Create app.py

app.py

@@ -0,0 +1,101 @@
+import os
+import streamlit as st
+import faiss
+import openai
+from PyPDF2 import PdfReader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.vectorstores import FAISS
+from langchain.prompts import PromptTemplate
+from langchain.chains import RetrievalQA
+from langchain.document_loaders import GoogleDriveLoader
+
+# Set OpenAI API key
+openai.api_key = os.getenv("OPENAI_API_KEY")
+
+# Google Drive loader setup
+def load_documents_from_drive(drive_folder_id):
+    loader = GoogleDriveLoader(folder_id=drive_folder_id)
+    return loader.load()
+
+# Helper function to process documents into chunks
+def process_documents(documents):
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
+    chunks = []
+    for doc in documents:
+        reader = PdfReader(doc.file_path)
+        text = "".join([page.extract_text() for page in reader.pages])
+        chunks.extend(text_splitter.split_text(text))
+    return chunks
+
+# Function to build FAISS index
+def build_faiss_index(chunks):
+    embeddings = OpenAIEmbeddings()
+    vectorstore = FAISS.from_texts(chunks, embeddings)
+    return vectorstore
+
+# Streamlit app setup
+def main():
+    st.title("Legal Document Assistance")
+    st.sidebar.title("Settings")
+
+    # Input for Google Drive folder ID
+    drive_folder_id = st.sidebar.text_input("Google Drive Folder ID", "")
+
+    # Initialize FAISS index
+    if st.sidebar.button("Load and Process Documents"):
+        st.write("Loading documents...")
+        try:
+            documents = load_documents_from_drive(drive_folder_id)
+            st.write(f"Loaded {len(documents)} documents.")
+
+            chunks = process_documents(documents)
+            st.write(f"Processed into {len(chunks)} chunks.")
+
+            vectorstore = build_faiss_index(chunks)
+            st.session_state.vectorstore = vectorstore
+
+            st.write("FAISS index built successfully!")
+        except Exception as e:
+            st.error(f"Error: {str(e)}")
+
+    # User query input
+    query = st.text_input("Enter your legal query:")
+
+    if query and "vectorstore" in st.session_state:
+        vectorstore = st.session_state.vectorstore
+        retriever = vectorstore.as_retriever()
+
+        prompt_template = PromptTemplate(
+            input_variables=["context", "question"],
+            template="You are a legal assistant. Given the context: {context}, answer the question: {question} succinctly.",
+        )
+
+        qa_chain = RetrievalQA(retriever=retriever, prompt_template=prompt_template)
+        response = qa_chain.run(query)
+
+        st.write("Generated Response:")
+        st.write(response)
+
+        # Generate and display downloadable PDF
+        if st.button("Generate PDF"):
+            from fpdf import FPDF
+
+            pdf = FPDF()
+            pdf.add_page()
+            pdf.set_font("Arial", size=12)
+            pdf.multi_cell(0, 10, f"Query: {query}\n\nResponse: {response}")
+
+            pdf_file_path = "response.pdf"
+            pdf.output(pdf_file_path)
+
+            with open(pdf_file_path, "rb") as f:
+                st.download_button(
+                    label="Download PDF",
+                    data=f,
+                    file_name="response.pdf",
+                    mime="application/pdf",
+                )
+
+if __name__ == "__main__":
+    main()