 

170 yield io.BytesIO(file.data_stream.read()) 

171 

172 

173def read_blob(file: Blob) -> str: 

174 """ 

175 Read the contents of a Git Blob object and decode it as a string. 

176 

177 Args: 

178 file (Blob): The Git Blob object to read. 

179 

180 Returns: 

181 str: The contents of the Blob as a string. 

182 """ 

183 with open_blob(file) as f: 

184 return f.read().decode() 

185 

186 

187def get_file_for_commit(filename: str, commit_version: str = "latest", repo: Repo = None) -> str: 

188 """ 

189 Get the contents of a file in the Git repository at a specific commit version. 

190 

191 Args: 

192 filename (str): The path of the file to retrieve. 

193 commit_version (str, optional): The commit hash or branch name. Defaults to "latest" (latest commit). 

194 repo (Repo, optional): An existing Git repository instance. If provided, uses the given instance. 

195 

196 Returns: 

197 str: The contents of the file as a string. 

198 """ 

199 repo = find_git_repo(repo, at=filename) 

200 commit = latest_commit(repo) if commit_version == "latest" else commit_by_id(commit_version, repo) 

201 

202 file_path = str(Path(filename).resolve()) 

203 # relative to the .git folder: 

204 relative_file_path = file_path.removeprefix(f"{repo.working_dir}/") 

205 

206 file_at_commit = commit.tree / relative_file_path 

207 return read_blob(file_at_commit) 

208 

209 

210def get_file_for_version(filename: str, version: str, prompt_description: str = "", with_git: bool = True) -> str: 

211 """ 

212 Get the contents of a file based on the version specified. 

213 

214 Args: 

215 filename (str): The path of the file to retrieve. 

216 version (str): The version specifier, which can be "current", "stdin", or a commit hash/branch name. 

217 prompt_description (str, optional): A description to display when asking for input from stdin. 

218 

219 Returns: 

220 str: The contents of the file as a string. 

221 """ 

222 if not with_git or version == "current": 

223 return Path(filename).read_text() 

224 elif version == "stdin": # pragma: no cover 

225 print_if_interactive( 

226 f"[blue]Please paste your define tables ({prompt_description}) code below " 

227 f"and press ctrl-D when finished.[/blue]", 

228 file=sys.stderr, 

229 ) 

230 result = sys.stdin.read() 

231 print_if_interactive("[blue]---[/blue]", file=sys.stderr) 

232 return result 

233 elif with_git: 

234 try: 

235 return get_file_for_commit(filename, version) 

236 except (git.exc.GitError, gitdb.exc.ODBError) as e: 

237 raise FileNotFoundError(f"{filename}@{version}") from e 

238 

239 

240def extract_file_version_and_path( 

241 file_path_or_git_tag: Optional[str], default_version: str = "stdin" 

242) -> tuple[str, str | None]: 

243 """ 

244 Extract the file version and path from the given input. 

245 

246 Args: 

247 file_path_or_git_tag (str, optional): The input string containing the file path and/or Git tag. 

248 default_version (str, optional): The default version to use if no version is specified. Defaults to "stdin". 

249 

250 Returns: 

251 tuple[str, str | None]: A tuple containing the extracted version and file path (or None if not specified). 

252 

253 Examples: 

254 myfile.py (implies @current) 

255 

256 myfile.py@latest 

257 myfile.py@my-branch 

258 myfile.py@b3f24091a9 

259 

260 @latest (implies no path, e.g. in case of ALTER to copy previously defined path) 

261 """ 

262 if not file_path_or_git_tag: 

263 return default_version, "" 

264 

265 if file_path_or_git_tag == "-": 

266 return "stdin", "-" 

267 

268 if file_path_or_git_tag.startswith("@"): 

269 file_version = file_path_or_git_tag.strip("@") 

270 file_path = None 

271 elif "@" in file_path_or_git_tag: 

272 file_path, file_version = file_path_or_git_tag.split("@") 

273 else: 

274 file_version = default_version # `latest` for before; `current` for after. 

275 file_path = file_path_or_git_tag 

276 

277 return file_version, file_path 

278 

279 

280def extract_file_versions_and_paths( 

281 filename_before: Optional[str], filename_after: Optional[str] 

282) -> tuple[tuple[str, str | None], tuple[str, str | None]]: 

283 """ 

284 Extract the file versions and paths based on the before and after filenames. 

285 

286 Args: 

287 filename_before (str, optional): The path of the file before the change (or None). 

288 filename_after (str, optional): The path of the file after the change (or None). 

289 

290 Returns: 

291 tuple[tuple[str, str | None], tuple[str, str | None]]: 

292 A tuple of two tuples, each containing the version and path of the before and after files. 

293 """ 

294 version_before, filepath_before = extract_file_version_and_path( 

295 filename_before, 

296 default_version="current" 

297 if filename_after and filename_before and filename_after != filename_before 

298 else "latest", 

299 ) 

300 version_after, filepath_after = extract_file_version_and_path(filename_after, default_version="current") 

301 

302 if not (filepath_before or filepath_after): 

303 raise ValueError("Please supply at least one file name.") 

304 elif not filepath_after: 

305 filepath_after = filepath_before 

306 elif not filepath_before: 

307 filepath_before = filepath_after 

308 

309 return (version_before, filepath_before), (version_after, filepath_after) 

310 

311 

312def get_absolute_path_info(filename: Optional[str], version: str, git_root: Optional[Path] = None) -> tuple[bool, str]: 

313 """ 

314 Get absolute path information for the file based on the version and Git root. 

315 

316 Args: 

317 filename (str, optional): The path of the file to check (or None). 

318 version (str): The version specifier, which can be "stdin", "current", or a commit hash/branch name. 

319 git_root (Path, optional): The root directory of the Git repository. If None, it will be determined. 

320 

321 Returns: 

322 tuple[bool, str]: A tuple containing a boolean indicating if the file exists and the absolute path to the file. 

323 """ 

324 if version == "stdin": 

325 return True, "" 

326 elif filename is None: 

327 # can't deal with this, not stdin and no file should show file missing error later. 

328 return False, "" 

329 

330 if git_root is None: 

331 git_root = find_git_root() or Path(os.getcwd()) 

332 

333 path = Path(filename) 

334 path_via_git = git_root / filename 

335 

336 if path.exists(): 

337 exists = True 

338 absolute_path = str(path.resolve()) 

339 elif path_via_git.exists(): 

340 exists = True 

341 absolute_path = str(path_via_git.resolve()) 

342 else: 

343 exists = False 

344 absolute_path = "" 

345 

346 return exists, absolute_path 

347 

348 

349def ensure_no_migrate_on_real_db( 

350 code: str, db_names: typing.Iterable[str] = ("db", "database"), fix: typing.Optional[bool] = False 

351) -> str: 

352 """ 

353 Ensure that the code does not contain actual migrations on a real database. 

354 

355 It does this by removing definitions of 'db' and database. This can be changed by customizing `db_names`. 

356 It also removes local imports to prevent irrelevant code being executed. 

357 

358 Args: 

359 code (str): The code to check for database migrations. 

360 db_names (Iterable[str], optional): Names of variables representing the database. 

361 Defaults to ("db", "database"). 

362 fix (bool, optional): If True, removes the migration code. Defaults to False. 

363 

364 Returns: 

365 str: The modified code with migration code removed if fix=True, otherwise the original code. 

366 """ 

367 variables = find_defined_variables(code) 

368 

369 found_variables = set() 

370 

371 for db_name in db_names: 

372 if db_name in variables: 

373 if fix: 

374 code = remove_specific_variables(code, db_names) 

375 else: 

376 found_variables.add(db_name) 

377 

378 if found_variables: 

379 if len(found_variables) == 1: 

380 var = next(iter(found_variables)) 

381 message = f"Variable {var} defined in code! " 

382 else: # pragma: no cover 

383 var = ", ".join(found_variables) 

384 message = f"Variables {var} defined in code! " 

385 raise ValueError( 

386 f"{message} Please remove this or use --magic to prevent performing actual migrations on your database." 

387 ) 

388 

389 if has_local_imports(code): 

390 if fix: 

391 code = remove_local_imports(code) 

392 else: 

393 raise ValueError("Local imports are used in this file! Please remove these or use --magic.") 

394 

395 return code 

396 

397 

398MAX_RETRIES = 30 

399 

400# todo: overload more methods 

401 

402TEMPLATE_PYDAL = """ 

403from pydal import * 

404from pydal.objects import * 

405from pydal.validators import * 

406 

407from pydal2sql_core import generate_sql 

408 

409 

410# from pydal import DAL 

411db = database = DummyDAL(None, migrate=False) 

412 

413tables = $tables 

414db_type = '$db_type' 

415 

416$extra 

417 

418$code_before 

419 

420db_old = db 

421db_new = db = database = DummyDAL(None, migrate=False) 

422 

423$extra 

424 

425$code_after 

426 

427if not tables: 

428 tables = set(db_old._tables + db_new._tables) - _special_tables 

429 

430if not tables: 

431 raise ValueError('no-tables-found') 

432 

433for table in tables: 

434 print('-- start ', table, '--', file=_file) 

435 if table in db_old and table in db_new: 

436 print(generate_sql(db_old[table], db_new[table], db_type=db_type), file=_file) 

437 elif table in db_old: 

438 print(f'DROP TABLE {table};', file=_file) 

439 else: 

440 print(generate_sql(db_new[table], db_type=db_type), file=_file) 

441 print('-- END OF MIGRATION --', file=_file) 

442 """ 

443 

444TEMPLATE_TYPEDAL = """ 

445from pydal import * 

446from pydal.objects import * 

447from pydal.validators import * 

448from typedal import * 

449 

450from pydal2sql_core import generate_sql 

451 

452 

453# from typedal import TypeDAL as DAL 

454db = database = DummyDAL(None, migrate=False) 

455 

456tables = $tables 

457db_type = '$db_type' 

458 

459$extra 

460 

461$code_before 

462 

463db_old = db 

464db_new = db = database = DummyDAL(None, migrate=False) 

465 

466$extra 

467 

468$code_after 

469 

470if not tables: 

471 tables = set(db_old._tables + db_new._tables) - _special_tables 

472 

473if not tables: 

474 raise ValueError('no-tables-found') 

475 

476 

477for table in tables: 

478 print('-- start ', table, '--', file=_file) 

479 if table in db_old and table in db_new: 

480 print(generate_sql(db_old[table], db_new[table], db_type=db_type), file=_file) 

481 elif table in db_old: 

482 print(f'DROP TABLE {table};', file=_file) 

483 else: 

484 print(generate_sql(db_new[table], db_type=db_type), file=_file) 

485 

486 print('-- END OF MIGRATION --', file=_file) 

487 """ 

488 

489 

490def sql_to_function_name(sql_statement: str) -> str: 

491 """ 

492 Extract action (CREATE, ALTER, DROP) and table name from the SQL statement. 

493 """ 

494 match = re.findall(r"(CREATE|ALTER|DROP)\s+TABLE\s+['\"]?(\w+)['\"]?", sql_statement.lower(), re.IGNORECASE) 

495 

496 if not match: 

497 # raise ValueError("Invalid SQL statement. Unable to extract action and table name.") 

498 return "unknown_migration" 

499 

500 action, table_name = match[0] 

501 

502 # Generate a function name with the specified format 

503 return f"{action}_{table_name}" 

504 

505 

506def _setup_generic_edwh_migrate(file: Path, is_typedal: bool) -> None: 

507 contents = ( 

508 "from edwh_migrate import migration\n" 

509 + ("from typedal import TypeDAL" if is_typedal else "from pydal import DAL") 

510 + "\n" 

511 ) 

512 

513 with file.open("w") as f: 

514 f.write(textwrap.dedent(contents)) 

515 

516 rich.print(f"[green] New migrate file {file} created [/green]") 

517 

518 

519def _build_edwh_migration(contents: str, cls: str, date: str, existing: Optional[str] = None) -> str: 

520 sql_func_name = sql_to_function_name(contents) 

521 func_name = "_placeholder_" 

522 

523 for n in range(1, 1000): 

524 func_name = f"{sql_func_name}_{date}_{str(n).zfill(3)}" 

525 

526 if existing and f"def {func_name}" in existing: 

527 if contents.replace(" ", "").replace("\n", "") in existing.replace(" ", "").replace("\n", ""): 

528 rich.print(f"[yellow] migration {func_name} already exists, skipping! [/yellow]") 

529 return "" 

530 elif func_name.startswith("alter"): 

531 # bump number because alter migrations are different 

532 continue 

533 else: 

534 rich.print( 

535 f"[red] migration {func_name} already exists [bold]with different contents[/bold], skipping! [/red]" 

536 ) 

537 return "" 

538 else: 

539 # okay function name, stop incrementing 

540 break 

541 

542 contents = textwrap.indent(contents.strip(), " " * 16) 

543 return textwrap.dedent( 

544 f''' 

545 

546 @migration 

547 def {func_name}(db: {cls}): 

548 db.executesql(""" 

549{contents} 

550 """) 

551 db.commit() 

552 

553 return True 

554 ''' 

555 ) 

556 

557 

558def _build_edwh_migrations(contents: str, is_typedal: bool, output: Optional[Path] = None) -> str: 

559 cls = "TypeDAL" if is_typedal else "DAL" 

560 date = datetime.now().strftime("%Y%m%d") # yyyymmdd 

561 

562 existing = output.read_text() if output and output.exists() else None 

563 

564 return "".join( 

565 _build_edwh_migration(migration, cls, date, existing) 

566 for migration in contents.split("-- END OF MIGRATION --") 

567 if migration.strip() 

568 ) 

569 

570 

571def _handle_output( 

572 file: io.StringIO, 

573 output_file: Path | str | io.StringIO | None, 

574 output_format: SUPPORTED_OUTPUT_FORMATS = DEFAULT_OUTPUT_FORMAT, 

575 is_typedal: bool = False, 

576) -> None: 

577 file.seek(0) 

578 contents = file.read() 

579 

580 if isinstance(output_file, str): 

581 output_file = Path(output_file) 

582 

583 if output_format == "edwh-migrate": 

584 contents = _build_edwh_migrations(contents, is_typedal, output_file if isinstance(output_file, Path) else None) 

585 elif output_format in {"default", "sql"} or not output_format: 

586 contents = "\n".join(contents.split("-- END OF MIGRATION --")) 

587 else: 

588 raise ValueError( 

589 f"Unknown format {output_format}. " f"Please choose one of {typing.get_args(_SUPPORTED_OUTPUT_FORMATS)}" 

590 ) 

591 

592 if isinstance(output_file, Path): 

593 if output_format == "edwh-migrate" and (not output_file.exists() or output_file.stat().st_size == 0): 

594 _setup_generic_edwh_migrate(output_file, is_typedal) 

595 

596 if contents.strip(): 

597 with output_file.open("a") as f: 

598 f.write(contents) 

599 

600 rich.print(f"[green] Written migration(s) to {output_file} [/green]") 

601 else: 

602 rich.print(f"[yellow] Nothing to write to {output_file} [/yellow]") 

603 

604 elif isinstance(output_file, io.StringIO): 

605 output_file.write(contents) 

606 else: 

607 # no file, just print to stdout: 

608 print(contents.strip()) 

609 

610 

611IMPORT_IN_STR = re.compile(r'File "<string>", line (\d+), in <module>') 

612 

613 

614def _handle_import_error(code: str, error: ImportError) -> str: 

615 # error is deeper in a package, find the related import in the code: 

616 tb_lines = traceback.format_exc().splitlines() 

617 

618 for line in tb_lines: 

619 if matches := IMPORT_IN_STR.findall(line): 

620 # 'File "<string>", line 15, in <module>' 

621 line_no = int(matches[0]) - 1 

622 lines = code.split("\n") 

623 return lines[line_no] 

624 

625 # I don't know how to trigger this case: 

626 raise ValueError("Faulty import could not be automatically deleted") from error # pragma: no cover 

627 

628 

629MISSING_RELATIONSHIP = re.compile(r"Cannot resolve reference (\w+) in \w+ definition") 

630 

631 

632def _handle_relation_error(error: KeyError) -> tuple[str, str]: 

633 if not (table := MISSING_RELATIONSHIP.findall(str(error))): 

634 # other error, raise again 

635 raise error 

636 

637 t = table[0] 

638 

639 return ( 

640 t, 

641 """ 

642 db.define_table('%s', redefine=True) 

643 """ 

644 % t, 

645 ) 

646 

647 

648def handle_cli( 

649 code_before: str, 

650 code_after: str, 

651 db_type: Optional[str] = None, 

652 tables: Optional[list[str] | list[list[str]]] = None, 

653 verbose: Optional[bool] = False, 

654 noop: Optional[bool] = False, 

655 magic: Optional[bool] = False, 

656 function_name: Optional[str | tuple[str, ...]] = "define_tables", 

657 use_typedal: bool | typing.Literal["auto"] = "auto", 

658 output_format: SUPPORTED_OUTPUT_FORMATS = DEFAULT_OUTPUT_FORMAT, 

659 output_file: Optional[str | Path | io.StringIO] = None, 

660) -> bool: 

661 """ 

662 Handle user input for generating SQL migration statements based on before and after code. 

663 

664 Args: 

665 code_before (str): The code representing the state of the database before the change. 

666 code_after (str, optional): The code representing the state of the database after the change. 

667 db_type (str, optional): The type of the database (e.g., "postgres", "mysql", etc.). Defaults to None. 

668 tables (list[str] or list[list[str]], optional): The list of tables to generate SQL for. Defaults to None. 

669 verbose (bool, optional): If True, print the generated code. Defaults to False. 

670 noop (bool, optional): If True, only print the generated code but do not execute it. Defaults to False. 

671 magic (bool, optional): If True, automatically add missing variables for execution. Defaults to False. 

672 function_name (str, optional): The name of the function where the tables are defined. Defaults: "define_tables". 

673 use_typedal: replace pydal imports with TypeDAL? 

674 output_format: defaults to just SQL, edwh-migrate migration syntax also supported 

675 output_file: append the output to a file instead of printing it? 

676 

677 # todo: prefix (e.g. public.) 

678 

679 Returns: 

680 bool: True if SQL migration statements are generated and executed successfully, False otherwise. 

681 """ 

682 # todo: better typedal checking 

683 if use_typedal == "auto": 

684 use_typedal = "typedal" in code_before.lower() or "typedal" in code_after.lower() 

685 

686 if function_name: 

687 define_table_functions: set[str] = set(function_name) if isinstance(function_name, tuple) else {function_name} 

688 else: 

689 define_table_functions = set() 

690 

691 template = TEMPLATE_TYPEDAL if use_typedal else TEMPLATE_PYDAL 

692 

693 to_execute = string.Template(textwrap.dedent(template)) 

694 

695 code_before = ensure_no_migrate_on_real_db(code_before, fix=magic) 

696 code_after = ensure_no_migrate_on_real_db(code_after, fix=magic) 

697 extra_code = "" 

698 

699 generated_code = to_execute.substitute( 

700 { 

701 "tables": flatten(tables or []), 

702 "db_type": db_type or "", 

703 "code_before": textwrap.dedent(code_before), 

704 "code_after": textwrap.dedent(code_after), 

705 "extra": extra_code, 

706 } 

707 ) 

708 if verbose or noop: 

709 rich.print(generated_code, file=sys.stderr) 

710 

711 if noop: 

712 # done 

713 return True 

714 

715 err: typing.Optional[Exception] = None 

716 catch: dict[str, Any] = {} 

717 retry_counter = MAX_RETRIES 

718 

719 magic_vars = {"_file", "DummyDAL", "_special_tables"} 

720 special_tables: set[str] = {"typedal_cache", "typedal_cache_dependency"} if use_typedal else set() 

721 

722 while retry_counter: 

723 retry_counter -= 1 

724 try: 

725 if verbose: 

726 rich.print(generated_code, file=sys.stderr) 

727 

728 # 'catch' is used to add and receive globals from the exec scope. 

729 # another argument could be added for locals, but adding simply {} changes the behavior negatively. 

730 # so for now, only globals is passed. 

731 catch["_file"] = io.StringIO() # <- every print should go to this file, so we can handle it afterwards 

732 catch["DummyDAL"] = ( 

733 DummyTypeDAL if use_typedal else DummyDAL 

734 ) # <- use a fake DAL that doesn't actually run queries 

735 catch["_special_tables"] = special_tables # <- e.g. typedal_cache, auth_user 

736 # note: when adding something to 'catch', also add it to magic_vars!!! 

737 

738 exec(generated_code, catch) # nosec: B102 

739 _handle_output(catch["_file"], output_file, output_format, is_typedal=use_typedal) 

740 return True # success! 

741 except ValueError as e: 

742 if str(e) != "no-tables-found": # pragma: no cover 

743 rich.print(f"[yellow]{e}[/yellow]", file=sys.stderr) 

744 return False 

745 

746 if define_table_functions: 

747 any_found = False 

748 for function_name in define_table_functions: 

749 define_tables = find_function_to_call(generated_code, function_name) 

750 

751 # if define_tables function is found, add call to it at end of code 

752 if define_tables is not None: 

753 generated_code = add_function_call(generated_code, function_name, multiple=True) 

754 any_found = True 

755 

756 if any_found: 

757 # hurray! 

758 continue 

759 

760 # else: no define_tables or other method to use found. 

761 

762 print(f"No tables found in the top-level or {function_name} function!", file=sys.stderr) 

763 if use_typedal: 

764 print( 

765 "Please use `db.define` or `database.define`, " 

766 "or if you really need to use an alias like my_db.define, " 

767 "add `my_db = db` at the top of the file or pass `--db-name mydb`.", 

768 file=sys.stderr, 

769 ) 

770 else: 

771 print( 

772 "Please use `db.define_table` or `database.define_table`, " 

773 "or if you really need to use an alias like my_db.define_tables, " 

774 "add `my_db = db` at the top of the file or pass `--db-name mydb`.", 

775 file=sys.stderr, 

776 ) 

777 print(f"You can also specify a --function to use something else than {function_name}.", file=sys.stderr) 

778 

779 return False 

780 

781 except NameError as e: 

782 err = e 

783 # something is missing! 

784 missing_vars = find_missing_variables(generated_code) - magic_vars 

785 if not magic: 

786 rich.print( 

787 f"Your code is missing some variables: {missing_vars}. Add these or try --magic", 

788 file=sys.stderr, 

789 ) 

790 return False 

791 

792 # postponed: this can possibly also be achieved by updating the 'catch' dict 

793 # instead of injecting in the string. 

794 extra_code = extra_code + "\n" + textwrap.dedent(generate_magic_code(missing_vars)) 

795 

796 code_before = remove_if_falsey_blocks(code_before) 

797 code_after = remove_if_falsey_blocks(code_after) 

798 

799 generated_code = to_execute.substitute( 

800 { 

801 "tables": flatten(tables or []), 

802 "db_type": db_type or "", 

803 "extra": textwrap.dedent(extra_code), 

804 "code_before": textwrap.dedent(code_before), 

805 "code_after": textwrap.dedent(code_after), 

806 } 

807 ) 

808 except ImportError as e: 

809 # should include ModuleNotFoundError 

810 err = e 

811 # if we catch an ImportError, we try to remove the import and retry 

812 if not e.path: 

813 # code exists in code itself 

814 code_before = remove_import(code_before, e.name or "") 

815 code_after = remove_import(code_after, e.name or "") 

816 else: 

817 to_remove = _handle_import_error(generated_code, e) 

818 code_before = code_before.replace(to_remove, "\n") 

819 code_after = code_after.replace(to_remove, "\n") 

820 

821 generated_code = to_execute.substitute( 

822 { 

823 "tables": flatten(tables or []), 

824 "db_type": db_type or "", 

825 "extra": textwrap.dedent(extra_code), 

826 "code_before": textwrap.dedent(code_before), 

827 "code_after": textwrap.dedent(code_after), 

828 } 

829 ) 

830 

831 except KeyError as e: 

832 err = e 

833 table_name, table_definition = _handle_relation_error(e) 

834 special_tables.add(table_name) 

835 extra_code = extra_code + "\n" + textwrap.dedent(table_definition) 

836 

837 generated_code = to_execute.substitute( 

838 { 

839 "tables": flatten(tables or []), 

840 "db_type": db_type or "", 

841 "extra": textwrap.dedent(extra_code), 

842 "code_before": textwrap.dedent(code_before), 

843 "code_after": textwrap.dedent(code_after), 

844 } 

845 ) 

846 except Exception as e: 

847 err = e 

848 # otherwise: give up 

849 retry_counter = 0 

850 finally: 

851 # reset: 

852 typing.TYPE_CHECKING = False 

853 

854 if retry_counter < 1: # pragma: no cover 

855 rich.print(f"[red]Code could not be fixed automagically![/red]. Error: {err or '?'}", file=sys.stderr) 

856 return False 

857 

858 # idk when this would happen, but something definitely went wrong here: 

859 return False # pragma: no cover 

860 

861 

862def core_create( 

863 filename: Optional[str] = None, 

864 tables: Optional[list[str]] = None, 

865 db_type: Optional[SUPPORTED_DATABASE_TYPES_WITH_ALIASES] = None, 

866 magic: bool = False, 

867 noop: bool = False, 

868 verbose: bool = False, 

869 function: Optional[str | tuple[str, ...]] = None, 

870 output_format: Optional[SUPPORTED_OUTPUT_FORMATS] = DEFAULT_OUTPUT_FORMAT, 

871 output_file: Optional[str | Path] = None, 

872) -> bool: 

873 """ 

874 Generates SQL migration statements for creating one or more tables, based on the code in a given source file. 

875 

876 Args: 

877 filename: The filename of the source file to parse. This code represents the final state of the database. 

878 tables: A list of table names to generate SQL for. 

879 If None, the function will attempt to process all tables found in the code. 

880 db_type: The type of the database. If None, the function will attempt to infer it from the code. 

881 magic: If True, automatically add missing variables for execution. 

882 noop: If True, only print the generated code but do not execute it. 

883 verbose: If True, print the generated code and additional debug information. 

884 function: The name of the function where the tables are defined. 

885 If None, the function will use 'define_tables'. 

886 output_format: defaults to just SQL, edwh-migrate migration syntax also supported 

887 output_file: append the output to a file instead of printing it? 

888 

889 Returns: 

890 bool: True if SQL migration statements are generated and (if not in noop mode) executed successfully, 

891 False otherwise. 

892 

893 Raises: 

894 ValueError: If the source file cannot be found or if no tables could be found in the code. 

895 """ 

896 git_root = find_git_root() or Path(os.getcwd()) 

897 

898 functions: set[str] = set() 

899 if function: # pragma: no cover 

900 if isinstance(function, tuple): 

901 functions.update(function) 

902 else: 

903 functions.add(function) 

904 

905 if filename and ":" in filename: 

906 # e.g. models.py:define_tables 

907 filename, _function = filename.split(":", 1) 

908 functions.add(_function) 

909 

910 file_version, file_path = extract_file_version_and_path( 

911 filename, default_version="current" if filename else "stdin" 

912 ) 

913 file_exists, file_absolute_path = get_absolute_path_info(file_path, file_version, git_root) 

914 

915 if not file_exists: 

916 raise FileNotFoundError(f"Source file {filename} could not be found.") 

917 

918 text = get_file_for_version(file_absolute_path, file_version, prompt_description="table definition") 

919 

920 return handle_cli( 

921 "", 

922 text, 

923 db_type=db_type, 

924 tables=tables, 

925 verbose=verbose, 

926 noop=noop, 

927 magic=magic, 

928 function_name=tuple(functions), 

929 output_format=output_format, 

930 output_file=output_file, 

931 ) 

932 

933 

934def core_alter( 

935 filename_before: Optional[str] = None, 

936 filename_after: Optional[str] = None, 

937 tables: Optional[list[str]] = None, 

938 db_type: Optional[SUPPORTED_DATABASE_TYPES_WITH_ALIASES] = None, 

939 magic: bool = False, 

940 noop: bool = False, 

941 verbose: bool = False, 

942 function: Optional[str] = None, 

943 output_format: Optional[SUPPORTED_OUTPUT_FORMATS] = DEFAULT_OUTPUT_FORMAT, 

944 output_file: Optional[str | Path] = None, 

945) -> bool: 

946 """ 

947 Generates SQL migration statements for altering the database, based on the code in two given source files. 

948 

949 Args: 

950 filename_before: The filename of the source file before changes. 

951 This code represents the initial state of the database. 

952 filename_after: The filename of the source file after changes. 

953 This code represents the final state of the database. 

954 tables: A list of table names to generate SQL for. 

955 If None, the function will attempt to process all tables found in the code. 

956 db_type: The type of the database. If None, the function will attempt to infer it from the code. 

957 magic: If True, automatically add missing variables for execution. 

958 noop: If True, only print the generated code but do not execute it. 

959 verbose: If True, print the generated code and additional debug information. 

960 function: The name of the function where the tables are defined. 

961 If None, the function will use 'define_tables'. 

962 output_format: defaults to just SQL, edwh-migrate migration syntax also supported 

963 output_file: append the output to a file instead of printing it? 

964 

965 Returns: 

966 bool: True if SQL migration statements are generated and (if not in noop mode) executed successfully, 

967 False otherwise. 

968 

969 Raises: 

970 ValueError: If either of the source files cannot be found, if no tables could be found in the code, 

971 or if the codes before and after are identical. 

972 """ 

973 git_root = find_git_root(filename_before) or find_git_root(filename_after) 

974 

975 functions: set[str] = set() 

976 if function: # pragma: no cover 

977 functions.add(function) 

978 

979 if filename_before and ":" in filename_before: 

980 # e.g. models.py:define_tables 

981 filename_before, _function = filename_before.split(":", 1) 

982 functions.add(_function) 

983 

984 if filename_after and ":" in filename_after: 

985 # e.g. models.py:define_tables 

986 filename_after, _function = filename_after.split(":", 1) 

987 functions.add(_function) 

988 

989 before, after = extract_file_versions_and_paths(filename_before, filename_after) 

990 

991 version_before, filename_before = before 

992 version_after, filename_after = after 

993 

994 # either ./file exists or /file exists (seen from git root): 

995 

996 before_exists, before_absolute_path = get_absolute_path_info(filename_before, version_before, git_root) 

997 after_exists, after_absolute_path = get_absolute_path_info(filename_after, version_after, git_root) 

998 

999 if not (before_exists and after_exists): 

1000 message = "" 

1001 message += "" if before_exists else f"Path {filename_before} does not exist! " 

1002 if filename_before != filename_after: 

1003 message += "" if after_exists else f"Path {filename_after} does not exist!" 

1004 raise FileNotFoundError(message) 

1005 

1006 try: 

1007 code_before = get_file_for_version( 

1008 before_absolute_path, 

1009 version_before, 

1010 prompt_description="current table definition", 

1011 with_git=git_root is not None, 

1012 ) 

1013 code_after = get_file_for_version( 

1014 after_absolute_path, 

1015 version_after, 

1016 prompt_description="desired table definition", 

1017 with_git=git_root is not None, 

1018 ) 

1019 

1020 if not (code_before and code_after): 

1021 message = "" 

1022 message += "" if code_before else "Before code is empty (Maybe try `pydal2sql create`)! " 

1023 message += "" if code_after else "After code is empty! " 

1024 raise ValueError(message) 

1025 

1026 if code_before == code_after: 

1027 raise ValueError("Both contain the same code!") 

1028 

1029 except ValueError as e: 

1030 rich.print(f"[yellow] alter failed ({e}), trying create! [/yellow]", file=sys.stderr) 

1031 try: 

1032 return core_create( 

1033 filename_after or filename_before, 

1034 tables, 

1035 db_type, 

1036 magic, 

1037 noop, 

1038 verbose, 

1039 tuple(functions), 

1040 output_format, 

1041 output_file, 

1042 ) 

1043 except Exception: # pragma: no cover 

1044 return False 

1045 

1046 return handle_cli( 

1047 code_before, 

1048 code_after, 

1049 db_type=db_type, 

1050 tables=tables, 

1051 verbose=verbose, 

1052 noop=noop, 

1053 magic=magic, 

1054 function_name=tuple(functions), 

1055 output_format=output_format, 

1056 output_file=output_file, 

1057 )