optimized code

app.py

@@ -1,27 +1,16 @@
 import os
-import openai
 from langchain_community.embeddings import OpenAIEmbeddings
 from langchain_community.vectorstores import FAISS
-from langchain_community.llms import OpenAI
-from langchain.chains import ConversationChain
-from langchain_community.document_loaders import PyPDFLoader
-from langchain.memory import ConversationBufferMemory
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain_openai import ChatOpenAI
 from langchain.memory import ConversationSummaryMemory
-
 import gradio as gr
 from PyPDF2 import PdfReader
 from langchain.agents import initialize_agent, Tool
-from langchain.schema import HumanMessage, AIMessage
 from langchain_core.exceptions import OutputParserException
 
 apiKey = os.getenv("OPENAI_API_KEY")
 
-# 
-# apiKey = open("key.txt", "r").readline().strip('\n') 
-
-
 
 # Load PDF
 def read_pdf(file_paths):
@@ -32,40 +21,33 @@ def read_pdf(file_paths):
             text = ""
             for page in reader.pages:
                 text += page.extract_text()
-            combined_text += text + "\n\n"  # Add a newline for separation between files
+            combined_text += text + "\n\n"  
     return combined_text
 
-# Load legal document (Constitution of India)
 pdf_file_path = ["property_law.pdf","ipc.pdf","constitution_of_india.pdf"]
 document_text = read_pdf(pdf_file_path)
 
-# Split the text into smaller chunks (e.g., 1,000 characters each)
 text_splitter = RecursiveCharacterTextSplitter(
-    chunk_size=1000,  # Adjust this size based on your needs
-    chunk_overlap=100  # To keep some overlap between chunks for better context
+    chunk_size=1000,  
+    chunk_overlap=100 
 )
 
-# Split the document text
 chunks = text_splitter.split_text(document_text)
 
-# Initialize embeddings and FAISS vector store
 embeddings = OpenAIEmbeddings(openai_api_key=apiKey)
 vector_db = FAISS.from_texts(chunks, embeddings)
 
 
 
 
-# Function to retrieve relevant content from vector DB
 def retrieve_from_db(query):
     results = vector_db.similarity_search(query, k=1)
     return results[0].page_content
 
 
-# Initialize OpenAI LLM
 llm = ChatOpenAI(openai_api_key=apiKey)
 
 
-# Define agent tools
 tools = [
     Tool(
         name="DocumentRetriever",
@@ -75,7 +57,6 @@ tools = [
 ]
 
 
-# Initialize memory and agent
 memory = ConversationSummaryMemory(llm=llm)
 agent = initialize_agent(
     tools=tools,
@@ -86,41 +67,32 @@ agent = initialize_agent(
 )
 
 
-# Function to interact with the agent and store conversation
 def chatbot(input_text, chat_history):
     try:    
         
-        # Run the agent with the input and memory history
         response = agent.run(input_text)
 
-        # Check if the response is "N/A" and replace with a custom message
         if response == "N/A":
             response = "Sorry, I couldn't understand your question. Please ask a specific question regarding IPC, Transfer of Property and Constitution of India."
         
-        # Store the assistant's response in memory
         memory.save_context({"user": input_text}, {"assistant": response})
 
-        # Update chat history with the new response
         chat_history.append([input_text, response])
     
         
         return chat_history
     
     except OutputParserException as e:
-        # Handle the exception and notify the user
         error_message = "Sorry, I couldn't understand your question. Please ask a specific question regarding IPC, Transfer of Property and Constitution of India."
-        # Append the error message to chat history
         chat_history.append([error_message, input_text])
         print("Error:", str(e))
 
         return chat_history
 
-# Gradio UI
 def gradio_interface():
     with gr.Blocks() as demo:
         gr.Markdown("# Legal Query Chatbot")
 
-         # Create chat UI with custom class
         with gr.Column():
             chatbot_ui = gr.Chatbot()
             user_input = gr.Textbox(show_label=True, placeholder="Enter your INDIAN PENAL CODE, TRANSFER OF PROPERTY, CONSTITUTION OF INDIA query here...")
@@ -131,7 +103,6 @@ def gradio_interface():
     return demo
 
 
-# Run the app
 app = gradio_interface()
 
 if __name__ == "__main__":