feedback_tool.py

import psycopg2
import random
import streamlit as st
from transformers import BartForConditionalGeneration, BartTokenizer

# establishing connection to the database used
database_name = "postgres"
database_user = "souadyakubu"
database_password = "souadseny"
database_host = "localhost"
database_port = "5432"

my_connection = psycopg2.connect(database = database_name , user = database_user , password = database_password, host = database_host, port = database_port)
cursor = my_connection.cursor()


# making a feedback table 
with my_connection.cursor() as cursor:
    cursor.execute('''
        CREATE TABLE IF NOT EXISTS feedback (
            id SERIAL PRIMARY KEY,
            context TEXT,
            question TEXT,
            system_generated TEXT,
            rater_id INTEGER,
            rank INTEGER
        )
    ''')
    my_connection.commit()

@st.cache(allow_output_mutation=True)
def generate_questions(model_choice, context):
    pass

def feedback_log(context, question, system, rater_id, rank):
    # putting the feedback data into the PostgreSQL database
    with my_connection.cursor() as cursor:
        cursor.execute(
            "INSERT INTO feedback (context, question, system_generated, rater_id, rank) VALUES (%s, %s, %s, %s, %s)",
            (context, question, system, rater_id, rank)
        )
        my_connection.commit()

st.title("Feedback Collection Tool")

    # collecting contexts 
    # contexts = st.text_area("Enter the contexts (one per line):").split("\n")

    #selecting random context
    # random_context = random.choice(contexts)

    #selecting some questions for the random context 
    # random_questions = generate_questions(model_choice, random_context)

    # for demonstration purposes, I'll assume we already have a random context and its generated questions.
random_context = "This is a sample context."
random_questions = ["What is your name?", "What are your hobbies?", "Tell me about your work experience."]

st.subheader("Context:")
st.write(random_context)

st.subheader("Generated Questions:")
for i, question in enumerate(random_questions):
    st.write(f"{i + 1}. {question}")
        
    #User Feedback
selected_question_index = st.selectbox("Select the best question:", [i + 1 for i in range(len(random_questions))])

rank = st.slider("Rate the question (1 = worst, 5 = best):", 1, 5)

if st.button("Submit Feedback"):
    selected_question = random_questions[selected_question_index - 1]
    system_that_generated = "hyechanjun/interview-question-remake"
    rater_id = 1  #can be changed based on user authentication
    feedback_log(random_context, selected_question, system_that_generated, rater_id, rank)
    st.success("Feedback submitted successfully!")