EE21's picture
Update app.py
da1f55e
raw
history blame
4.38 kB
import streamlit as st
import PyPDF2
from extractive_summarization import summarize_with_textrank, summarize_with_lsa
from abstractive_model import summarize_with_bart
from keyword_extraction import extract_keywords
from keyphrase_extraction import extract_sentences_with_obligations
#from blanc import BlancHelp
# Set page to wide mode
st.set_page_config(layout="wide")
# Function to handle file upload and return its content
def load_pdf(file):
pdf_reader = PyPDF2.PdfReader(file)
pdf_text = ""
for page_num in range(len(pdf_reader.pages)):
pdf_text += pdf_reader.pages[page_num].extract_text() or ""
return pdf_text
# Main app
def main():
st.title("Terms of Service Summarizer")
# Layout: 3 columns
col1, col2, col3 = st.columns([1, 3, 2], gap="large")
# Left column: Radio buttons for summarizer choice
with col1:
radio_options = ['Abstractive', 'Extractive (TextRank)', "Extractive (Latent Semantic Analysis)", 'Keyword Extraction', 'Keyphrase Extraction']
help_text = "Abstractive: Abstractive summarization generates a summary that may contain words not present in the original text. " \
"It uses a fine-tuned model on BART-large-CNN.<br>" \
"Extractive: Extractive summarization selects and extracts sentences or phrases directly from the original text to create a summary using the TextRank algorithm.<br>" \
"Keyword Extraction: Keyword extraction identifies and extracts important keywords or terms from the text using the Rake algorithm. " \
"These keywords can be used for various purposes such as content analysis and SEO.<br>" \
"Keyphrase Extraction: Keyphrase extraction is similar to keyword extraction but focuses on identifying multi-word phrases or expressions that are significant in the text using the Rake algorithm."
radio_selection = st.radio("Choose type of summarizer:", radio_options, help=help_text)
# Middle column: Text input and File uploader
with col2:
user_input = st.text_area("Enter your text here:")
uploaded_file = st.file_uploader("Upload a PDF", type="pdf")
if st.button("Summarize"):
if uploaded_file and user_input:
st.warning("Please provide either text input or a PDF file, not both.")
return
elif uploaded_file:
# Extract text from PDF
file_content = load_pdf(uploaded_file)
st.write("PDF uploaded successfully.")
elif user_input:
file_content = user_input
else:
st.warning("Please upload a PDF or enter some text to summarize.")
return
# Perform extractive summarization
if radio_selection == "Extractive (TextRank)":
summary = summarize_with_textrank(file_content)
st.session_state.summary = summary
# Perform extractive summarization
if radio_selection == "Extractive (Latent Semantic Analysis)":
summary = summarize_with_lsa(file_content)
st.session_state.summary = summary
# Perform extractive summarization
if radio_selection == "Abstractive (Fine-tuned BART)":
summary = summarize_with_bart(file_content)
st.session_state.summary = summary
# Perform extractive summarization
if radio_selection == "Abstractive (BART-large-CNN)":
summary = summarize_with_bart(file_content)
st.session_state.summary = summary
# Perform Keyword Extraction
if radio_selection == "Keyword Extraction (RAKE)":
summary = extract_keywords(file_content)
st.session_state.summary = summary
# Perform Keyphrase Extraction
if radio_selection == "Keyphrase Extraction (RAKE)":
summary = extract_sentences_with_obligations(file_content)
st.session_state.summary = summary
# Right column: Displaying text after pressing 'Summarize'
with col3:
st.write("Summary:")
if 'summary' in st.session_state:
st.write(st.session_state.summary)
if __name__ == "__main__":
main()