Spaces:
Sleeping
Sleeping
File size: 7,095 Bytes
a86e213 0ba415f a86e213 345d028 a86e213 bafdc7e a86e213 c3903ae 28e0c43 c3903ae a86e213 bafdc7e ae29644 bafdc7e 28e0c43 bafdc7e c3903ae 4069a9c bafdc7e c3903ae b0e0109 c050d3e 719373a bafdc7e 25199b3 bafdc7e d9d604b 6ac2c35 25199b3 bafdc7e c3903ae 16842d6 c3903ae bafdc7e 25199b3 bafdc7e 16842d6 bafdc7e 27478f5 bafdc7e 27478f5 bafdc7e 39ddee7 c3903ae bafdc7e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 |
import streamlit as st
import pandas as pd
from io import StringIO
from util.injection import process_scores_multiple
from util.model import AzureAgent, GPTAgent
import os
# Set up the Streamlit interface
st.title('Result Generation')
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():
keys = ["model_submitted", "api_key", "endpoint_url", "deployment_name", "temperature", "max_tokens",
"data_processed", "group_name", "occupation", "privilege_label", "protect_label", "num_run",
"uploaded_file", "occupation_submitted","sample_size","charateristics"]
defaults = [False, "", "https://safeguard-monitor.openai.azure.com/", "gpt35-1106", 0.0, 150, False, "Gender",
"Programmer", "Male", "Female", 1, None, False,1,"This candidate's performance during the internship at our institution was evaluated to be at the 50th percentile among current employees."]
for key, default in zip(keys, defaults):
if key not in st.session_state:
st.session_state[key] = default
if not st.session_state.get('password_correct', False):
check_password()
else:
st.sidebar.success("Password Verified. Proceed with the demo.")
st.sidebar.title('Model Settings')
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
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("resume_subsampled.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:
# categories = ["HR", "DESIGNER", "INFORMATION-TECHNOLOGY", "TEACHER", "ADVOCATE", "BUSINESS-DEVELOPMENT",
# "HEALTHCARE", "FITNESS", "AGRICULTURE", "BPO", "SALES", "CONSULTANT", "DIGITAL-MEDIA",
# "AUTOMOBILE", "CHEF", "FINANCE", "APPAREL", "ENGINEERING", "ACCOUNTANT", "CONSTRUCTION",
# "PUBLIC-RELATIONS", "BANKING", "ARTS", "AVIATION"]
categories = list(df["Occupation"].unique())
st.session_state.occupation = st.selectbox("Occupation", options=categories, index=categories.index(st.session_state.occupation) if st.session_state.occupation in categories else 0)
st.session_state.sample_size = st.number_input("Sample Size", 1, len(df), st.session_state.sample_size)
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.charateristics = st.text_area("Characteristics", value=st.session_state.charateristics)
st.session_state.num_run = st.number_input("Number of Runs", 1, 10, st.session_state.num_run)
df = df[df["Occupation"] == st.session_state.occupation]
df = df.sample(n=st.session_state.sample_size)
st.write('Data:', df)
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}
preprocessed_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.occupation3)#,st.session_state.charateristics)
st.session_state.data_processed = True # Mark as processed
st.write('Processed Data:', preprocessed_df)
# Allow downloading of the evaluation results
st.download_button(
label="Download Generation Results",
data=preprocessed_df.to_csv().encode('utf-8'),
file_name='generation_results.csv',
mime='text/csv',
)
if st.button("Reset Experiment Settings"):
st.session_state.sample_size = 1
st.session_state.charateristics = "This candidate's performance during the internship at our institution was evaluated to be at the 50th percentile among current employees."
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
|