add leaderboard filter

README.md

@@ -16,6 +16,11 @@ Check out the configuration reference at https://huggingface.co/docs/hub/spaces-
 
 
 # de-identification-leaderboard
+## run web
+```
+streamlit run app.py
+```
+
 
 ##  leaderboard data
 score wil save to huggingface dataset

app.py

@@ -12,6 +12,7 @@ from huggingface_hub import hf_hub_download
 from utils.upload_hub import upload_scores_to_hub, file_name_decode
 from utils.Evaluation_answer_txt import Evaluation_answer_txt
 from utils.hub_info import get_model_size
+from utils.filterable_dataframe import filterable_dataframe
 
 # st.set_page_config(layout="wide")
 st.set_page_config(layout="centered")
@@ -78,10 +79,14 @@ try:
         df = get_leaderboard_df()
 
         # remove the detail result column
-        default_columns = [c for c in df.columns if c not in ['MICRO precision', 'MICRO recall', 'MACRO precision', 'MACRO recall',  'detail result']]
+        default_columns = [c for c in df.columns if c not in ['file name', 'submitter', 'MICRO precision', 'MICRO recall', 'MACRO precision', 'MACRO recall',  'detail result']]
         selected_columns = st.multiselect('Select columns to display', df.columns, default=default_columns)
 
-        leaderboard_df =    st.dataframe(df[selected_columns], selection_mode='multi-row', on_select='rerun', key='leaderboard')
+        # add filterable dataframe
+        filtered_df = filterable_dataframe(df)
+        
+        leaderboard_df =    st.dataframe(filtered_df[selected_columns], selection_mode='multi-row', on_select='rerun', key='leaderboard')
+
         
         st.subheader("Detail Result")
         det_ind = st.session_state.leaderboard['selection']['rows']

utils/filterable_dataframe.py

@@ -0,0 +1,37 @@
+import streamlit as st
+
+
+
+
+def filterable_dataframe(df):
+    # 多選框選擇需要篩選的列
+    selected_columns = st.multiselect('Select columns to filter', options=df.columns)
+
+    filters = {}
+
+    # 動態生成篩選條件
+    for column in selected_columns:
+        with st.expander(f'Filter by {column}', expanded=True):
+            if df[column].dtype == 'object':
+                unique_values = df[column].unique()
+                selected_values = st.multiselect(f'Select values for {column}', options=list(unique_values))
+                search_text = st.text_input(f'Search in {column}')
+                if search_text:
+                    selected_values = [value for value in selected_values if search_text.lower() in value.lower()]
+                if selected_values:
+                    filters[column] = selected_values
+            else:
+                min_value, max_value = df[column].min(), df[column].max()
+                selected_range = st.slider(f'Select range for {column}', min_value, max_value, (min_value, max_value))
+                if selected_range != (min_value, max_value):
+                    filters[column] = selected_range
+
+    # 根據篩選條件過濾數據框
+    filtered_df = df
+    for column, value in filters.items():
+        if isinstance(value, list):
+            filtered_df = filtered_df[filtered_df[column].isin(value)]
+        elif isinstance(value, tuple):
+            filtered_df = filtered_df[(filtered_df[column] >= value[0]) & (filtered_df[column] <= value[1])]
+
+    return filtered_df