app.py

import streamlit as st
from transformers import AutoTokenizer, AutoModelForQuestionAnswering, pipeline

# Set page configuration
st.set_page_config(
    page_title="Question Answering App",
    page_icon="❓",
    layout="centered",
    initial_sidebar_state="auto",
)

# Page title with custom style
st.markdown(
    """
    <h1 style="text-align: center; color: #4CAF50; font-family: 'Helvetica';">
        📚 Question Answering App
    </h1>
    <p style="text-align: center; color: #777; font-size: 18px;">
        Enter a context and question to get precise answers powered by AI.
    </p>
    """,
    unsafe_allow_html=True,
)

# Sidebar for model settings and context input
st.sidebar.header("Model Settings")
model_checkpoint = st.sidebar.text_input(
    "Model Checkpoint", "Diezu/viedumrc", help="Specify the model checkpoint to use."
)
model_checkpoint1 = 'Diezu/viedumrc'
question_answerer = pipeline("question-answering", model=model_checkpoint1)
st.sidebar.markdown(
    """
    <small>Using model: <code>Diezu/viedumrc</code>.</small>
    """,
    unsafe_allow_html=True,
)

context_sidebar = st.sidebar.text_area(
    "Context",
    "",
    help="Enter the context that contains information for answering questions.",
    height=200,
    placeholder="Provide context for your question...",
)

# # Load the tokenizer and model with error handling
# try:
#     tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)
#     model = AutoModelForQuestionAnswering.from_pretrained(model_checkpoint)
#     question_answerer = pipeline("question-answering", model=model, tokenizer=tokenizer)
# except Exception as e:
#     st.error(f"Failed to load model or tokenizer: {e}", icon="🚨")
#     st.stop()

# Main application
st.markdown(
    """
    <h2 style="color: #2196F3;">Provide Context and Question</h2>
    """,
    unsafe_allow_html=True,
)

# Input: question
question = st.text_input(
    "Question",
    "",
    help="Write the question you want to ask about the provided context.",
    placeholder="What is your question?",
)

# Button to get answer
if st.button("Get Answer"):
    if context_sidebar.strip() == "" or question.strip() == "":
        st.warning("Please provide both context and a question!", icon="⚠️")
    else:
        try:
            # Using context from the sidebar and question from the main section
            result = question_answerer(question=question, context=context_sidebar)
            st.success("Answer Found!", icon="✅")
            st.markdown(
                f"""
                <div style="background-color: #f1f8ff; border-left: 4px solid #2196F3; padding: 10px; margin-top: 10px;">
                    <strong>Answer:</strong> {result['answer']}
                </div>
                """,
                unsafe_allow_html=True,
            )
        except Exception as e:
            st.error(f"Error while processing: {e}", icon="🚨")

# Footer with custom style
st.markdown(
    """
    <hr>
    <footer style="text-align: center; font-size: small; color: #888;">
        Built with ❤️ using <strong>Streamlit</strong> and <strong>Transformers</strong>.
        <br>
        <small><a href="https://github.com/DINHD1" target="_blank">GitHub</a></small>
    </footer>
    """,
    unsafe_allow_html=True,
)

# import streamlit as st
# from transformers import pipeline

# st.title('Question Answering')
# model_checkpoint = 'Diezu/dieumrc'
# question_answerer = pipeline("question-answering", model=model_checkpoint)

# context = st.text_area('CONTEXT')
# question = st.text_input('QUESTION')

# if st.button('ANSWER'):
#     i=question_answerer(question,context)
#     #st.write(i['answer'])
#     st.markdown(i['answer']) 
#     #st.write('<h1 style="font-size: 50px;">i['answer']</h1>')