app.py

# from langchain_ollama import ChatOllama
from langchain_google_genai import ChatGoogleGenerativeAI
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_core.chat_history import InMemoryChatMessageHistory
from langchain_core.runnables.history import RunnableWithMessageHistory
from dotenv import load_dotenv

import uuid
import gradio as gr

load_dotenv()  # Load environment variables from .env file

store = {}

def get_session_history(session_id):
    """
    Function to get the session history.
    This function can be modified to retrieve history from a database or any other storage.
    """
    if( session_id not in store):
        store[session_id] = InMemoryChatMessageHistory()
    return store[session_id] 
    
#set up llama3.2:latest model with Ollama
# llm = ChatOllama(model="llama3.2:latest", temperature=2.0)
llm=ChatGoogleGenerativeAI(
    model="gemini-2.0-flash"


)

#Create a chat prompt template
prompt = ChatPromptTemplate.from_messages(
    messages=[
        # ("system", "You are a helpful assistant. Answer the user's questions to the best of your ability."),
        ("system", "You are a friendly and helpful assistant named Zenbot. Start by greeting the user and then answer their questions."),
        MessagesPlaceholder(variable_name="history"),
        ("human", "{input}")
    ]
)

#Combine the prompt and the model into a chain. 1st version
# chain = prompt | llm
# response = chain.invoke({"question": "How to install gradio?"})
# print(response.content)

#chain with message history
# This allows the model to remember previous interactions in the session
chain = RunnableWithMessageHistory(
    runnable= prompt | llm,
    get_session_history=get_session_history,
    input_messages_key="input",
    history_messages_key="history",
)


#Function to handle user input and generate a response. 1st version
# def chatbot(question):
#     if(question.strip() == ""):
#         return "Please enter a valid question." 

#     response = chain.invoke({"question": question})
#     return response.content


def chatbot(user_input,history_state,temprature_slider,  session_id=str(uuid.uuid4())):
    """
    Function to handle user input and generate a response.
    It maintains the session history and returns the response.
    """
    
    if(user_input.strip() == ""):
        return "Please enter a valid question."
        
    llm.temperature = temprature_slider  # Set the temperature for the model
    
    response = chain.invoke(
        {"input": user_input},
        config={"configurable": {"session_id": session_id}}
    ).content
    
    if history_state is None:
        history_state = []
    history_state.append((user_input, response))


    return str(store[session_id]) #, "Temprature : " + str(temprature_slider)

def clear(session_id):
    print("Clearing conversation history")
    store[session_id] = InMemoryChatMessageHistory()

#Create a Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("Zensar chatbot")
    history_state = gr.State(value=None)  # To keep track of the conversation history
    session_id = gr.State(value=str(uuid.uuid4()))  # Unique session ID for each user
    input_box = gr.Textbox(label="Ask a question", placeholder="Type your question here...")
    temprature_slider = gr.Slider(
        label="Temperature",
        minimum=0.0,
        maximum=1.0,
        value=0.5,
        step=0.1,
        interactive=True,
    )
    output_box = gr.Textbox(label="Answer", interactive=False)
    submit_button = gr.Button("Submit")
    clear_button = gr.ClearButton(components=[input_box, output_box, temprature_slider])
   

    submit_button.click(
        fn=chatbot, 
        inputs=[input_box,history_state,temprature_slider, session_id], 
        outputs=output_box
    )

    clear_button.click(
        fn=clear,
        inputs=[session_id],

    )

   

# Launch the Gradio app
demo.launch()