Update

app.py

@@ -1,20 +1,26 @@
 import streamlit as st
 import pandas as pd
 
+
 def load_data():
     return pd.read_csv("benchmark_data.csv")
 
 def case_insensitive_search(data, query, column):
-    if query:
+    if query: 
         return data[data[column].str.lower().str.contains(query.lower())]
     return data
 
+def display_table(data, rows_per_page=10):
+    container = st.container()
+    with container:
+        height = min(40 + rows_per_page * 38, 800)  
+        st.dataframe(data, height=height)
+
 def main():
     st.title("Multihop-RAG Benchmark 💡")
 
     data = load_data()
 
-    # 筛选条件
     st.sidebar.header("Search Options")
     chat_model_query = st.sidebar.text_input("Chat Model")
     embedding_model_query = st.sidebar.text_input("Embedding Model")
@@ -25,42 +31,28 @@ def main():
         data = case_insensitive_search(data, chat_model_query, 'chat_model')
     if embedding_model_query:
         data = case_insensitive_search(data, embedding_model_query, 'embedding_model')
-    if chunk_query:
+    if chunk_query:  
         data = case_insensitive_search(data, chunk_query, 'chunk')
     if frame_query:
         data = case_insensitive_search(data, frame_query, 'framework')
 
-    # 创建多级列索引
-    columns = [
-        ("Settings", "Framework"),
-        ("Settings", "Chat Model"),
-        ("Settings", "Embedding Model"),
-        ("Settings", "Chunk"),
-        ("Retrieval Metrics", "MRR@10"),
-        ("Retrieval Metrics", "Hit@10"),
-        ("Response Metrics", "Accuracy")
-    ]
+    st.write("Displaying results for MRR@10, Hit@10, and Accuracy across different frameworks, embedding models, chat models, and chunks.")
+    display_table(data) 
 
-    # 选择并设置多级列索引
-    data = data[['framework', 'chat_model', 'embedding_model', 'chunk', 'MRR@10', 'Hit@10', 'Accuracy']]
-    data.columns = pd.MultiIndex.from_tuples(columns)
+    # if st.sidebar.checkbox("Show Metrics Distribution"):
+    #     st.subheader("Metrics Distribution")
+    #     st.bar_chart(data[['MRR@10', 'Hit@10', 'Accuracy']]) 
 
-    # 展示数据
-    st.dataframe(data.style.set_table_styles([{
-        'selector': 'th',
-        'props': [('background-color', '#f4f4f4'), ('font-weight', 'bold')]
-    }]), height=600)
 
-    if st.sidebar.checkbox("Show Metrics Distribution"):
-        st.subheader("Metrics Distribution")
-        st.bar_chart(data[['MRR@10', 'Hit@10', 'Accuracy']])
-
-    # 引用
     st.sidebar.header("Citation")
     st.sidebar.info(
         "Please cite this dataset as:\n"
-        "Tang, Yixuan, and Yi Yang. MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries. ArXiv, 2024, /abs/2401.15391."
+        "Tang, Yixuan, and Yi Yang. MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries. ArXiv, 2024,  /abs/2401.15391."
     )
-
+    st.markdown("---")
+    st.caption("For citation, please use: 'Tang, Yixuan, and Yi Yang. MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries. ArXiv, 2024,  /abs/2401.15391. '")
+    st.markdown("---")
+    st.caption("For results self-reporting, please send an email to ytangch@connect.ust.hk")
+ 
 if __name__ == "__main__":
     main()