app.py

import streamlit as st
from huggingface_hub import InferenceClient
import os
import pickle

st.title("Transcription Summarization")

# Base URL and API Key
API_KEY = os.environ.get('HUGGINGFACE_API_KEY')
BASE_URL = "https://api-inference.huggingface.co/models/"

model_links = {
    "LegacyLift🚀": BASE_URL + "Qwen/QwQ-32B-Preview",
    "ModernMigrate⭐": BASE_URL + "mistralai/Mixtral-8x7B-Instruct-v0.1",
    "RetroRecode🔄": BASE_URL + "microsoft/Phi-3-mini-4k-instruct"
}

model_info = {
    "LegacyLift🚀": {
        'description': "The LegacyLift model is a **Large Language Model (LLM)** for problem-solving, content writing, and daily tips.",
        'logo': './11.jpg'
    },
    "ModernMigrate⭐": {
        'description': "The ModernMigrate model excels in coding, logical reasoning, and high-speed inference.",
        'logo': './2.jpg'
    },
    "RetroRecode🔄": {
        'description': "The RetroRecode  is ideal for critical development, practical knowledge, and serverless inference.",
        'logo': './3.jpg'
    },
}

# Function Definitions
def format_prompt(message, conversation_history, custom_instructions=None):
    """Formats the input prompt."""
    prompt = f"\[INST\] {custom_instructions} \[/INST\]\n\[CONV_HISTORY\]\n"
    for role, content in conversation_history:
        prompt += f"{role.upper()}: {content}\n"
    prompt += f"\[/CONV_HISTORY\]\n\[INST\] {message} \[/INST\]\n\[RESPONSE\]\n"
    return prompt

def reset_conversation():
    """Resets the conversation."""
    st.session_state.conversation = []
    st.session_state.messages = []
    st.session_state.chat_state = "reset"

def load_conversation_history():
    """Loads conversation history from a file."""
    history_file = "conversation_history.pickle"
    return pickle.load(open(history_file, "rb")) if os.path.exists(history_file) else []

def save_conversation_history(conversation_history):
    """Saves conversation history to a file."""
    with open("conversation_history.pickle", "wb") as f:
        pickle.dump(conversation_history, f)

# Sidebar UI
models = list(model_links.keys())
selected_model = st.sidebar.selectbox("Select Model", models)
temp_values = st.sidebar.slider('Select Temperature', 0.0, 1.0, 0.5)
st.sidebar.button('Reset Chat', on_click=reset_conversation)

st.sidebar.write(f"Chatting with **{selected_model}**")
st.sidebar.markdown(model_info[selected_model]['description'])
st.sidebar.image(model_info[selected_model]['logo'])

# Load session state
if "prev_option" not in st.session_state:
    st.session_state.prev_option = selected_model

if st.session_state.prev_option != selected_model:
    st.session_state.messages = []
    st.session_state.prev_option = selected_model

if "chat_state" not in st.session_state:
    st.session_state.chat_state = "normal"

# Load conversation history
if "messages" not in st.session_state:
    st.session_state.messages = load_conversation_history()

# Main Chat
repo_id = model_links[selected_model]
st.subheader(f"{selected_model}")

if st.session_state.chat_state == "normal":
    for message in st.session_state.messages:
        with st.chat_message(message["role"]):
            st.markdown(message["content"])

    if prompt := st.chat_input(f"Hi I'm {selected_model}, How can I help you today?"):
        custom_instruction = (
            "Analyze this transcript with precision. Remove commas in claim numbers, preserve exact numbers and "
            "dates (dd/mm/yy). Extract claim numbers as single entities."
            """1. Only include information explicitly stated

            2. Mark unclear information as "UNCLEAR"
            
            3. Preserve exact numbers, dates (in dd/mm/yy format), and claims
            
            4. Focus on factual content
            
            IMPORTANT REQUIREMENTS:
            
            - Format all dates as dd/mm/yy
            
            - Extract and list all claim numbers mentioned
            
            - Maintain exact numbers and statistics as stated
            
            - Do not make assumptions about unclear information
            
            Please analyze the following transcript and structure your response as follows:
            
            PARTICIPANTS:
            
            - List all participants and their roles (if mentioned)
            
            CONTEXT:
            
            - Meeting purpose
            
            - Duration (if mentioned)
            
            - Meeting date/time
            
            KEY POINTS:
            
            - Main topics discussed
            
            - Decisions made
            
            - Important numbers/metrics mentioned
            
            - Claims discussed
            
            ACTION ITEMS:
            
            - Specific tasks assigned
            
            - Who is responsible
            
            - Deadlines (in dd/mm/yy format)
            
            FOLLOW UP:
            
            - Scheduled next meetings
            
            - Pending items
            
            - Required approvals or confirmations"""


        )
        conversation_history = [(msg["role"], msg["content"]) for msg in st.session_state.messages]
        formatted_text = format_prompt(prompt, conversation_history, custom_instruction)

        with st.chat_message("user"):
            st.markdown(prompt)

        st.session_state.messages.append({"role": "user", "content": prompt})

        with st.chat_message("assistant"):
            try:
                client = InferenceClient(model=repo_id)
                response = client.text_generation(
                    formatted_text,
                    temperature=temp_values,
                    max_new_tokens=1024,
                    stream=True
                )
                response_text = ''.join(response)  # Collect and concatenate the response
                response_text = response_text.replace(",", "")  # Remove commas in claim numbers
                st.markdown(response_text)
                st.session_state.messages.append({"role": "assistant", "content": response_text})
                save_conversation_history(st.session_state.messages)
            except Exception as e:
                st.error(f"An error occurred: {e}")

elif st.session_state.chat_state == "reset":
    st.session_state.chat_state = "normal"
    st.experimental_rerun()