Summary

class skore.project.summary.Summary(data=None, index=None, columns=None, dtype=None, copy=None)

Metadata and metrics for all reports persisted in a project at a given moment.

A summary object is an extended pandas.DataFrame, containing all the metadata and metrics of the reports that have been persisted in a project. It implements a custom HTML representation, that allows user to filter its reports using a parallel coordinates plot. In a Jupyter Notebook, this representation automatically replaces the standard pandas.DataFrame one and displays an interactive plot.

The parallel coordinates plot is interactive, such the user can select a filter path and retrieve the corresponding reports.

Outside a Jupyter Notebook, the summary object can be used as a standard pandas.DataFrame object. This means that it can be questioned, divided etc., using the standard pandas.DataFrame functions.

Refer to pandas.DataFrame for the standard methods and attributes.

static factory(project, /)

Construct a summary object from project at a given moment.

Parameters:

projectUnion[skore_local_project.Project, skore_hub_project.Project]

The project from which the summary object is to be constructed.

Notes

This function is not intended for direct use. Instead simply use the accessor skore.Project.summarize().

reports(*, filter=True, return_as='list')

Return the reports referenced by the summary object from the project.

Parameters:

filterbool, optional

Filter the reports to return with the user query string selection, default True.

return_asLiteral[“list”, “comparison”], optional

Return reports as flat list or comparison report, default list.