app.py

import streamlit as st
from openai import OpenAI
import os
import sys
from dotenv import load_dotenv, dotenv_values
load_dotenv()

# initialize the client
client = OpenAI(
  base_url="https://api-inference.huggingface.co/v1",
  api_key=os.environ.get('HUGGINGFACEHUB_API_TOKEN')  # Replace with your token
)

# Create supported models
model_links = {
    "Mixtral-8x7B-Instruct-v0.1": "mistralai/Mixtral-8x7B-Instruct-v0.1",
    "Aya-23-35B": "CohereForAI/aya-23-35B",
    "Mistral-Nemo-Instruct-2407": "mistralai/Mistral-Nemo-Instruct-2407",
    "Nous-Hermes-2-Mixtral-8x7B-DPO": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
    "Mistral-7B-Instruct-v0.1": "mistralai/Mistral-7B-Instruct-v0.1",
    "Mistral-7B-Instruct-v0.2": "mistralai/Mistral-7B-Instruct-v0.2",
    "Mistral-7B-Instruct-v0.3": "mistralai/Mistral-7B-Instruct-v0.3",
    "Mistral-Small-Instruct-2409": "mistralai/Mistral-Small-Instruct-2409",
    "EuroLLM-9B-Instruct": "utter-project/EuroLLM-9B-Instruct",
    "EuroLLM-9B": "utter-project/EuroLLM-9B",
    "Athene-V2-Chat": "Nexusflow/Athene-V2-Chat",
    
}

def reset_conversation():
    #st.session_state.conversation = []
    st.session_state.messages = []
    return None

def ask_assistant_stream(st_model, st_messages, st_temp_value, st_max_tokens):
    response={}
    try:
        stream = client.chat.completions.create(
            model=st_model,
            messages=[
                {"role": m["role"], "content": m["content"]}
                for m in st_messages
            ],
            temperature=st_temp_value,
            stream=True,
            max_tokens=st_max_tokens,
        )
        response["stream"] = stream
    
    except Exception as e:
        pass

    return response

# Define the available models & Create the sidebar with the dropdown for model selection
models =[key for key in model_links.keys()]
selected_model = st.sidebar.selectbox("Select Model", models)

# Create a temperature slider
temp_values = st.sidebar.slider('Select a temperature value', 0.0, 1.0, (0.5))

# Create a max_token slider
max_token_value = st.sidebar.slider('Select a max_token value', 1000, 9000, (5000))

#Add reset button to clear conversation
st.sidebar.button('Reset Chat', on_click=reset_conversation) #Reset button

# Create model description
st.sidebar.write(f"You're now chatting with **{selected_model}**")
st.sidebar.markdown("*Generated content may be inaccurate or false.*")

# Edit dialog for editing a message
@st.dialog("Edit Message")
def edit_message(position):
    returnText = st.text_area("message:", value = st.session_state.messages[position-1]["content"])
    
    if st.button("Save"):
        st.session_state.messages[position-1]["content"] = returnText
        st.rerun()
    if st.session_state.messages[position-1]["role"] == "user":
        if st.button("Save & Retry"):
            st.session_state.messages[position-1]["content"] = returnText
            del st.session_state.messages[position:]
            st.session_state.instant_request = True
            st.rerun()

def remove_message(position):
    st.toast("try to remove message no: " + str(position-1) + " and "+ str(position))
    del st.session_state.messages[position-2:position]

def ask_assistant_write_stream():
    # Display assistant response in chat message container
    assistant = ask_assistant_stream(model_links[selected_model], st.session_state.messages, temp_values, max_token_value)
    pos = len(st.session_state.messages)+1
    if "stream" in assistant: 
        with st.chat_message("assistant"):
            col1, col2 = st.columns([9,1])
            response = col1.write_stream(assistant["stream"])
            col2.button("", icon = ":material/edit:", key="button_edit_message_"+str(pos), args=[pos], on_click=edit_message)
            col2.button("", icon = ":material/delete:", key="button_remove_message_"+str(pos), args=[pos], on_click=remove_message)
    else:
        with st.chat_message("assistant"):
            col1, col2 = st.columns([9,1])
            response = col1.write("Failure!")
            col2.button("", icon = ":material/delete:", key="button_remove_message_"+str(pos), args=[pos], on_click=remove_message)

    st.session_state.messages.append({"role": "assistant", "content": response})
    


st.subheader(f'{selected_model}')

# Initialize chat history
if "messages" not in st.session_state:
    st.session_state.messages = []

# Display chat messages from history on app rerun
pos = 0
for message in st.session_state.messages:
    pos=pos+1
    with st.chat_message(message["role"]):
        col1, col2 = st.columns([9,1])
        col1.markdown(message["content"])
        col2.button("", icon = ":material/edit:", key="button_edit_message_"+str(pos), args=[pos], on_click=edit_message)
        if message["role"] == "assistant":
            col2.button("", icon = ":material/delete:", key="button_remove_message_"+str(pos), args=[pos], on_click=remove_message)

if "instant_request" not in st.session_state:
    st.session_state.instant_request = False


if st.session_state.instant_request:
    ask_assistant_write_stream()
    st.session_state.instant_request = False

# Accept user input
if prompt := st.chat_input(f"Hi I'm {selected_model}, ask me a question"):
    # Display user message in chat message container and Add user message to chat history
    pos = len(st.session_state.messages)+1
    with st.chat_message("user"):
        col1, col2 = st.columns([9,1])
        col1.markdown(prompt)
        col2.button("", icon = ":material/edit:", key="button_edit_message_"+str(pos), args=[pos], on_click=edit_message)
    st.session_state.messages.append({"role": "user", "content": prompt})
    
    # Display assistant response in chat message container
    ask_assistant_write_stream()