load_scores

load_scores(path, layout='auto', registry=None)

Read a benchmark score CSV and validate its metric ids against the registry.

Parameters

Name Type Description Default
path str | Path Path to the CSV file. required
layout str "auto" (default) detects the layout from the header: a header whose columns are exactly tool, dataset, metric and score (in any order) is read as long, anything else as wide. Pass "wide" or "long" to force a layout. 'auto'
registry Registry | None Optional Registry instance. Defaults to a fresh registry over the bundled metrics. None

Returns

Type Description
Scores

Raises

Type Description
UnknownMetricError If any metric id in the file does not resolve to a metric card.
ValueError If the file is empty, a forced layout does not match the header, a long file is missing a required column, or a long file carries a duplicate tool-dataset-metric row.