Module

pixano.cli.data

Pixano data import/export CLI (spec §9) — a thin shell over the shared io core.

Functions

def

 export_command 
(data_dir, dataset, destination, format, media, include_record_field)
  → None

Export a dataset from the library (JSONL v2 output re-imports identically).

Parameter	Type	Default
`data_dir`	`Path`	`typer.Argument(..., exists=True, file_okay=False, help='Pixano data directory.')`
`dataset`	`str`	`typer.Argument(..., help='Dataset name or id in the library.')`
`destination`	`Path`	`typer.Argument(..., help='Destination directory.')`
`format`	`str`	`typer.Option('pixano_jsonl', '--format', help='Export format.')`
`media`	`str`	`typer.Option('files', '--media', help='files (dump embedded bytes) or uris (write URIs verbatim).')`
`include_record_field`	`list[str]`	`typer.Option([], '--include-record-field', help='Include a Record base field in exported JSONL attrs (repeatable). Choices: status, created_at, updated_at, comment.')`

def

 fix_creation_dates 
(data_dir)
  → None

Fill empty ``creation_date`` fields of all datasets by using their oldest record timestamp.

Parameter	Type	Default
`data_dir`	`Path`	`typer.Argument(..., exists=True, file_okay=False, help='Path to data directory.')`

def

 formats_command 
()
  → None

List the registered data formats.

def

 import_command 
(data_dir, source, format, spec_file, name, workspace, mode, media, namespace, episodes, max_frames, dry_run, yes, importer_spec, info_py)
  → None

Import a dataset: analyze, confirm the plan, then ingest atomically.

Parameter	Type	Default
`data_dir`	`Path`	`typer.Argument(..., exists=True, file_okay=False, help='Pixano data directory.')`
`source`	`str`	`typer.Argument(..., help='Source directory, or a Hugging Face dataset id (org/name).')`
`format`	`str`	`typer.Option('auto', '--format', help='Data format (auto = detect).')`
`spec_file`	`Optional[Path]`	`typer.Option(None, '--spec', exists=True, dir_okay=False, help='dataset.yaml import spec (default: <source>/dataset.yaml).')`
`name`	`str`	`typer.Option('', '--name', help='Target dataset name (overrides the spec).')`
`workspace`	`str`	`typer.Option('', '--workspace', help='Workspace preset (image, video, image_vqa, ...).')`
`mode`	`str`	`typer.Option('create', '--mode', help='create, overwrite, or add.')`
`media`	`Optional[str]`	`typer.Option(None, '--media', help='Media storage: embed (default) or uri.')`
`namespace`	`Optional[str]`	`typer.Option(None, '--namespace', help='Id namespace (default: source identity).')`
`episodes`	`Optional[str]`	`typer.Option(None, '--episodes', help="LeRobot: episode subset, e.g. '0:4' or '1,3'.")`
`max_frames`	`Optional[int]`	`typer.Option(None, '--max-frames', help='LeRobot: cap extracted frames per episode (uniform stride).')`
`dry_run`	`bool`	`typer.Option(False, '--dry-run', help='Analyze and print the plan without importing.')`
`yes`	`bool`	`typer.Option(False, '--yes', '-y', help='Skip the plan confirmation prompt.')`
`importer_spec`	`Optional[str]`	`typer.Option(None, '--importer', help="Advanced: custom importer as 'path/to/file.py:ClassName'.")`
`info_py`	`Optional[str]`	`typer.Option(None, '--info-py', help="Advanced: DatasetInfo as 'path/to/file.py:attribute'.")`

def

 jobs_command 
(data_dir, action, job_id)
  → None

Inspect the shared import/export job store (the same one the GUI polls).

Parameter	Type	Default
`data_dir`	`Path`	`typer.Argument(..., exists=True, file_okay=False, help='Pixano data directory.')`
`action`	`str`	`typer.Argument('list', help='list, show, cancel, resume, or rollback.')`
`job_id`	`str`	`typer.Argument('', help='Job id (for show/cancel).')`

def

 optimize_dataset 
(data_dir, dataset)
  → None

Index the standard filter columns and compact a dataset (fast filtered pagination).

Parameter	Type	Default
`data_dir`	`Path`	`typer.Argument(..., help='Pixano data directory.')`
`dataset`	`str`	`typer.Argument(..., help='Dataset name (library folder name).')`