plexus.cli.dataset.datasets module

plexus.cli.dataset.datasets.apply_item_pipeline_to_dataframe(dataframe, item_config, client)

plexus.cli.dataset.datasets.build_associated_dataset_from_feedback_ids(*, client: PlexusDashboardClient, scorecard_identifier: str, score_identifier: str, feedback_item_ids: Sequence[str], source_report_block_id: str | None = None, eligibility_rule: str = 'unanimous non-contradiction', task_id: str | None = None) → Dict[str, Any]

Build an associated dataset directly from explicit feedback item IDs.

One strict label source is supported: FeedbackItem.finalAnswerValue.

plexus.cli.dataset.datasets.create_client() → PlexusDashboardClient

Create a client and log its configuration

async plexus.cli.dataset.datasets.create_initial_data_source_version(client, data_source)

Create the initial version for a DataSource that doesn’t have one yet.

plexus.cli.dataset.datasets.get_amplify_bucket()

Get the S3 bucket name from environment variables or fall back to reading amplify_outputs.json.