add hybrid summarization

app.py

@@ -1,26 +1,26 @@
 import streamlit as st
 import re
+import PyPDF2
+import matplotlib.pyplot as plt
+import io
+from wordcloud import WordCloud
+from PIL import Image
+
 from rouge import Rouge
 from datasets import load_dataset
-import PyPDF2
 from extractive_summarization import summarize_with_textrank, summarize_with_lsa
 from abstractive_summarization import summarize_with_bart_cnn, summarize_with_bart_ft, summarize_with_led, summarize_with_t5
 from keyword_extraction import extract_keywords
 from keyphrase_extraction import extract_sentences_with_obligations
-from wordcloud import WordCloud
-import matplotlib.pyplot as plt
-from PIL import Image
-import io
-#from blanc import BlancHelp
+from hybrid_summarization import summarize_hybrid
 
-
-# Load in ToS
+#-------------------------------------------------------------------#
+# Load in ToS-Summaries dataset
 dataset = load_dataset("EE21/ToS-Summaries")
 
 # Extract titles or identifiers for the ToS
 tos_titles = [f"Document {i}" for i in range(len(dataset['train']))]
-
-        
+    
 # Set page to wide mode
 st.set_page_config(layout="wide")
 
@@ -41,7 +41,7 @@ def main():
 
     # Left column: Radio buttons for summarizer choice
     with col1:
-        radio_options = ["Abstractive (LongT5)", "Abstractive (LED)", 'Abstractive (BART Fine-tuned)', "Abstractive (BART-large-CNN)", 'Extractive (TextRank)', 
+        radio_options = ["Hybrid (RAKE + BART Fine-tuned)", "Abstractive (LongT5)", "Abstractive (LED)", 'Abstractive (BART Fine-tuned)', "Abstractive (BART-large-CNN)", 'Extractive (TextRank)', 
                          "Extractive (Latent Semantic Analysis)", 'Keyphrase Extraction (RAKE)', 'Keyword Extraction (RAKE)']
         
         help_text = "Abstractive: Abstractive summarization generates a summary that may contain words not present in the original text. " \
@@ -77,6 +77,11 @@ def main():
                 st.warning("Please upload a PDF, enter some text, or select a document to summarize.")
                 return
 
+            # Perform hybrid summarization
+            if radio_selection == "Hybrid (RAKE + BART Fine-tuned)":
+                summary = summarize_hybrid(file_content)
+                st.session_state.summary = summary
+
             # Perform extractive summarization
             if radio_selection == "Extractive (TextRank)":
                 summary = summarize_with_textrank(file_content)
@@ -87,22 +92,22 @@ def main():
                 summary = summarize_with_lsa(file_content)
                 st.session_state.summary = summary
 
-            # Perform extractive summarization
+            # Perform abstractive summarization
             if radio_selection == "Abstractive (BART Fine-tuned)":
                 summary = summarize_with_bart_ft(file_content)
                 st.session_state.summary = summary
 
-            # Perform extractive summarization
+            # Perform abstractive summarization
             if radio_selection == "Abstractive (BART-large-CNN)":
                 summary = summarize_with_bart_cnn(file_content)
                 st.session_state.summary = summary
 
-            # Perform extractive summarization
+            # Perform abstractive summarization
             if radio_selection == "Abstractive (LongT5)":
                 summary = summarize_with_t5(file_content)
                 st.session_state.summary = summary
 
-            # Perform extractive summarization
+            # Perform abstractive summarization
             if radio_selection == "Abstractive (LED)":
                 summary = summarize_with_led(file_content)
                 st.session_state.summary = summary

hybrid_summarization.py

@@ -0,0 +1,6 @@
+from keyphrase_extraction import extract_sentences_with_obligations
+from abstractive_summarization import summarize_with_bart_ft
+
+def summarize_hybrid(text):
+    extract = extract_sentences_with_obligations(text)
+    return summarize_with_bart_ft(extract)