divbase-cli 0.1.0.dev1__tar.gz → 0.1.0.dev3__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (31) hide show
  1. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/.gitignore +4 -0
  2. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/PKG-INFO +4 -3
  3. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/pyproject.toml +10 -3
  4. divbase_cli-0.1.0.dev3/src/divbase_cli/__init__.py +1 -0
  5. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_commands/auth_cli.py +4 -9
  6. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_commands/dimensions_cli.py +4 -8
  7. divbase_cli-0.1.0.dev3/src/divbase_cli/cli_commands/file_cli.py +459 -0
  8. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_commands/query_cli.py +3 -7
  9. divbase_cli-0.1.0.dev3/src/divbase_cli/cli_commands/shared_args_options.py +20 -0
  10. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_commands/task_history_cli.py +3 -8
  11. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_commands/user_config_cli.py +14 -44
  12. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_commands/version_cli.py +16 -24
  13. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_config.py +18 -7
  14. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_exceptions.py +37 -22
  15. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/config_resolver.py +10 -10
  16. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/divbase_cli.py +1 -1
  17. divbase_cli-0.1.0.dev3/src/divbase_cli/retries.py +34 -0
  18. divbase_cli-0.1.0.dev3/src/divbase_cli/services/__init__.py +0 -0
  19. divbase_cli-0.1.0.dev3/src/divbase_cli/services/pre_signed_urls.py +446 -0
  20. divbase_cli-0.1.0.dev3/src/divbase_cli/services/project_versions.py +77 -0
  21. divbase_cli-0.1.0.dev3/src/divbase_cli/services/s3_files.py +355 -0
  22. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/user_auth.py +26 -13
  23. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/user_config.py +20 -9
  24. divbase_cli-0.1.0.dev3/src/divbase_cli/utils.py +47 -0
  25. divbase_cli-0.1.0.dev1/src/divbase_cli/__init__.py +0 -1
  26. divbase_cli-0.1.0.dev1/src/divbase_cli/cli_commands/file_cli.py +0 -245
  27. divbase_cli-0.1.0.dev1/src/divbase_cli/pre_signed_urls.py +0 -169
  28. divbase_cli-0.1.0.dev1/src/divbase_cli/services.py +0 -219
  29. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/README.md +0 -0
  30. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/cli_commands/__init__.py +0 -0
  31. {divbase_cli-0.1.0.dev1 → divbase_cli-0.1.0.dev3}/src/divbase_cli/display_task_history.py +0 -0
@@ -17,6 +17,7 @@ sample_metadata_*.tsv
17
17
  *.vcf
18
18
  *.vcf.gz
19
19
  *.vcf.gz.csi
20
+ *.vcf.gz.tbi
20
21
  !tests/fixtures/*.vcf.gz
21
22
  tests/fixtures/temp*
22
23
  tests/fixtures/merged*
@@ -28,6 +29,9 @@ bcftools_divbase_job_config.json
28
29
  vcf_dimensions.tsv
29
30
  mock*.tsv
30
31
  task_records*.json
32
+ split_scaffold_files.txt
33
+ scripts/benchmarking/*.yaml
34
+ scripts/benchmarking/results
31
35
 
32
36
  #MacOS artifacts
33
37
  .DS_Store
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: divbase-cli
3
- Version: 0.1.0.dev1
3
+ Version: 0.1.0.dev3
4
4
  Summary: Command Line Interface for Divbase
5
5
  Project-URL: Homepage, https://divbase.scilifelab.se
6
6
  Project-URL: Documentation, https://scilifelabdatacentre.github.io/divbase
@@ -18,8 +18,9 @@ Classifier: Programming Language :: Python :: 3.12
18
18
  Classifier: Programming Language :: Python :: 3.13
19
19
  Classifier: Programming Language :: Python :: 3.14
20
20
  Requires-Python: >=3.12
21
- Requires-Dist: divbase-lib==0.1.0.dev1
22
- Requires-Dist: httpx<2,>=0.28.1
21
+ Requires-Dist: divbase-lib==0.1.0.dev3
22
+ Requires-Dist: httpx<1,>=0.28.1
23
+ Requires-Dist: stamina>=25.2.0
23
24
  Requires-Dist: typer<1,>=0.21.1
24
25
  Description-Content-Type: text/markdown
25
26
 
@@ -8,9 +8,10 @@ authors = [
8
8
  ]
9
9
  requires-python = ">=3.12"
10
10
  dependencies = [
11
- "divbase-lib==0.1.0.dev1",
12
11
  "typer>=0.21.1,<1",
13
- "httpx>=0.28.1,<2",
12
+ "httpx>=0.28.1,<1",
13
+ "stamina>=25.2.0",
14
+ "divbase-lib==0.1.0.dev3",
14
15
  ]
15
16
  dynamic = ["version"]
16
17
 
@@ -44,4 +45,10 @@ build-backend = "hatchling.build"
44
45
  path = "src/divbase_cli/__init__.py"
45
46
 
46
47
  [tool.uv.sources]
47
- divbase-lib = { workspace = true }
48
+ divbase-lib = { workspace = true }
49
+
50
+ [tool.ruff]
51
+ extend = "../../ruff.toml"
52
+
53
+ [tool.ruff.lint.flake8-tidy-imports.banned-api]
54
+ "divbase_api" = { msg = "divbase-cli cannot use code from divbase-api. Move shared code to divbase-lib." }
@@ -0,0 +1 @@
1
+ __version__ = "0.1.0.dev3"
@@ -3,13 +3,11 @@ CLI subcommand for managing user auth with DivBase server.
3
3
  """
4
4
 
5
5
  from datetime import datetime
6
- from pathlib import Path
7
6
 
8
7
  import typer
9
8
  from pydantic import SecretStr
10
9
  from typing_extensions import Annotated
11
10
 
12
- from divbase_cli.cli_commands.user_config_cli import CONFIG_FILE_OPTION
13
11
  from divbase_cli.cli_config import cli_settings
14
12
  from divbase_cli.cli_exceptions import AuthenticationError
15
13
  from divbase_cli.user_auth import (
@@ -30,7 +28,6 @@ def login(
30
28
  email: str,
31
29
  password: Annotated[str, typer.Option(prompt=True, hide_input=True)],
32
30
  divbase_url: str = typer.Option(cli_settings.DIVBASE_API_URL, help="DivBase server URL to connect to."),
33
- config_file: Path = CONFIG_FILE_OPTION,
34
31
  force: bool = typer.Option(False, "--force", "-f", help="Force login again even if already logged in"),
35
32
  ):
36
33
  """
@@ -43,7 +40,7 @@ def login(
43
40
  secret_password = SecretStr(password)
44
41
  del password # avoid user passwords showing up in error messages etc...
45
42
 
46
- config = load_user_config(config_file)
43
+ config = load_user_config()
47
44
 
48
45
  if not force:
49
46
  session_expires_at = check_existing_session(divbase_url=divbase_url, config=config)
@@ -55,7 +52,7 @@ def login(
55
52
  print("Login cancelled.")
56
53
  return
57
54
 
58
- login_to_divbase(email=email, password=secret_password, divbase_url=divbase_url, config_path=config_file)
55
+ login_to_divbase(email=email, password=secret_password, divbase_url=divbase_url)
59
56
  print(f"Logged in successfully as: {email}")
60
57
 
61
58
 
@@ -69,13 +66,11 @@ def logout():
69
66
 
70
67
 
71
68
  @auth_app.command("whoami")
72
- def whoami(
73
- config_file: Path = CONFIG_FILE_OPTION,
74
- ):
69
+ def whoami():
75
70
  """
76
71
  Return information about the currently logged-in user.
77
72
  """
78
- config = load_user_config(config_file)
73
+ config = load_user_config()
79
74
  logged_in_url = config.logged_in_url
80
75
 
81
76
  # TODO - move logged in check to the make_authenticated_request function?
@@ -1,11 +1,9 @@
1
1
  import logging
2
- from pathlib import Path
3
2
 
4
3
  import typer
5
4
  import yaml
6
5
 
7
- from divbase_cli.cli_commands.user_config_cli import CONFIG_FILE_OPTION
8
- from divbase_cli.cli_commands.version_cli import PROJECT_NAME_OPTION
6
+ from divbase_cli.cli_commands.shared_args_options import PROJECT_NAME_OPTION
9
7
  from divbase_cli.config_resolver import resolve_project
10
8
  from divbase_cli.user_auth import make_authenticated_request
11
9
  from divbase_lib.api_schemas.vcf_dimensions import DimensionsShowResult
@@ -22,11 +20,10 @@ dimensions_app = typer.Typer(
22
20
  @dimensions_app.command("update")
23
21
  def update_dimensions_index(
24
22
  project: str | None = PROJECT_NAME_OPTION,
25
- config_file: Path = CONFIG_FILE_OPTION,
26
23
  ) -> None:
27
24
  """Calculate and add the dimensions of a VCF file to the dimensions index file in the project."""
28
25
 
29
- project_config = resolve_project(project_name=project, config_path=config_file)
26
+ project_config = resolve_project(project_name=project)
30
27
 
31
28
  response = make_authenticated_request(
32
29
  method="PUT",
@@ -53,14 +50,13 @@ def show_dimensions_index(
53
50
  )
54
51
  ),
55
52
  project: str | None = PROJECT_NAME_OPTION,
56
- config_file: Path = CONFIG_FILE_OPTION,
57
53
  ) -> None:
58
54
  """
59
55
  Show the dimensions index file for a project.
60
56
  When running --unique-scaffolds, the sorting separates between numeric and non-numeric scaffold names.
61
57
  """
62
58
 
63
- project_config = resolve_project(project_name=project, config_path=config_file)
59
+ project_config = resolve_project(project_name=project)
64
60
 
65
61
  response = make_authenticated_request(
66
62
  method="GET",
@@ -82,7 +78,7 @@ def show_dimensions_index(
82
78
  else:
83
79
  print(
84
80
  f"No entry found for filename: {filename}. Please check that the filename is correct and that it is a VCF file (extension: .vcf or .vcf.gz)."
85
- "\nHint: use 'divbase-cli files list' to view all files in the project."
81
+ "\nHint: use 'divbase-cli files ls' to view all files in the project."
86
82
  )
87
83
  return
88
84
 
@@ -0,0 +1,459 @@
1
+ """
2
+ Command line interface for managing files in a DivBase project's store on DivBase.
3
+
4
+ TODO - Download all files option.
5
+ TODO - skip checked option (aka skip files that already exist in same local dir with correct checksum).
6
+ """
7
+
8
+ from pathlib import Path
9
+ from zoneinfo import ZoneInfo
10
+
11
+ import typer
12
+ from rich import print
13
+ from rich.table import Table
14
+ from typing_extensions import Annotated
15
+
16
+ from divbase_cli.cli_commands.shared_args_options import FORMAT_AS_TSV_OPTION, PROJECT_NAME_OPTION
17
+ from divbase_cli.cli_exceptions import UnsupportedFileNameError, UnsupportedFileTypeError
18
+ from divbase_cli.config_resolver import ensure_logged_in, resolve_download_dir, resolve_project
19
+ from divbase_cli.services.s3_files import (
20
+ download_files_command,
21
+ get_file_info_command,
22
+ list_files_command,
23
+ restore_objects_command,
24
+ soft_delete_objects_command,
25
+ stream_file_command,
26
+ upload_files_command,
27
+ )
28
+ from divbase_cli.utils import format_file_size, print_rich_table_as_tsv
29
+ from divbase_lib.divbase_constants import SUPPORTED_DIVBASE_FILE_TYPES, UNSUPPORTED_CHARACTERS_IN_FILENAMES
30
+
31
+ file_app = typer.Typer(no_args_is_help=True, help="Download/upload/list files to/from the project's store on DivBase.")
32
+
33
+ NO_FILES_SPECIFIED_MSG = "No files specified for the command, exiting..."
34
+
35
+
36
+ @file_app.command("ls")
37
+ def list_files(
38
+ format_output_as_tsv: bool = FORMAT_AS_TSV_OPTION,
39
+ prefix_filter: str | None = typer.Option(
40
+ None,
41
+ "--prefix",
42
+ "-p",
43
+ help="Optional prefix to filter the listed files by name (only list files starting with this prefix).",
44
+ ),
45
+ include_results_files: bool = typer.Option(
46
+ False,
47
+ "--include-results-files",
48
+ "-r",
49
+ help="If set, will also show DivBase query results files which are hidden by default.",
50
+ ),
51
+ project: str | None = PROJECT_NAME_OPTION,
52
+ ):
53
+ """
54
+ list all currently available files in the project's DivBase store.
55
+
56
+ You can optionally filter the listed files by providing a prefix.
57
+ By default, DivBase query results files are hidden from the listing. Use the --include-results-files option to include them.
58
+ To see information about the versions of each file, use the 'divbase-cli files info [FILE_NAME]' command instead
59
+ """
60
+ project_config = resolve_project(project_name=project)
61
+ logged_in_url = ensure_logged_in(desired_url=project_config.divbase_url)
62
+
63
+ files = list_files_command(
64
+ divbase_base_url=logged_in_url,
65
+ project_name=project_config.name,
66
+ prefix_filter=prefix_filter,
67
+ include_results_files=include_results_files,
68
+ )
69
+
70
+ if not files:
71
+ print("No files found in the project's store on DivBase.")
72
+ return
73
+
74
+ table = Table(title=f"Files in [bold]{project_config.name}'s [/bold] DivBase Store:")
75
+ table.add_column("Name", justify="left", style="cyan")
76
+ table.add_column("File size", justify="left", style="magenta", no_wrap=True)
77
+ table.add_column("Upload date (CET)", justify="left", style="green", no_wrap=True)
78
+ table.add_column("MD5 checksum", justify="left", style="yellow")
79
+
80
+ for file_details in files:
81
+ cet_timestamp = file_details.last_modified.astimezone(ZoneInfo("CET")).strftime("%Y-%m-%d %H:%M:%S %Z")
82
+ file_size = format_file_size(size_bytes=file_details.size)
83
+ table.add_row(
84
+ file_details.name,
85
+ file_size,
86
+ cet_timestamp,
87
+ file_details.etag,
88
+ )
89
+
90
+ if not format_output_as_tsv:
91
+ print(table)
92
+ else:
93
+ print_rich_table_as_tsv(table=table)
94
+
95
+
96
+ @file_app.command("info")
97
+ def file_info(
98
+ file_name: str = typer.Argument(..., help="Name of the file to get information about."),
99
+ format_output_as_tsv: bool = FORMAT_AS_TSV_OPTION,
100
+ project: str | None = PROJECT_NAME_OPTION,
101
+ ):
102
+ """
103
+ Get detailed information about a specific file in the project's DivBase store.
104
+
105
+ This includes all versions of the file and whether the file is currently marked as soft deleted.
106
+ """
107
+ project_config = resolve_project(project_name=project)
108
+ logged_in_url = ensure_logged_in(desired_url=project_config.divbase_url)
109
+
110
+ file_info = get_file_info_command(
111
+ divbase_base_url=logged_in_url,
112
+ project_name=project_config.name,
113
+ object_name=file_name,
114
+ )
115
+
116
+ if not file_info.versions:
117
+ # API should never return an object with no versions, but just in case
118
+ print("No available versions for this file.")
119
+ return
120
+
121
+ if file_info.is_currently_deleted:
122
+ print(
123
+ "[bold red]Warning: This file is marked as soft deleted.\n[/bold red]"
124
+ + "[italic] To restore a deleted file, use the 'divbase-cli files restore' command.\n"
125
+ + " Or upload the file again using the 'divbase-cli files upload' command.\n[/italic]",
126
+ )
127
+
128
+ table = Table(
129
+ title=f"Available Versions for '[bold]{file_info.object_name}[/bold]'",
130
+ caption="Versions shown are ordered with the latest/current version first/at the top",
131
+ )
132
+ table.add_column("File size", justify="left", style="magenta", no_wrap=True)
133
+ table.add_column("Upload date (CET)", justify="left", style="green", no_wrap=True)
134
+ table.add_column("MD5 checksum", justify="left", style="yellow")
135
+ table.add_column("Version ID", justify="left", style="cyan")
136
+
137
+ for version in file_info.versions:
138
+ cet_timestamp = version.last_modified.astimezone(ZoneInfo("CET")).strftime("%Y-%m-%d %H:%M:%S %Z")
139
+ file_size = format_file_size(size_bytes=version.size)
140
+ table.add_row(
141
+ file_size,
142
+ cet_timestamp,
143
+ version.etag,
144
+ version.version_id,
145
+ )
146
+
147
+ if not format_output_as_tsv:
148
+ print(table)
149
+ else:
150
+ print_rich_table_as_tsv(table=table)
151
+
152
+
153
+ @file_app.command("download")
154
+ def download_files(
155
+ files: list[str] = typer.Argument(
156
+ None, help="Space separated list of files/objects to download from the project's store on DivBase."
157
+ ),
158
+ file_list: Path | None = typer.Option(None, "--file-list", help="Text file with list of files to upload."),
159
+ download_dir: str = typer.Option(
160
+ None,
161
+ help="""Directory to download the files to.
162
+ If not provided, defaults to what you specified in your user config.
163
+ If also not specified in your user config, downloads to the current directory.
164
+ You can also specify "." to download to the current directory.""",
165
+ ),
166
+ disable_verify_checksums: Annotated[
167
+ bool,
168
+ typer.Option(
169
+ "--disable-verify-checksums",
170
+ help="Turn off checksum verification which is on by default. "
171
+ "Checksum verification means all downloaded files are verified against their MD5 checksums."
172
+ "It is recommended to leave checksum verification enabled unless you have a specific reason to disable it.",
173
+ ),
174
+ ] = False,
175
+ project_version: str | None = typer.Option(
176
+ default=None,
177
+ help="User defined version of the project's at which to download the files. If not provided, downloads the latest version of all selected files.",
178
+ ),
179
+ project: str | None = PROJECT_NAME_OPTION,
180
+ ):
181
+ """
182
+ Download files from the project's store on DivBase.
183
+
184
+ This can be done by either:
185
+ 1. providing a list of files paths directly in the command line
186
+ 2. providing a text file with a list of files to download (new file on each line).
187
+
188
+ To download the latest version of a file, just provide its name. "file1" "file2" etc.
189
+ To download a specific/older version of a file, use the format: "file_name:version_id"
190
+ You can get a file's version id using the 'divbase-cli file info [FILE_NAME]' command.
191
+ You can mix and match latest and specific versions in the same command.
192
+ E.g. to download the latest version of file1 and version "3xcdsdsdiw829x"
193
+ of file2: 'divbase-cli files download file1 file2:3xcdsdsdiw829x'
194
+ """
195
+ project_config = resolve_project(project_name=project)
196
+ logged_in_url = ensure_logged_in(desired_url=project_config.divbase_url)
197
+ download_dir_path = resolve_download_dir(download_dir=download_dir)
198
+
199
+ raw_files_input = _resolve_file_inputs(files=files, file_list=file_list)
200
+
201
+ download_results = download_files_command(
202
+ divbase_base_url=logged_in_url,
203
+ project_name=project_config.name,
204
+ raw_files_input=raw_files_input,
205
+ download_dir=download_dir_path,
206
+ verify_checksums=not disable_verify_checksums,
207
+ project_version=project_version,
208
+ )
209
+
210
+ if download_results.successful:
211
+ print("[green bold]Successfully downloaded the following files:[/green bold]")
212
+ for success in download_results.successful:
213
+ print(f"- '{success.object_name}' downloaded to: '{success.file_path.resolve()}'")
214
+ if download_results.failed:
215
+ print("[red bold]ERROR: Failed to download the following files:[/red bold]")
216
+ for failed in download_results.failed:
217
+ print(f"[red]- '{failed.object_name}': Exception: '{failed.exception}'[/red]")
218
+
219
+ raise typer.Exit(1)
220
+
221
+
222
+ @file_app.command("stream")
223
+ def stream_file(
224
+ file_name: str = typer.Argument(..., help="Name of the file you want to stream."),
225
+ version_id: str | None = typer.Option(
226
+ default=None,
227
+ help="Specify this if you want to look at an older/specific version of the file. "
228
+ "If not provided, the latest version of the file is used. "
229
+ "To get a file's version ids, use the 'divbase-cli file info [FILE_NAME]' command.",
230
+ ),
231
+ project: str | None = PROJECT_NAME_OPTION,
232
+ ):
233
+ """
234
+ Stream a file's content to standard output.
235
+
236
+ This allows your to pipe the output to other tools like 'less', 'head', 'zcat' and 'bcftools'.
237
+
238
+ Examples:
239
+ - View a file: divbase-cli files stream my_file.tsv | less
240
+ - View a gzipped file: divbase-cli files stream my_file.vcf.gz | zcat | less
241
+ - Run a bcftools command: divbase-cli files stream my_file.vcf.gz | bcftools view -h - # The "-" tells bcftools to read from standard input
242
+ """
243
+ project_config = resolve_project(project_name=project)
244
+ logged_in_url = ensure_logged_in(desired_url=project_config.divbase_url)
245
+
246
+ stream_file_command(
247
+ divbase_base_url=logged_in_url,
248
+ project_name=project_config.name,
249
+ file_name=file_name,
250
+ version_id=version_id,
251
+ )
252
+
253
+
254
+ @file_app.command("upload")
255
+ def upload_files(
256
+ files: list[Path] | None = typer.Argument(None, help="Space separated list of files to upload."),
257
+ upload_dir: Path | None = typer.Option(None, "--upload-dir", help="Directory to upload all files from."),
258
+ file_list: Path | None = typer.Option(None, "--file-list", help="Text file with list of files to upload."),
259
+ disable_safe_mode: Annotated[
260
+ bool,
261
+ typer.Option(
262
+ "--disable-safe-mode",
263
+ help="Turn off safe mode which is on by default. Safe mode adds 2 extra bits of security by first calculating the MD5 checksum of each file that you're about to upload:"
264
+ "(1) Checks if any of the files you're about to upload already exist (by comparing name and checksum) and if so stops the upload process."
265
+ "(2) Sends the file's checksum when the file is uploaded so the server can verify the upload was successful (by calculating and comparing the checksums)."
266
+ "It is recommended to leave safe mode enabled unless you have a specific reason to disable it.",
267
+ ),
268
+ ] = False,
269
+ project: str | None = PROJECT_NAME_OPTION,
270
+ ):
271
+ """
272
+ Upload files to your project's store on DivBase:
273
+
274
+ To provide files to upload you can either:
275
+ 1. provide a list of files paths directly in the command line
276
+ 2. provide a directory to upload
277
+ 3. provide a text file with or a file list.
278
+ """
279
+ project_config = resolve_project(project_name=project)
280
+ logged_in_url = ensure_logged_in(desired_url=project_config.divbase_url)
281
+
282
+ if bool(files) + bool(upload_dir) + bool(file_list) > 1:
283
+ print("Please specify only one of --files, --upload_dir, or --file-list.")
284
+ raise typer.Exit(1)
285
+
286
+ all_files: set[Path] = set()
287
+ if files:
288
+ all_files.update(files)
289
+ if upload_dir:
290
+ all_files.update([p for p in upload_dir.iterdir() if p.is_file()])
291
+ if file_list:
292
+ with open(file_list) as f:
293
+ for line in f:
294
+ path = Path(line.strip())
295
+ if path.is_file():
296
+ all_files.add(path)
297
+
298
+ if not all_files:
299
+ print(NO_FILES_SPECIFIED_MSG)
300
+ raise typer.Exit(1)
301
+
302
+ _check_for_unsupported_files(all_files)
303
+
304
+ uploaded_results = upload_files_command(
305
+ project_name=project_config.name,
306
+ divbase_base_url=logged_in_url,
307
+ all_files=list(all_files),
308
+ safe_mode=not disable_safe_mode,
309
+ )
310
+
311
+ if uploaded_results.successful:
312
+ print("[green bold] The following files were successfully uploaded: [/green bold]")
313
+ for object in uploaded_results.successful:
314
+ print(f"- '{object.object_name}' created from file at: '{object.file_path.resolve()}'")
315
+
316
+ if uploaded_results.failed:
317
+ print("[red bold]ERROR: Failed to upload the following files:[/red bold]")
318
+ for failed in uploaded_results.failed:
319
+ print(f"[red]- '{failed.object_name}': Exception: '{failed.exception}'[/red]")
320
+
321
+ raise typer.Exit(1)
322
+
323
+
324
+ @file_app.command("rm")
325
+ def remove_files(
326
+ files: list[str] | None = typer.Argument(
327
+ None, help="Space seperated list of files/objects in the project's store on DivBase to delete."
328
+ ),
329
+ file_list: Path | None = typer.Option(None, "--file-list", help="Text file with list of files to delete."),
330
+ dry_run: bool = typer.Option(
331
+ False, "--dry-run", help="If set, will not actually delete the files, just print what would be deleted."
332
+ ),
333
+ project: str | None = PROJECT_NAME_OPTION,
334
+ ):
335
+ """
336
+ Soft delete files from the project's store on DivBase
337
+
338
+ To provide files to delete you can either:
339
+ 1. provide a list of file names directly in the command line
340
+ 2. provide a text file with a list of files to delete.
341
+
342
+ Note that deleting a non existent file will be treated as a successful deletion.
343
+ """
344
+ project_config = resolve_project(project_name=project)
345
+ logged_in_url = ensure_logged_in(desired_url=project_config.divbase_url)
346
+
347
+ all_files = _resolve_file_inputs(files=files, file_list=file_list)
348
+
349
+ if dry_run:
350
+ print("Dry run mode enabled. The following files would have been deleted:")
351
+ for file in all_files:
352
+ print(f"- '{file}'")
353
+ return
354
+
355
+ deleted_files = soft_delete_objects_command(
356
+ divbase_base_url=logged_in_url,
357
+ project_name=project_config.name,
358
+ all_files=all_files,
359
+ )
360
+
361
+ if deleted_files:
362
+ print("Deleted files:")
363
+ for file in deleted_files:
364
+ print(f"- '{file}'")
365
+ else:
366
+ print("No files were deleted.")
367
+
368
+
369
+ @file_app.command("restore")
370
+ def restore_soft_deleted_files(
371
+ files: list[str] | None = typer.Argument(
372
+ None, help="Space seperated list of files/objects in the project's store on DivBase to restore."
373
+ ),
374
+ file_list: Path | None = typer.Option(None, "--file-list", help="Text file with list of files to restore."),
375
+ project: str | None = PROJECT_NAME_OPTION,
376
+ ):
377
+ """
378
+ Restore soft deleted files from the project's store on DivBase
379
+
380
+ To provide files to restore you can either:
381
+ 1. provide a list of files directly in the command line.
382
+ 2. provide a text file with a list of files to restore (new file on each line).
383
+
384
+ NOTE: Attempts to restore a file that is not soft deleted will be considered successful and the file will remain live. This means you can repeatedly run this command on the same file and get the same response.
385
+ """
386
+ project_config = resolve_project(project_name=project)
387
+ logged_in_url = ensure_logged_in(desired_url=project_config.divbase_url)
388
+
389
+ all_files = _resolve_file_inputs(files=files, file_list=file_list)
390
+
391
+ restored_objects_response = restore_objects_command(
392
+ divbase_base_url=logged_in_url,
393
+ project_name=project_config.name,
394
+ all_files=all_files,
395
+ )
396
+
397
+ if restored_objects_response.restored:
398
+ print("Restored files:")
399
+ for file in restored_objects_response.restored:
400
+ print(f"- '{file}'")
401
+
402
+ if restored_objects_response.not_restored:
403
+ print("[bold red]WARNING: Some files could not be restored:[/bold red]")
404
+ for file in restored_objects_response.not_restored:
405
+ print(f"[red]- '{file}'[/red]")
406
+
407
+ print(
408
+ "Possible reasons for failed restores:\n"
409
+ "1. The object does not exist in the bucket (e.g., a typo in the name).\n"
410
+ "2. The object was hard-deleted and is unrecoverable.\n"
411
+ "3. An unexpected server error occurred during the restore attempt."
412
+ )
413
+
414
+
415
+ def _resolve_file_inputs(files: list[str] | None, file_list: Path | None) -> list[str]:
416
+ """Helper function to resolve file inputs from command line arguments."""
417
+ if bool(files) + bool(file_list) > 1:
418
+ print("Please specify only one of --files or --file-list.")
419
+ raise typer.Exit(1)
420
+
421
+ all_files = set()
422
+ if files:
423
+ all_files.update(files)
424
+ if file_list:
425
+ with open(file_list) as f:
426
+ for line in f:
427
+ all_files.add(line.strip())
428
+
429
+ if not all_files:
430
+ print(NO_FILES_SPECIFIED_MSG)
431
+ raise typer.Exit(1)
432
+ return list(all_files)
433
+
434
+
435
+ def _check_for_unsupported_files(all_files: set[Path]) -> None:
436
+ """
437
+ Helper fn to check if any of the files to be uploaded are not supported by DivBase.
438
+ Raises error if so.
439
+
440
+ This can be to prevent users from:
441
+ 1. Accidentally uploading unsupported files.
442
+ 2. Uploading files that have characters that we reserve for actions like filtering/querying on DivBase.
443
+ (e.g. the syntax "[file_name]:[version_id]" can be used to download specific file versions).
444
+
445
+ This is not a security feature, just for UX purposes.
446
+ """
447
+ unsupported_file_types, unsupported_chars = [], []
448
+ for file_path in all_files:
449
+ if not any(file_path.name.endswith(supported) for supported in SUPPORTED_DIVBASE_FILE_TYPES):
450
+ unsupported_file_types.append(file_path)
451
+
452
+ if any(char in file_path.name for char in UNSUPPORTED_CHARACTERS_IN_FILENAMES):
453
+ unsupported_chars.append(file_path)
454
+
455
+ if unsupported_file_types:
456
+ raise UnsupportedFileTypeError(unsupported_files=unsupported_file_types)
457
+
458
+ if unsupported_chars:
459
+ raise UnsupportedFileNameError(unsupported_files=unsupported_chars)
@@ -17,13 +17,11 @@ TODO:
17
17
  """
18
18
 
19
19
  import logging
20
- from pathlib import Path
21
20
 
22
21
  import typer
23
22
  from rich import print
24
23
 
25
- from divbase_cli.cli_commands.user_config_cli import CONFIG_FILE_OPTION
26
- from divbase_cli.cli_commands.version_cli import PROJECT_NAME_OPTION
24
+ from divbase_cli.cli_commands.shared_args_options import PROJECT_NAME_OPTION
27
25
  from divbase_cli.cli_config import cli_settings
28
26
  from divbase_cli.config_resolver import resolve_project
29
27
  from divbase_cli.user_auth import make_authenticated_request
@@ -75,7 +73,6 @@ def sample_metadata_query(
75
73
  ),
76
74
  metadata_tsv_name: str = METADATA_TSV_ARGUMENT,
77
75
  project: str | None = PROJECT_NAME_OPTION,
78
- config_file: Path = CONFIG_FILE_OPTION,
79
76
  ) -> None:
80
77
  """
81
78
  Query the tsv sidecar metadata file for the VCF files in the project's data store on DivBase.
@@ -86,7 +83,7 @@ def sample_metadata_query(
86
83
  TODO: handle when the name of the sample column is something other than Sample_ID
87
84
  """
88
85
 
89
- project_config = resolve_project(project_name=project, config_path=config_file)
86
+ project_config = resolve_project(project_name=project)
90
87
 
91
88
  request_data = SampleMetadataQueryRequest(tsv_filter=filter, metadata_tsv_name=metadata_tsv_name)
92
89
 
@@ -116,7 +113,6 @@ def pipe_query(
116
113
  command: str = BCFTOOLS_ARGUMENT,
117
114
  metadata_tsv_name: str = METADATA_TSV_ARGUMENT,
118
115
  project: str | None = PROJECT_NAME_OPTION,
119
- config_file: Path = CONFIG_FILE_OPTION,
120
116
  ) -> None:
121
117
  """
122
118
  Submit a query to run on the DivBase API. A single, merged VCF file will be added to the project on success.
@@ -129,7 +125,7 @@ def pipe_query(
129
125
  TODO consider handling the bcftools command whitelist checks also on the CLI level since the error messages are nicer looking?
130
126
  TODO consider moving downloading of missing files elsewhere, since this is now done before the celery task
131
127
  """
132
- project_config = resolve_project(project_name=project, config_path=config_file)
128
+ project_config = resolve_project(project_name=project)
133
129
 
134
130
  request_data = BcftoolsQueryRequest(tsv_filter=tsv_filter, command=command, metadata_tsv_name=metadata_tsv_name)
135
131