feat(app): update app functionality and connect to server - Adithya S K

app.py

@@ -1,3 +1,4 @@
+import os
 import re
 import streamlit as st
 import requests
@@ -6,22 +7,94 @@ from io import StringIO
 import plotly.graph_objs as go
 from huggingface_hub import HfApi
 from huggingface_hub.utils import RepositoryNotFoundError, RevisionNotFoundError
+from dotenv import load_dotenv
 
+load_dotenv()
+
+SERVER_URL = os.getenv("SERVER_URL")
+
+def get_data():
+    response = requests.get(SERVER_URL)
+    data = response.json()
+    return data
 
 def main():
+    
     st.set_page_config(page_title="Indic LLM Leaderboard", layout="wide")
-
+    
     title_column, refresh_column = st.columns([.92, 0.08])
     with title_column:
-        st.title("Indic LLM Leaderboard")
-        # st.title("\U0001F1EE\U0001F1F3 Indic LLM Leaderboard")
-        st.markdown("Leaderboard made with 🧐 [Easy Eval](hhttps://github.com/adithya-s-k/easy_eval) using [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard) benchmark suite.")
+        st.title("Indic LLM Leaderboard (α)")
+        st.markdown("The Indic Eval Leaderboard utilizes the [indic_eval](https://github.com/adithya-s-k/indic_eval) evaluation framework , incorporating SOTA translated benchmarks like ARC, Hellaswag, MMLU, among others. Supporting 7 Indic languages, it offers a comprehensive platform for assessing model performance and comparing results within the Indic language modeling landscape.")
     with refresh_column:
-        st.button("Refresh", type="primary")
+        if st.button("Refresh", type="primary"):
+            data = get_data()
     
     Leaderboard_tab, About_tab ,FAQ_tab, Submit_tab = st.tabs(["🏅 Leaderboard", "📝 About" , "❗FAQ","🚀 Submit"])
     
     with Leaderboard_tab:
+        data = get_data()
+        
+        table_data = []
+        all_models = []
+        
+        for item in data:
+            model_name = item.get("name")
+            language = item.get("language")
+            try:
+                ALL = item["result"]["all"]["acc_norm"]
+            except KeyError:
+                ALL = None
+            try:
+                ARC_Easy = item["result"]["ARC-Easy"]["acc_norm"]
+            except KeyError:
+                ARC_Easy = None
+            try:
+                ARC_Challenge = item["result"]["ARC-Challenge"]["acc_norm"]
+            except KeyError:
+                ARC_Challenge = None
+            try:
+                Hellaswag = item["result"]["Hellaswag"]["acc_norm"]
+            except KeyError:
+                Hellaswag = None
+            try:
+                Boolq = item["result"]["Boolq"]["acc_norm"]
+            except KeyError:
+                Boolq = None
+            try:
+                MMLU = item["result"]["MMLU"]["acc_norm"]
+            except KeyError:
+                MMLU = None
+            try:
+                Winograde = item["result"]["Winograde"]["acc_norm"]
+            except KeyError:
+                Winograde = None
+            try:
+                Translation = item["result"]["Translation"]["acc_norm"]
+            except KeyError:
+                Translation = None
+            try:
+                Generation = item["result"]["Generation"]["acc_norm"]
+            except KeyError:
+                Generation = None
+            
+            all_models.append(model_name)
+            table_data.append({
+                "Model Name": model_name,
+                "Language": language,
+                "Avergae": ALL,
+                "ARC-Easy": ARC_Easy,
+                "ARC-Challenge": ARC_Challenge,
+                "Hellaswag": Hellaswag,
+                "Boolq": Boolq,
+                "MMLU": MMLU,
+                "Winograde": Winograde,
+                "Translation": Translation,
+                "Generation": Generation
+            })
+
+        df = pd.DataFrame(table_data)
+        
         title = st.text_input('Model Name', placeholder=" 🔍 Search for your model (separate multiple queries with `;`) and press ENTER...")
         
         col1, col2 = st.columns(2)
@@ -32,15 +105,50 @@ def main():
         with col2:
             language_options = st.multiselect(
                 'Pick Languages',
-                ['Kannada', 'Hindi', 'Tamil', 'Telegu','Gujarathi','Marathi','Malayalam'],['Kannada', 'Hindi', 'Tamil', 'Telegu','Gujarathi','Marathi','Malayalam'])
-        
-        df = pd.DataFrame(columns=benchmark_options)
-        
-        st.dataframe(df, use_container_width=True ) 
+                ['kannada', 'hindi', 'tamil', 'telegu','gujarathi','marathi','malayalam'],['kannada', 'hindi', 'tamil', 'telegu','gujarathi','marathi','malayalam'])
+
+        if title:
+            if ';' in title:
+                model_names = [name.strip() for name in title.split(';')]
+                filtered_df = df[df['Model Name'].isin(model_names)]
+            else:
+                filtered_df = df[df['Model Name'].str.contains(title, case=False, na=False)]
             
+            filtered_df = filtered_df[filtered_df['Language'].isin(language_options)]
+            filtered_df = filtered_df[df.columns.intersection(['Model Name', 'Language'] + benchmark_options)]
+
+            # Calculate average across selected benchmark columns
+            filtered_df['Average'] = filtered_df[benchmark_options].mean(axis=1)
+
+            # Display the filtered DataFrame
+            st.dataframe(filtered_df, use_container_width=True)
+        elif benchmark_options or language_options:
+            filtered_df = df[df['Language'].isin(language_options)]
+            filtered_df = filtered_df[df.columns.intersection(['Model Name', 'Language'] + benchmark_options)]
+
+            # Calculate average across selected benchmark columns
+            filtered_df['Average'] = filtered_df[benchmark_options].mean(axis=1)
+
+            st.dataframe(filtered_df, use_container_width=True)
+
+        # Multiselect for comparing models
         compare_models = st.multiselect(
-            'Pick Modles to compare them',
-            ['CognitiveLab/Ambari-7b', 'CognitiveLab/Ambari-7b-Hindi'])
+            'Pick Models to compare them',
+            df['Model Name'].unique()
+        )
+
+        # Display DataFrame for selected models and their scores
+        if compare_models:
+            compare_data = []
+            for model in compare_models:
+                model_data = df[df['Model Name'] == model]
+                compare_data.append(model_data)
+            if compare_data:
+                compare_df = pd.concat(compare_data)
+                compare_df['Average'] = compare_df[benchmark_options].mean(axis=1) # Calculate average
+                st.dataframe(compare_df, use_container_width=True)
+                
+
 
     # About tab
     with About_tab:
@@ -50,20 +158,31 @@ def main():
             ### Indic Eval
             
             ### Contribute
-            
         ''')
         
     # FAQ tab
     with FAQ_tab:
         st.markdown('''
-                    ### FAQ 
-                    ''')
-    
-    # Subit tab
+            ### FAQ 
+                        
+            ### SUBMISSIONS
+            
+            
+            ### RESULTS
+            
+            
+            ### EDITING SUBMISSIONS
+            
+            
+            ### OTHER
+        ''')
+
+    # Submit tab
     with Submit_tab:
         st.markdown('''
-                    ### Submit Your Model
-                    ''')
+            ### Submit Your Model
+        ''')
+
         
     with st.expander(label="📙 Citation"):
         code = '''
@@ -77,8 +196,5 @@ def main():
         '''
         st.code(code, language='python')
         
-    
-    
-    
 if __name__ == "__main__":
     main()