Function melt_scores

visualize_instability.py:45–58

(runs: List[Dict])

def melt_scores(runs: List[Dict]) -> pd.DataFrame:
    rows = []
    for run in runs:
        for assignment, score in run["scores"].items():
            rows.append(
                {
                    "agent": run["agent"],
                    "file": run["file"],
                    "assignment": assignment,
                    "averageScore": score,
                    "finalScore": run["finalScore"],
                }
            )
    return pd.DataFrame(rows)


def plot_instability(df: pd.DataFrame, output: Path) -> None: