Update app.py

app.py

@@ -1,11 +1,11 @@
 import gradio as gr
-from langchain.chains import RetrievalQA
+from langchain.chains import RetrievalQAWithSourcesChain  # Import RetrievalQAWithSourcesChain instead of RetrievalQA
 from langchain.llms import HuggingFaceHub  # Import HuggingFaceHub for Jais
-from langchain.vectorstores import Chroma
 from langchain.embeddings import HuggingFaceEmbeddings
 import os
 import preprocess  # Import the preprocess module
 import create_database  # Import the create_database module
+from langchain_chroma import Chroma  # Import Chroma from langchain_chroma
 
 
 # --- Preprocessing and Database Creation ---
@@ -22,9 +22,7 @@ embedding_model = HuggingFaceEmbeddings(model_name="sentence-transformers/all-Mi
 vector_db = Chroma(persist_directory="db", embedding_function=embedding_model)
 retriever = vector_db.as_retriever(search_kwargs={"k": 3})
 
-
 # Load Jais-13B from Hugging Face
-
 def initialize_llm():
     """Initializes the Hugging Face LLM using the HUGGINGFACEHUB_API_TOKEN environment variable."""
     huggingfacehub_api_token = os.environ.get("HUGGINGFACEHUB_API_TOKEN")
@@ -41,12 +39,15 @@ def initialize_llm():
 
     
 # Create the RetrievalQA chain
-qa_chain = RetrievalQA(llm=initialize_llm, retriever=retriever)
+# chain_type="stuff" is required in LangChain 0.0.200 or later
+qa_chain = RetrievalQAWithSourcesChain.from_chain_type(llm=initialize_llm, chain_type="stuff", retriever=retriever) 
 
 # --- Gradio Interface ---
 
 def chatbot_interface(question):
-    return qa_chain.run(question)
+    result = qa_chain({"question": question})
+    # Only return the answer, not the sources
+    return result['answer'] 
 
 iface = gr.Interface(
     fn=chatbot_interface,