update

pages/1_Injection.py

@@ -8,6 +8,19 @@ from util.model import AzureAgent, GPTAgent
 st.title('Result Generation')
 st.sidebar.title('Model Settings')
 
+def check_password():
+    def password_entered():
+        if password_input == os.getenv('PASSWORD'):
+            st.session_state['password_correct'] = True
+        else:
+            st.error("Incorrect Password, please try again.")
+
+    password_input = st.text_input("Enter Password:", type="password")
+    submit_button = st.button("Submit", on_click=password_entered)
+
+    if submit_button and not st.session_state.get('password_correct', False):
+        st.error("Please enter a valid password to access the demo.")
+
 
 # Define a function to manage state initialization
 def initialize_state():
@@ -21,77 +34,82 @@ def initialize_state():
             st.session_state[key] = default
 
 
-initialize_state()
-
-# Model selection and configuration
-model_type = st.sidebar.radio("Select the type of agent", ('GPTAgent', 'AzureAgent'))
-st.session_state.api_key = st.sidebar.text_input("API Key", type="password", value=st.session_state.api_key)
-st.session_state.endpoint_url = st.sidebar.text_input("Endpoint URL", value=st.session_state.endpoint_url)
-st.session_state.deployment_name = st.sidebar.text_input("Model Name", value=st.session_state.deployment_name)
-api_version = '2024-02-15-preview' if model_type == 'GPTAgent' else ''
-st.session_state.temperature = st.sidebar.slider("Temperature", 0.0, 1.0, st.session_state.temperature, 0.01)
-st.session_state.max_tokens = st.sidebar.number_input("Max Tokens", 1, 1000, st.session_state.max_tokens)
-
-if st.sidebar.button("Reset Model Info"):
-    initialize_state()  # Reset all state to defaults
-    st.experimental_rerun()
-
-if st.sidebar.button("Submit Model Info"):
-    st.session_state.model_submitted = True
-
-# Ensure experiment settings are only shown if model info is submitted
-if st.session_state.model_submitted:
-    df = None
-    file_options = st.radio("Choose file source:", ["Upload", "Example"])
-    if file_options == "Example":
-        df = pd.read_csv("prompt_test.csv")
-    else:
-        st.session_state.uploaded_file = st.file_uploader("Choose a file")
-        if st.session_state.uploaded_file is not None:
-            data = StringIO(st.session_state.uploaded_file.getvalue().decode("utf-8"))
-            df = pd.read_csv(data)
-    if df is not None:
-
-        st.write('Data:', df)
-
-        # Button to add a new row
-
-        st.session_state.occupation = st.text_input("Occupation", value=st.session_state.occupation)
-        st.session_state.group_name = st.text_input("Group Name", value=st.session_state.group_name)
-        st.session_state.privilege_label = st.text_input("Privilege Label", value=st.session_state.privilege_label)
-        st.session_state.protect_label = st.text_input("Protect Label", value=st.session_state.protect_label)
-        st.session_state.num_run = st.number_input("Number of Runs", 1, 10, st.session_state.num_run)
-
-        if st.button('Process Data') and not st.session_state.data_processed:
-            # Initialize the correct agent based on model type
-            if model_type == 'AzureAgent':
-                agent = AzureAgent(st.session_state.api_key, st.session_state.endpoint_url,
-                                   st.session_state.deployment_name)
-            else:
-                agent = GPTAgent(st.session_state.api_key, st.session_state.endpoint_url,
-                                 st.session_state.deployment_name, api_version)
-
-            # Process data and display results
-            with st.spinner('Processing data...'):
-                parameters = {"temperature": st.session_state.temperature, "max_tokens": st.session_state.max_tokens}
-                df = process_scores_multiple(df, st.session_state.num_run, parameters, st.session_state.privilege_label,st.session_state.protect_label, agent, st.session_state.group_name,st.session_state.occupation)
-                st.session_state.data_processed = True  # Mark as processed
-
-            st.write('Processed Data:', df)
-
-            # Allow downloading of the evaluation results
-            st.download_button(
-                label="Download Generation Results",
-                data=df.to_csv().encode('utf-8'),
-                file_name='generation_results.csv',
-                mime='text/csv',
-            )
-
-        if st.button("Reset Experiment Settings"):
-            st.session_state.occupation = "Programmer"
-            st.session_state.group_name = "Gender"
-            st.session_state.privilege_label = "Male"
-            st.session_state.protect_label = "Female"
-            st.session_state.num_run = 1
-            st.session_state.data_processed = False
-            st.session_state.uploaded_file = None
+if not st.session_state.get('password_correct', False):
+    check_password()
+else:
+    st.sidebar.success("Password Verified. Proceed with the demo.")
+
+    initialize_state()
+
+    # Model selection and configuration
+    model_type = st.sidebar.radio("Select the type of agent", ('GPTAgent', 'AzureAgent'))
+    st.session_state.api_key = st.sidebar.text_input("API Key", type="password", value=st.session_state.api_key)
+    st.session_state.endpoint_url = st.sidebar.text_input("Endpoint URL", value=st.session_state.endpoint_url)
+    st.session_state.deployment_name = st.sidebar.text_input("Model Name", value=st.session_state.deployment_name)
+    api_version = '2024-02-15-preview' if model_type == 'GPTAgent' else ''
+    st.session_state.temperature = st.sidebar.slider("Temperature", 0.0, 1.0, st.session_state.temperature, 0.01)
+    st.session_state.max_tokens = st.sidebar.number_input("Max Tokens", 1, 1000, st.session_state.max_tokens)
+
+    if st.sidebar.button("Reset Model Info"):
+        initialize_state()  # Reset all state to defaults
+        st.experimental_rerun()
+
+    if st.sidebar.button("Submit Model Info"):
+        st.session_state.model_submitted = True
+
+    # Ensure experiment settings are only shown if model info is submitted
+    if st.session_state.model_submitted:
+        df = None
+        file_options = st.radio("Choose file source:", ["Upload", "Example"])
+        if file_options == "Example":
+            df = pd.read_csv("prompt_test.csv")
+        else:
+            st.session_state.uploaded_file = st.file_uploader("Choose a file")
+            if st.session_state.uploaded_file is not None:
+                data = StringIO(st.session_state.uploaded_file.getvalue().decode("utf-8"))
+                df = pd.read_csv(data)
+        if df is not None:
+
+            st.write('Data:', df)
+
+            # Button to add a new row
+
+            st.session_state.occupation = st.text_input("Occupation", value=st.session_state.occupation)
+            st.session_state.group_name = st.text_input("Group Name", value=st.session_state.group_name)
+            st.session_state.privilege_label = st.text_input("Privilege Label", value=st.session_state.privilege_label)
+            st.session_state.protect_label = st.text_input("Protect Label", value=st.session_state.protect_label)
+            st.session_state.num_run = st.number_input("Number of Runs", 1, 10, st.session_state.num_run)
+
+            if st.button('Process Data') and not st.session_state.data_processed:
+                # Initialize the correct agent based on model type
+                if model_type == 'AzureAgent':
+                    agent = AzureAgent(st.session_state.api_key, st.session_state.endpoint_url,
+                                       st.session_state.deployment_name)
+                else:
+                    agent = GPTAgent(st.session_state.api_key, st.session_state.endpoint_url,
+                                     st.session_state.deployment_name, api_version)
+
+                # Process data and display results
+                with st.spinner('Processing data...'):
+                    parameters = {"temperature": st.session_state.temperature, "max_tokens": st.session_state.max_tokens}
+                    df = process_scores_multiple(df, st.session_state.num_run, parameters, st.session_state.privilege_label,st.session_state.protect_label, agent, st.session_state.group_name,st.session_state.occupation)
+                    st.session_state.data_processed = True  # Mark as processed
+
+                st.write('Processed Data:', df)
+
+                # Allow downloading of the evaluation results
+                st.download_button(
+                    label="Download Generation Results",
+                    data=df.to_csv().encode('utf-8'),
+                    file_name='generation_results.csv',
+                    mime='text/csv',
+                )
+
+            if st.button("Reset Experiment Settings"):
+                st.session_state.occupation = "Programmer"
+                st.session_state.group_name = "Gender"
+                st.session_state.privilege_label = "Male"
+                st.session_state.protect_label = "Female"
+                st.session_state.num_run = 1
+                st.session_state.data_processed = False
+                st.session_state.uploaded_file = None

pages/2_Evaluation.py

@@ -1,63 +1,83 @@
+import os
+
 import streamlit as st
 import pandas as pd
 from io import StringIO
 from util.evaluation import statistical_tests,calculate_correlations,calculate_divergences
 
+def check_password():
+    def password_entered():
+        if password_input == os.getenv('PASSWORD'):
+            st.session_state['password_correct'] = True
+        else:
+            st.error("Incorrect Password, please try again.")
+
+    password_input = st.text_input("Enter Password:", type="password")
+    submit_button = st.button("Submit", on_click=password_entered)
+
+    if submit_button and not st.session_state.get('password_correct', False):
+        st.error("Please enter a valid password to access the demo.")
+
 def app():
     st.title('Result Evaluation')
 
-    # Allow users to upload a CSV file with processed results
-    uploaded_file = st.file_uploader("Upload your processed CSV file", type="csv")
-    if uploaded_file is not None:
-        data = StringIO(uploaded_file.getvalue().decode('utf-8'))
-        df = pd.read_csv(data)
-
-        # Add ranks for each score within each row
-        ranks = df[['Privilege_Avg_Score', 'Protect_Avg_Score', 'Neutral_Avg_Score']].rank(axis=1, ascending=False)
-
-        df['Privilege_Rank'] = ranks['Privilege_Avg_Score']
-        df['Protect_Rank'] = ranks['Protect_Avg_Score']
-        df['Neutral_Rank'] = ranks['Neutral_Avg_Score']
-
-        st.write('Uploaded Data:', df)
-
-        if st.button('Evaluate Data'):
-            with st.spinner('Evaluating data...'):
-                # Existing statistical tests
-                test_results = statistical_tests(df)
-                st.write('Test Results:', test_results)
-                # evaluation_results = result_evaluation(test_results)
-                # st.write('Evaluation Results:', evaluation_results)
-
-                # New correlation calculations
-                correlation_results = calculate_correlations(df)
-                st.write('Correlation Results:', correlation_results)
-
-                # New divergence calculations
-                divergence_results = calculate_divergences(df)
-                st.write('Divergence Results:', divergence_results)
-
-                # Flatten the results for combining
-                flat_test_results = {f"{key1}_{key2}": value2 for key1, value1 in test_results.items() for key2, value2
-                                     in (value1.items() if isinstance(value1, dict) else {key1: value1}.items())}
-                flat_correlation_results = {f"Correlation_{key1}": value1 for key1, value1 in
-                                            correlation_results.items()}
-                flat_divergence_results = {f"Divergence_{key1}": value1 for key1, value1 in divergence_results.items()}
-
-                # Combine all results
-                results_combined = {**flat_test_results, **flat_correlation_results, **flat_divergence_results}
-
-                # Convert to DataFrame for download
-                results_df = pd.DataFrame(list(results_combined.items()), columns=['Metric', 'Value'])
-
-                st.write('Combined Results:', results_df)
-
-                st.download_button(
-                    label="Download Evaluation Results",
-                    data=results_df.to_csv(index=False).encode('utf-8'),
-                    file_name='evaluation_results.csv',
-                    mime='text/csv',
-                )
+    if not st.session_state.get('password_correct', False):
+        check_password()
+    else:
+        st.sidebar.success("Password Verified. Proceed with the demo.")
+
+        # Allow users to upload a CSV file with processed results
+        uploaded_file = st.file_uploader("Upload your processed CSV file", type="csv")
+        if uploaded_file is not None:
+            data = StringIO(uploaded_file.getvalue().decode('utf-8'))
+            df = pd.read_csv(data)
+
+            # Add ranks for each score within each row
+            ranks = df[['Privilege_Avg_Score', 'Protect_Avg_Score', 'Neutral_Avg_Score']].rank(axis=1, ascending=False)
+
+            df['Privilege_Rank'] = ranks['Privilege_Avg_Score']
+            df['Protect_Rank'] = ranks['Protect_Avg_Score']
+            df['Neutral_Rank'] = ranks['Neutral_Avg_Score']
+
+            st.write('Uploaded Data:', df)
+
+            if st.button('Evaluate Data'):
+                with st.spinner('Evaluating data...'):
+                    # Existing statistical tests
+                    test_results = statistical_tests(df)
+                    st.write('Test Results:', test_results)
+                    # evaluation_results = result_evaluation(test_results)
+                    # st.write('Evaluation Results:', evaluation_results)
+
+                    # New correlation calculations
+                    correlation_results = calculate_correlations(df)
+                    st.write('Correlation Results:', correlation_results)
+
+                    # New divergence calculations
+                    divergence_results = calculate_divergences(df)
+                    st.write('Divergence Results:', divergence_results)
+
+                    # Flatten the results for combining
+                    flat_test_results = {f"{key1}_{key2}": value2 for key1, value1 in test_results.items() for key2, value2
+                                         in (value1.items() if isinstance(value1, dict) else {key1: value1}.items())}
+                    flat_correlation_results = {f"Correlation_{key1}": value1 for key1, value1 in
+                                                correlation_results.items()}
+                    flat_divergence_results = {f"Divergence_{key1}": value1 for key1, value1 in divergence_results.items()}
+
+                    # Combine all results
+                    results_combined = {**flat_test_results, **flat_correlation_results, **flat_divergence_results}
+
+                    # Convert to DataFrame for download
+                    results_df = pd.DataFrame(list(results_combined.items()), columns=['Metric', 'Value'])
+
+                    st.write('Combined Results:', results_df)
+
+                    st.download_button(
+                        label="Download Evaluation Results",
+                        data=results_df.to_csv(index=False).encode('utf-8'),
+                        file_name='evaluation_results.csv',
+                        mime='text/csv',
+                    )
 
 if __name__ == "__main__":
     app()