File size: 2,399 Bytes
db1341d aa37927 db1341d aa37927 bdb77ab 561f24a aa37927 233c78c db1341d bdb77ab db1341d 561f24a db1341d aa37927 39cb364 db1341d bdb77ab db1341d 561f24a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 |
import os
import pandas as pd
import json
from src.display.utils import COLUMNS, EVAL_COLS, Tasks
from src.envs import EVAL_RESULTS_PATH
def get_leaderboard_df(eval_results_path, eval_requests_path, cols, benchmark_cols):
# Initialize an empty DataFrame
df = pd.DataFrame(columns=cols)
# Load evaluation results from JSON files
if os.path.exists(eval_results_path):
result_files = [
os.path.join(eval_results_path, f)
for f in os.listdir(eval_results_path)
if f.endswith('.json')
]
data_list = []
for file in result_files:
with open(file, 'r') as f:
data = json.load(f)
flattened_data = {}
flattened_data.update(data.get('config', {}))
flattened_data.update(data.get('results', {}))
data_list.append(flattened_data)
if data_list:
df = pd.DataFrame(data_list)
# Ensure DataFrame has all columns
for col in cols:
if col not in df.columns:
df[col] = None
# Convert 'average' column to float and handle errors
if 'average' in df.columns:
df['average'] = pd.to_numeric(df['average'], errors='coerce')
# Sort by 'average' column if it exists
if 'average' in df.columns:
df = df.sort_values(by=['average'], ascending=False)
return df
def get_evaluation_queue_df(eval_requests_path, eval_cols):
# Initialize empty DataFrames
finished_df = pd.DataFrame(columns=eval_cols)
running_df = pd.DataFrame(columns=eval_cols)
pending_df = pd.DataFrame(columns=eval_cols)
# Load evaluation requests from JSON files
if os.path.exists(eval_requests_path):
request_files = [
os.path.join(eval_requests_path, f)
for f in os.listdir(eval_requests_path)
if f.endswith('.json')
]
data_list = []
for file in request_files:
with open(file, 'r') as f:
data = json.load(f)
data_list.append(data)
if data_list:
df = pd.DataFrame(data_list)
# Split DataFrame based on status
finished_df = df[df['status'] == 'finished']
running_df = df[df['status'] == 'running']
pending_df = df[df['status'] == 'pending']
return finished_df, running_df, pending_df
|