import streamlit as st
import streamlit as st
from pprint import pprint
import subprocess
cmd = ["python", "-m", "spacy", "download", "en_core_web_sm"]
subprocess.run(cmd)
from spacy.cli import download
from Questgen import main, main2
from PyPDF2 import PdfReader
from transformers import pipeline
from PyPDF2 import PdfReader
import nltk
import pandas as pd
nltk.download('punkt')
# st.title(body='7 - Question Generation')


def get_pdf_text(pdf_docs):
    text = ""
    for pdf in pdf_docs:
        pdf_reader = PdfReader(pdf)
        for page in pdf_reader.pages:
            text += page.extract_text()
    return text


########################################################
# st.header(body='Proposition 1',divider='red')

# if st.toggle(label='Show Proposition 1'):
st.title('Generate Questions from PDFs')
file = st.file_uploader(label='Upload',accept_multiple_files=True)
pr = st.button(label='Process')
if pr:
    # pr = st.button(label='Process')
    raw_text = get_pdf_text(file)
    # questions = []
    # ge = main.QGen()
    ge = main2.QGen()
    payload = {
        'input_text' : raw_text,
        # 'max_questions':2,
    }
    output = ge.predict_mcq(payload=payload)
    st.header(body='*Generated Questions are:*', divider='orange')
    for question in output['questions']:
        st.subheader(body=f":orange[Q{question['id']}:] {question['question_statement']}", divider='blue')
        st.markdown(f"A: {question['answer']}")
        c = 0
        for option in question['options']:
            # st.markdown(f"{c}")
            c+=1
            if c==1:
                st.markdown(f"B: {option}")
            elif c==2:
                st.markdown(f"C: {option}")
            elif c==3:
                st.markdown(f"D: {option}")


    if output is not None:
        # Convert the dictionary to a DataFrame
        df = pd.DataFrame(output['questions'])

        # Convert the options from lists to strings
        # df['options'] = df['options'].apply(lambda x: ','.join(x))
        df = df.drop(labels=['options_algorithm','extra_options','context','question_type'],axis=1)
        # Convert the DataFrame to CSV
        csv = df.to_csv(index=False).encode('utf-8')
        st.download_button(
            label='Download Data',
            data=csv,
            file_name='Generated MCQs.csv',
            mime='text/csv'
        )

    if st.toggle(label='Show Raw Output'):
        st.write(output)