adding doc summarizer

app.py

@@ -1,4 +1,46 @@
+import PyPDF2
 import streamlit as st
+from dotenv import load_dotenv
+from transformers import pipeline
 
-x = st.slider('Select a value')
-st.write(x, 'squared is', x * x)
+
+def retrieve_pdf_text(pdf_file):
+    pdf_reader = PyPDF2.PdfReader(pdf_file)
+    text = ""
+    for page in pdf_reader.pages:
+        text += page.extract_text()
+    return text
+
+
+def main():
+    load_dotenv()
+    st.set_page_config(page_title='Document Summarizer', page_icon=':books:')
+    st.header("Summarize a PDF")
+    hf_name = "pszemraj/led-base-book-summary"
+
+    pdf_file = st.file_uploader("Upload a PDF file with Annual Report", type=["pdf"])
+    length = st.slider('Max summary length', 0, 3000, 1000)
+
+    # if a pdf file is uploaded
+    if pdf_file:
+        raw_text = retrieve_pdf_text(pdf_file)
+        if st.button("Run"):
+            with st.spinner("Summarizing.."):
+                summarizer = pipeline("summarization", hf_name)
+                result = summarizer(
+                    raw_text,
+                    min_length=8,
+                    max_length=length,
+                    no_repeat_ngram_size=3,
+                    encoder_no_repeat_ngram_size=3,
+                    repetition_penalty=3.5,
+                    num_beams=4,
+                    do_sample=False,
+                    early_stopping=True,
+                )
+                st.write(result[0]["summary_text"])
+
+
+
+if __name__ == '__main__':
+    main()