restricting the number of rows to display on app

src/display.py

@@ -4,7 +4,7 @@
 #from st_aggrid import GridOptionsBuilder, AgGrid
 from st_aggrid import GridOptionsBuilder, AgGrid
 import streamlit as st
-from .load_data import load_dataframe, sort_by
+from .load_data import load_dataframe, sort_by, show_dataframe_top, search_by_name
 from .plot import plot_radar_chart_name, plot_radar_chart_rows
 
 
@@ -12,16 +12,16 @@ def display_app():
     st.markdown("# Open LLM Leaderboard Viz")
     st.markdown("This is a visualization of the results in [open-llm-leaderboard/results](https://huggingface.co/datasets/open-llm-leaderboard/results)")
     st.markdown("To select a model, click on the checkbox beside its name.")
-
-    
-    
-    #container = st.container(height = 150)
+    st.markdown("This displays the top 100 models by default, but you can change that using the number input below.") 
+    st.markdown("By defalut as well, the maximum number of row you can display is 500, it is due to the problem with st__aggrid component loading.")
+    st.markdown("If your model doesn't show up, please search it by its name.")
 
     dataframe = load_dataframe()
 
-    sort_selection = st.selectbox(label = "Sort by:", options = list(dataframe.columns))
+    sort_selection = st.selectbox(label = "Sort by:", options = list(dataframe.columns), index = 7)
+    number_of_row = st.number_input("Number of top rows to display", min_value=100, max_value=500, value="min", step=100)
     ascending = True
-    indexes = None
+    
     if sort_selection is None:
         sort_selection = "model_name"
         ascending = True
@@ -29,16 +29,26 @@ def display_app():
         ascending = True
     else:
         ascending = False
+
+
     name = st.text_input(label = ":mag: Search by name")
-    if name is not None:
-        indexes = dataframe["model_name"].str.contains(name)
-        if len(indexes) > 0:
-            dataframe = dataframe[indexes]
+    len_name_input = len(name) 
+    if len_name_input > 0:
+        dataframe_by_search = search_by_name(name)
+        if len(dataframe_by_search) > 0:
+            #st.write("number of model name with name", len(dataframe_by_search))
+            dataframe = dataframe_by_search
         else:
             dataframe = load_dataframe()
 
     dataframe = sort_by(dataframe=dataframe, column_name=sort_selection, ascending= ascending)
     dataframe_display = dataframe.copy()
+    
+    if len_name_input == 0:
+        # Show every only top n row
+        dataframe_display = show_dataframe_top(number_of_row,dataframe_display)    
+        
+    
     dataframe_display[["ARC", "HellaSwag", "TruthfulQA", "Winogrande", "GSM8K" ,"MMLU", "Average"]] = dataframe[["ARC", "HellaSwag", "TruthfulQA", "Winogrande", "GSM8K" ,"MMLU", "Average"]].astype(float)
     dataframe_display[["ARC", "HellaSwag", "TruthfulQA", "Winogrande", "GSM8K" ,"MMLU", "Average"]] = dataframe_display[["ARC", "HellaSwag", "TruthfulQA", "Winogrande", "GSM8K" ,"MMLU", "Average"]] *100
     dataframe_display[["ARC", "HellaSwag", "TruthfulQA", "Winogrande", "GSM8K" ,"MMLU", "Average"]] = dataframe_display[["ARC", "HellaSwag", "TruthfulQA", "Winogrande", "GSM8K" ,"MMLU", "Average"]].round(2)
@@ -52,8 +62,6 @@ def display_app():
     column1,col3, column2 = st.columns([0.26, 0.05, 0.69], gap = "small")
 
     with column1:
-        #with container:
-        #st.dataframe(dataframe_display)
         grid_response = AgGrid(
     dataframe_display, 
     gridOptions=gridOptions,

src/load_data.py

@@ -13,6 +13,19 @@ def load_dataframe() -> pd.DataFrame:
     dataframe = dataframe.drop(columns = "Unnamed: 0")
     return dataframe
 
+@st.cache_data
+def show_dataframe_top(n:int , dataframe: pd.DataFrame) -> pd.DataFrame:
+    """
+    read only the n-th first row
+    Arguments
+    -n: an integer telling the number of row
+    -dataframe: the dataframe to slice
+    Returns
+    dataframe: a pd.DataFrame of the average scores of the LLMs on each task
+    """
+
+    return dataframe.head(n)
+
 @st.cache_data
 def sort_by(dataframe: pd.DataFrame, column_name: str, ascending:bool = False) -> pd.DataFrame:
     """
@@ -26,4 +39,19 @@ def sort_by(dataframe: pd.DataFrame, column_name: str, ascending:bool = False) -
     Returns:
     a sorted dataframe
     """
-    return dataframe.sort_values(by = column_name, ascending = ascending )
+    return dataframe.sort_values(by = column_name, ascending = ascending )
+
+@st.cache_data
+def search_by_name(name: str) -> pd.DataFrame:
+    """
+    Search a model by its name
+
+    Arguments:
+    - name: the name of the model or part of it
+
+    Returns:
+    a pandas Dataframe of every row that contains name
+    """
+    dataframe = load_dataframe()
+    indexes = dataframe["model_name"].str.contains(name)
+    return dataframe[indexes]