Create app.py

app.py

@@ -0,0 +1,119 @@
+import os
+import requests
+import streamlit as st
+from io import BytesIO
+from PyPDF2 import PdfReader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.embeddings import HuggingFaceEmbeddings
+from langchain.vectorstores import FAISS
+from transformers import pipeline
+import torch
+
+# Set up the page configuration as the first Streamlit command
+st.set_page_config(page_title="RAG-based PDF Chat", layout="centered", page_icon="📄")
+
+# Load the summarization pipeline model
+@st.cache_resource
+def load_summarization_pipeline():
+    summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
+    return summarizer
+
+summarizer = load_summarization_pipeline()
+
+# Dictionary of Hugging Face PDF URLs grouped by folders
+PDF_FOLDERS = {
+    "PPC and Administration": [
+        "https://huggingface.co/spaces/tahirsher/GenAI_Lawyers_Guide/tree/main/PPC%20and%20Administration",
+    ],
+    "IHC": [
+        "https://huggingface.co/spaces/tahirsher/GenAI_Lawyers_Guide/tree/main/IHC"
+    "LHC": [
+        "https://huggingface.co/spaces/tahirsher/GenAI_Lawyers_Guide/tree/main/LHC"
+    "Lahore High Court Rules and Orders": [
+        "https://huggingface.co/spaces/tahirsher/GenAI_Lawyers_Guide/tree/main/Lahore%20High%20Court%20Rules%20and%20Orders"
+    "PHC": [
+        "https://huggingface.co/spaces/tahirsher/GenAI_Lawyers_Guide/tree/main/PHC"
+    "SC": [
+        "https://huggingface.co/spaces/tahirsher/GenAI_Lawyers_Guide/tree/main/SC"
+    ],
+}
+
+# Helper function to convert Hugging Face blob URLs to direct download URLs
+def get_huggingface_raw_url(url):
+    if "huggingface.co" in url and "/blob/" in url:
+        return url.replace("/blob/", "/resolve/")
+    return url
+
+# Fetch and extract text from all PDFs in specified folders
+def fetch_pdf_text_from_folders(pdf_folders):
+    all_text = ""
+    for folder_name, urls in pdf_folders.items():
+        folder_text = f"\n[Folder: {folder_name}]\n"
+        for url in urls:
+            raw_url = get_huggingface_raw_url(url)
+            response = requests.get(raw_url)
+            if response.status_code == 200:
+                pdf_file = BytesIO(response.content)
+                try:
+                    pdf_reader = PdfReader(pdf_file)
+                    for page in pdf_reader.pages:
+                        page_text = page.extract_text()
+                        if page_text:
+                            folder_text += page_text
+                except Exception as e:
+                    st.error(f"Failed to read PDF from URL {url}: {e}")
+            else:
+                st.error(f"Failed to fetch PDF from URL: {url}")
+        all_text += folder_text
+    return all_text
+
+# Split text into manageable chunks
+@st.cache_data
+def get_text_chunks(text):
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=10000, chunk_overlap=1000)
+    chunks = text_splitter.split_text(text)
+    return chunks
+
+# Initialize embedding function
+embedding_function = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
+
+# Create a FAISS vector store with embeddings
+@st.cache_resource
+def load_or_create_vector_store(text_chunks):
+    vector_store = FAISS.from_texts(text_chunks, embedding=embedding_function)
+    return vector_store
+
+# Generate summary based on the retrieved text
+def generate_summary_with_huggingface(query, retrieved_text):
+    summarization_input = f"{query}\n\nRelated information:\n{retrieved_text}"
+    max_input_length = 1024
+    summarization_input = summarization_input[:max_input_length]
+    summary = summarizer(summarization_input, max_length=500, min_length=50, do_sample=False)
+    return summary[0]["summary_text"]
+
+# Generate response for user query
+def user_input(user_question, vector_store):
+    docs = vector_store.similarity_search(user_question)
+    context_text = " ".join([doc.page_content for doc in docs])
+    return generate_summary_with_huggingface(user_question, context_text)
+
+# Main function to run the Streamlit app
+def main():
+    st.title("📄 Gen AI Lawyers Guide")
+    raw_text = fetch_pdf_text_from_folders(PDF_FOLDERS)
+    text_chunks = get_text_chunks(raw_text)
+    vector_store = load_or_create_vector_store(text_chunks)
+
+    user_question = st.text_input("Ask a Question:", placeholder="Type your question here...")
+
+    if st.button("Get Response"):
+        if not user_question:
+            st.warning("Please enter a question before submitting.")
+        else:
+            with st.spinner("Generating response..."):
+                answer = user_input(user_question, vector_store)
+                st.markdown(f"**🤖 AI:** {answer}")
+
+if __name__ == "__main__":
+    main()
+