Fix latest_result_path_per_model variable and refactor

app.py

@@ -1,15 +1,18 @@
+from functools import partial
+
 import gradio as gr
 
 from src.constants import SUBTASKS, TASKS
 from src.details import update_subtasks_component, update_load_details_component, load_details_dataframes, \
     display_details, update_sample_idx_component, clear_details
-from src.results import fetch_result_paths, filter_latest_result_path_per_model, update_load_results_component, \
-    load_results_dataframes, display_results, update_tasks_component, clear_results
-
+from src.results import update_load_results_component, \
+    load_results_dataframes, display_results, update_tasks_component, clear_results, \
+    filter_latest_result_path_per_model, fetch_result_paths
 
 
 # if __name__ == "__main__":
 latest_result_path_per_model = filter_latest_result_path_per_model(fetch_result_paths())
+load_results_dataframes = partial(load_results_dataframes, result_path_per_model=latest_result_path_per_model)
 
 with gr.Blocks(fill_height=True) as demo:
     gr.HTML("<h1 style='text-align: center;'>Compare Results of the 🤗 Open LLM Leaderboard</h1>")

src/results.py

@@ -4,7 +4,6 @@ import gradio as gr
 import pandas as pd
 from huggingface_hub import HfFileSystem
 
-from app import latest_result_path_per_model
 from src.constants import RESULTS_DATASET_ID, TASKS
 
 
@@ -24,34 +23,25 @@ def filter_latest_result_path_per_model(paths):
     return {model_id: max(paths) for model_id, paths in d.items()}
 
 
-def get_result_path_from_model(model_id, result_path_per_model):
-    return result_path_per_model[model_id]
-
-
 def update_load_results_component():
     return gr.Button("Load Results", interactive=True)
 
 
-def load_data(result_path) -> pd.DataFrame:
+def load_results_dataframe(model_id, result_path_per_model=None):
+    if not model_id or not result_path_per_model:
+        return
+    result_path = result_path_per_model[model_id]
     fs = HfFileSystem()
     with fs.open(result_path, "r") as f:
         data = json.load(f)
-    return data
-
-
-def load_results_dataframe(model_id):
-    if not model_id:
-        return
-    result_path = get_result_path_from_model(model_id, latest_result_path_per_model)
-    data = load_data(result_path)
     model_name = data.get("model_name", "Model")
     df = pd.json_normalize([{key: value for key, value in data.items()}])
     # df.columns = df.columns.str.split(".")  # .split return a list instead of a tuple
     return df.set_index(pd.Index([model_name])).reset_index()
 
 
-def load_results_dataframes(*model_ids):
-    return [load_results_dataframe(model_id) for model_id in model_ids]
+def load_results_dataframes(*model_ids, result_path_per_model=None):
+    return [load_results_dataframe(model_id, result_path_per_model=result_path_per_model) for model_id in model_ids]
 
 
 def display_results(task, *dfs):