app.py

from huggingface_hub import InferenceClient
import gradio as gr
from datetime import datetime
from textblob import TextBlob  # for typo correction

API_URL = "https://api-inference.huggingface.co/models/"

# Initialize the InferenceClient
client = InferenceClient("mistralai/Mistral-7B-Instruct-v0.1")

def format_prompt(message, history):
    """Format the prompt for the text generation model."""
    prompt = "<s>"
    for user_prompt, bot_response in history:
        prompt += f"[INST] {user_prompt} [/INST]"
        prompt += f" {bot_response}</s> "
    prompt += f"[INST] {message} [/INST]"
    return prompt

def correct_typos(text):
    """Correct typos in the text using TextBlob."""
    corrected_text = str(TextBlob(text).correct())
    return corrected_text

def generate(prompt, history):
    """Generate a response using the text generation model."""
    # Correct typos in the prompt
    prompt = correct_typos(prompt)

    # Check if the prompt is asking who created the bot
    if "who created you" in prompt.lower():
        return "I was created by Aniket Kumar and many more."
    # Handle small talk
    elif "how are you" in prompt.lower():
        return "I'm an AI and don't have feelings, but I'm here to help you. How can I assist you today?"

    # Set up parameters for text generation
    generate_kwargs = dict(
        temperature=0.9,
        max_new_tokens=512,
        top_p=0.95,
        repetition_penalty=1.0,
        do_sample=True,
    )

    # Format the prompt
    formatted_prompt = format_prompt(prompt, history)

    # Generate the response
    stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
    output = ""
    for response in stream:
        output += response.token.text
        yield output
    return output

def create_interface():
    """Create the Gradio interface."""
    customCSS = """
    #component-7 { # this is the default element ID of the chat component
      height: 800px; # adjust the height as needed
      flex-grow: 1;
    }
    """

    with gr.Blocks(css=customCSS) as demo:
        gr.ChatInterface(
            generate,
        )

    demo.queue().launch(debug=True)

# Run the application
create_interface()