Small updates for the leaderboard

app.py

@@ -3,10 +3,10 @@ import pandas as pd
 from tqdm.auto import tqdm
 from utils import *
 import gradio as gr
+
 from huggingface_hub import HfApi, hf_hub_download
 from huggingface_hub.repocard import metadata_load
 
-
 class DeepRL_Leaderboard:
     def __init__(self) -> None:
         self.leaderboard= {} 
@@ -16,20 +16,17 @@ class DeepRL_Leaderboard:
             id = id.strip()
             title = title.strip()
             self.leaderboard.update({id:{'title':title,'data':get_data_per_env(id)}})
+            
     def get_data(self):
         return self.leaderboard
 
     def get_ids(self):
         return list(self.leaderboard.keys())
 
-          
-
 # CSS file for the
 with open('app.css','r') as f:
     BLOCK_CSS = f.read() 
 
-
-
 LOADED_MODEL_IDS = {}
 LOADED_MODEL_METADATA = {}
 
@@ -83,17 +80,17 @@ def get_data_per_env(rl_env):
 
 
 rl_leaderboard = DeepRL_Leaderboard()
-rl_leaderboard.add_leaderboard('CarRacing-v0'," The Car Racing 🏎️ Leaderboard")
-rl_leaderboard.add_leaderboard('MountainCar-v0',"The Mountain Car ⛰️ 🚗 Leaderboard")
+rl_leaderboard.add_leaderboard('CartPole-v1','The Cartpole-v1 Leaderboard')
 rl_leaderboard.add_leaderboard('LunarLander-v2',"The Lunar Lander 🌕 Leaderboard")
-rl_leaderboard.add_leaderboard('BipedalWalker-v3',"The BipedalWalker Leaderboard")
-rl_leaderboard.add_leaderboard('Taxi-v3','The Taxi-v3🚖 Leaderboard')
 rl_leaderboard.add_leaderboard('FrozenLake-v1-4x4-no_slippery','The FrozenLake-v1-4x4-no_slippery Leaderboard')
 rl_leaderboard.add_leaderboard('FrozenLake-v1-8x8-no_slippery','The FrozenLake-v1-8x8-no_slippery Leaderboard')
 rl_leaderboard.add_leaderboard('FrozenLake-v1-4x4','The FrozenLake-v1-4x4 Leaderboard')
 rl_leaderboard.add_leaderboard('FrozenLake-v1-8x8','The FrozenLake-v1-8x8 Leaderboard')
+rl_leaderboard.add_leaderboard('Taxi-v3','The Taxi-v3🚖 Leaderboard')
+rl_leaderboard.add_leaderboard('CarRacing-v0'," The Car Racing 🏎️ Leaderboard")
+rl_leaderboard.add_leaderboard('MountainCar-v0',"The Mountain Car ⛰️ 🚗 Leaderboard")
+rl_leaderboard.add_leaderboard('BipedalWalker-v3',"The BipedalWalker Leaderboard")
 rl_leaderboard.add_leaderboard('SpaceInvadersNoFrameskip-v4','The SpaceInvadersNoFrameskip-v4 Leaderboard')
-rl_leaderboard.add_leaderboard('CartPole-v1','The Cartpole-v1 Leaderboard')
 rl_leaderboard.add_leaderboard('Pong-PLE-v0','The Pong-PLE-v0 🎾 Leaderboard')
 rl_leaderboard.add_leaderboard('Walker2DBulletEnv-v0','The Walker2DBulletEnv-v0 🤖 Leaderboard')
 rl_leaderboard.add_leaderboard('AntBulletEnv-v0','The AntBulletEnv-v0 🕸️ Leaderboard')
@@ -102,15 +99,11 @@ RL_ENVS = rl_leaderboard.get_ids()
 RL_DETAILS = rl_leaderboard.get_data()
 
 
-
 def update_data(rl_env):
     global LOADED_MODEL_IDS,LOADED_MODEL_METADATA
     data = []
-    model_ids = [x for x in get_model_ids(rl_env) if x not in LOADED_MODEL_IDS[rl_env]]
+    model_ids = [x for x in get_model_ids(rl_env)] #if x not in LOADED_MODEL_IDS[rl_env]] # For now let's update all
     
-    #model_ids = [x for x in get_model_ids(rl_env) if x not in LOADED_MODEL_IDS[rl_env] or LOADED_MODEL_METADATA[x]!=get_metadata(x)] this still calls get_metadata() so won't work
-    #model_ids = [x for x in get_model_ids(rl_env)]
-
     LOADED_MODEL_IDS[rl_env]+=model_ids
 
     for model_id in tqdm(model_ids):
@@ -134,7 +127,6 @@ def update_data(rl_env):
     return pd.DataFrame.from_records(data)
 
 
-
 def update_data_per_env(rl_env):
     global RL_DETAILS
 
@@ -165,8 +157,6 @@ def update_data_per_env(rl_env):
         return html,dataframe,dataframe.empty   
 
 
-
-
 def get_info_display(dataframe,env_name,name_leaderboard,is_empty):
     if not is_empty:
         markdown = """
@@ -175,7 +165,7 @@ def get_info_display(dataframe,env_name,name_leaderboard,is_empty):
         <br>
         <p> This is a leaderboard of <b>{len_dataframe}</b> agents, from <b>{num_unique_users}</b> unique users, playing {env_name} 👩‍🚀. </p>
         <br>
-        <p> We use lower bound result to sort the models: mean_reward - std_reward. </p>
+        <p> We use <b>lower bound result to sort the models: mean_reward - std_reward.</b> </p>
         <br>    
         <p> You can click on the model's name to be redirected to its model card which includes documentation. </p>
         <br>