debug.py

import pandas as pd
from src.display.utils import BENCHMARK_COLS
from src.about import Tasks
from src.leaderboard.read_evals import get_raw_eval_results

print("Tasks definitions:")
for task in Tasks:
    print(f"- {task.name}: benchmark={task.value.benchmark}, metric={task.value.metric}, col_name={task.value.col_name}")

print("\nBenchmark columns:", BENCHMARK_COLS)

try:
    # Get raw results first
    raw_results = get_raw_eval_results("eval-results", "eval-queue")
    print("\nRaw results:")
    for result in raw_results:
        print("\nResult:")
        print("- eval_name:", result.eval_name)
        print("- results:", result.results)
        data_dict = result.to_dict()
        print("- data_dict:", data_dict)
    
    # Convert to DataFrame
    all_data_json = [v.to_dict() for v in raw_results]
    df = pd.DataFrame.from_records(all_data_json)
    print("\nDataFrame columns:", df.columns.tolist())
    print("\nDataFrame contents:")
    print(df)
except Exception as e:
    print("\nError:", str(e))
    import traceback
    traceback.print_exc()
    
    # Print raw data for debugging
    print("\nRaw data from results file:")
    import json
    with open("eval-results/results_1.json") as f:
        print(json.dumps(json.load(f), indent=2))