Update app.py

app.py

@@ -133,99 +133,4 @@
 # if __name__ == "__main__":
 #     main()
 
-from sentence_transformers import SentenceTransformer
-from transformers import pipeline
-import faiss
-import numpy as np
-import streamlit as st
-import os
-from dotenv import load_dotenv
-import google.generativeai as genai
-import torch
-
-# Set device to CPU
-device = "cpu"
-
-# Load models once
-embedding_model = SentenceTransformer("all-MiniLM-L6-v2-int8", device=device)  # Removed normalize_embeddings
-summarizer = pipeline("summarization", model="facebook/bart-large-cnn", device=-1)
-
-# Load API Key
-load_dotenv()
-api_key = os.getenv("API_KEY")
-genai.configure(api_key=api_key)
-gemini_model = genai.GenerativeModel(model_name="gemini-2.0-flash")
-
-# Cache FAISS index & document texts
-@st.cache_resource
-def load_faiss_index():
-    if not os.path.exists("faiss_index.idx") or not os.path.exists("doc_texts.npy"):
-        return None, None
-    index = faiss.read_index("faiss_index.idx")
-    doc_texts = np.load("doc_texts.npy", allow_pickle=True)
-    return index, doc_texts
-
-faiss_index, doc_texts = load_faiss_index()
-
-# Query Legal Documents
-def query_legal_documents(query: str, top_k=3):
-    if faiss_index is None or doc_texts is None:
-        st.error("FAISS index or document data not found.")
-        return []
-
-    query_embedding = embedding_model.encode([query])
-    faiss.normalize_L2(query_embedding)  # Normalize embeddings manually
-
-    distances, indices = faiss_index.search(query_embedding, top_k)
-
-    return [doc_texts[i] for i in indices[0] if i < len(doc_texts)]
-
-# Summarization Agent (Batch Processing)
-def summarize_text(text_sections):
-    texts = [section for section in text_sections]
-    summaries = summarizer(texts, max_length=100, min_length=30, do_sample=False)
-    return [summary["summary_text"] for summary in summaries]
-
-# LLM Agent (Skip if Summaries are Sufficient)
-def generate_llm_response(summary_text):
-    if len(summary_text) < 200:
-        return summary_text  # Skip LLM for short summaries
-    response = gemini_model.generate_content(summary_text)
-    return response.text
-
-# Streamlit App
-def main():
-    st.set_page_config(page_title="Legal Chatbot", layout="wide")
-    st.title("🧑‍⚖️ Legal Chatbot")
-    user_query = st.text_input("Enter your legal query:")
-
-    if st.button("Submit"):
-        if user_query:
-            st.info("Processing your request...")
-            retrieved_sections = query_legal_documents(user_query)
-
-            if not retrieved_sections:
-                st.warning("No relevant legal documents found.")
-                return
-
-            summarized_sections = summarize_text(retrieved_sections)
-            combined_summary = "\n".join(summarized_sections)
-            final_response = generate_llm_response(combined_summary)
-
-            st.markdown("### 📖 Retrieved Data")
-            for section in retrieved_sections:
-                st.markdown(f"🔹 {section}")
-
-            st.markdown("### ✨ Summarized Response")
-            for summary in summarized_sections:
-                st.markdown(f"✅ {summary}")
-
-            st.markdown("### 🤖 AI-Enhanced Response")
-            st.text_area("Final Answer:", final_response, height=150)
-
-if __name__ == "__main__":
-    main()
-
-
-