Upload 4 files

src/leaderboard_utils.py

@@ -95,9 +95,22 @@ def process_df(df: pd.DataFrame) -> pd.DataFrame:
         df[col] = pd.to_numeric(df[col], errors='coerce')
 
     # Remove repetition in model names
-    df[df.columns[0]] = df[df.columns[0]].str.replace('-t0.0', '', regex=True)
+    df[df.columns[0]] = df[df.columns[0]].str.replace(r'-t[0-1]\.\d+', '', regex=True)
     df[df.columns[0]] = df[df.columns[0]].apply(lambda x: '--'.join(set(x.split('--'))))
 
+    # Rename the first column to 'Model' if it starts with 'Unnamed'
+    if df.columns[0].startswith('Unnamed'):
+        df.rename(columns={df.columns[0]: 'Model'}, inplace=True)
+
+    # Define the desired column order
+    desired_columns = ['Model', '-, clemscore', 'all, Average % Played', 'all, Average Quality Score']
+
+    # Ensure the DataFrame has all the desired columns
+    existing_columns = [col for col in desired_columns if col in df.columns]
+
+    # Reorder other DataFrame columns
+    df = df[existing_columns + [col for col in df.columns if col not in existing_columns]]
+
     # Update column names
     custom_column_names = ['Model', 'Clemscore', '% Played', 'Quality Score']
     for i, col in enumerate(df.columns[4:]):  # Start Capitalizing from the 5th column
@@ -107,7 +120,7 @@ def process_df(df: pd.DataFrame) -> pd.DataFrame:
 
     # Rename columns
     df.columns = custom_column_names
-
+ 
     return df
 
 

src/plot_utils.py

@@ -4,7 +4,7 @@ import requests
 import json
 import gradio as gr
 
-from src.assets.text_content import SHORT_NAMES, TEXT_NAME, MULTIMODAL_NAME
+from src.assets.text_content import SHORT_NAMES, TEXT_NAME, MULTIMODAL_NAME, REGISTRY_URL
 from src.leaderboard_utils import get_github_data
 
 
@@ -131,8 +131,7 @@ def split_models(model_list: list):
     commercial_models = []
     
     # Load model registry data from main repo
-    model_registry_url = "https://raw.githubusercontent.com/clp-research/clembench/main/backends/model_registry.json"
-    response = requests.get(model_registry_url)
+    response = requests.get(REGISTRY_URL)
 
     if response.status_code == 200:
         json_data = json.loads(response.text)
@@ -149,7 +148,7 @@ def split_models(model_list: list):
                     break
 
     else:
-        print(f"Failed to read JSON file: Status Code : {response.status_code}")
+        print(f"Failed to read JSON file: {REGISTRY_URL} Status Code : {response.status_code}")
 
     open_models.sort(key=lambda o: o.upper())
     commercial_models.sort(key=lambda c: c.upper())

src/version_utils.py

@@ -27,7 +27,7 @@ def get_version_data():
 
     # Check if the JSON file request was successful
     if response.status_code != 200:
-        print(f"Failed to read JSON file: Status Code: {response.status_code}")
+        print(f"Failed to read JSON file {json_url}: Status Code: {response.status_code}")
         return None, None, None, None
 
     json_data = response.json()