Jump to content

Connect Leeroopedia MCP: Equip your AI agents to search best practices, build plans, verify code, diagnose failures, and look up hyperparameter defaults.

Implementation:Truera Trulens Session Get Leaderboard

From Leeroopedia
Knowledge Sources
Domains LLM_Evaluation, Data_Analysis
Last Updated 2026-02-14 08:00 GMT

Overview

Concrete tool for generating an app version comparison leaderboard from aggregated feedback scores, provided by the trulens-core library.

Description

The TruSession.get_leaderboard method queries the database for all recorded feedback results and aggregates them by application version. The result is a pandas DataFrame with one row per app version and columns for each feedback function's mean score.

Usage

Call on a TruSession instance after recording multiple app versions. Filter by app IDs or group by metadata keys for more targeted analysis.

Code Reference

Source Location

  • Repository: trulens
  • File: src/core/trulens/core/session.py
  • Lines: L846-874

Signature

class TruSession:
    def get_leaderboard(
        self,
        app_ids: Optional[List[AppID]] = None,
        group_by_metadata_key: Optional[str] = None,
        limit: Optional[int] = None,
        offset: Optional[int] = None,
    ) -> pd.DataFrame:
        """Get a leaderboard for the given apps.

        Args:
            app_ids: Filter by specific app IDs. If None, includes all apps.
            group_by_metadata_key: Group results by a metadata key.
            limit: Max number of records to aggregate.
            offset: Record row offset.

        Returns:
            DataFrame of apps with their feedback results aggregated.
        """

Import

from trulens.core.session import TruSession

session = TruSession()
leaderboard = session.get_leaderboard()

I/O Contract

Inputs

Name Type Required Description
app_ids List[AppID] No Filter by specific apps
group_by_metadata_key str No Group by a metadata key
limit int No Max records to aggregate
offset int No Record row offset

Outputs

Name Type Description
return pd.DataFrame Apps as rows, feedback scores as columns (aggregated means)

Usage Examples

Basic Leaderboard

from trulens.core.session import TruSession

session = TruSession()
leaderboard = session.get_leaderboard()
print(leaderboard)

Filtered Leaderboard

leaderboard = session.get_leaderboard(
    app_ids=["app_1", "app_2"],
    limit=100
)

Related Pages

Implements Principle

Requires Environment

Page Connections

Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment