attempt summary

app.py

@@ -1,6 +1,6 @@
 import streamlit as st
 from haystack.document_stores import InMemoryDocumentStore
-from haystack.nodes import FARMReader, PreProcessor, PDFToTextConverter, TfidfRetriever
+from haystack.nodes import TransformersSummarizer, PreProcessor, PDFToTextConverter, TfidfRetriever
 from haystack.schema import Document
 import logging
 
@@ -14,6 +14,7 @@ preprocessor = PreProcessor(
     split_respect_sentence_boundary=True,
     split_overlap=3
 )
+summarizer = TransformersSummarizer(model_name_or_path="google/pegasus-xsum")
 
 
 def pdf_to_document_store(pdf_files):
@@ -27,6 +28,11 @@ def pdf_to_document_store(pdf_files):
     preprocessed_docs = preprocessor.process(documents)
     document_store.write_documents(preprocessed_docs)
     return None
+
+def summarize(files):
+    pdf_to_document_store(files)
+    summary = summarizer.predict(documents=document_store.get_all_documents(), generate_single_summary=True)
+    st.write(summary)
     
 uploaded_files = st.file_uploader("Choose PDF files", accept_multiple_files=True)
 
@@ -34,7 +40,7 @@ if uploaded_files is not None:
     st.write(len(uploaded_files))
     if st.button('Summarize Documents'):
         document_store.delete_documents()
-        pdf_to_document_store(uploaded_files)
+        summarize(uploaded_files)
 
 if st.button('Calculate num of docs'):
     st.write(document_store.get_document_count())