Update app.py

app.py

@@ -44,22 +44,27 @@ def main():
     
     
     docs = split_docs(documents, 350, 40)
+    vectorstore = get_vectorstore(docs)
+    if "conversation_chain" not in st.session_state:
+        st.session_state.conversation_chain = create_conversational_rag_chain(vectorstore)
     
 
     if prompt := st.text_input("Enter your question:"):
         msgs = st.session_state.get("chat_history", StreamlitChatMessageHistory(key="special_app_key"))
         st.chat_message("human").write(prompt)
 
-        input_dict = {"input": prompt, "chat_history": msgs.messages}
-        config = {"configurable": {"session_id": "any"}}
-
-        response = st.session_state.conversation_chain.invoke(input_dict, config)
-        st.chat_message("ai").write(response["answer"])
+        if st.session_state.conversation_chain is not None:
+            input_dict = {"input": prompt, "chat_history": msgs.messages}
+            config = {"configurable": {"session_id": "any"}}
+            response = st.session_state.conversation_chain.invoke(input_dict, config)
+            st.chat_message("ai").write(response["answer"])
 
-        if "docs" in response and response["documents"]:
-            for index, doc in enumerate(response["documents"]):
-                with st.expander(f"Document {index + 1}"):
-                    st.write(doc)
+            if "docs" in response and response["documents"]:
+                for index, doc in enumerate(response["documents"]):
+                    with st.expander(f"Document {index + 1}"):
+                        st.write(doc)
+        else:
+            st.error("Conversation chain is not available.")
 
         st.session_state["chat_history"] = msgs
 
@@ -112,5 +117,19 @@ def create_conversational_rag_chain(vectorstore):
     )
     return conversation_chain
 
+
+def get_vectorstore(text_chunks):
+    model_name = "sentence-transformers/all-mpnet-base-v2"
+    model_kwargs = {'device': 'cpu'}
+    encode_kwargs = {'normalize_embeddings': True}
+    embeddings = HuggingFaceEmbeddings(
+        model_name=model_name,
+        model_kwargs=model_kwargs,
+        encode_kwargs=encode_kwargs
+    )
+    vectorstore = Chroma.from_documents(
+        documents=text_chunks, embedding=embeddings, persist_directory="docs/chroma/")
+    return vectorstore     
+
 if __name__ == "__main__":
     main()