# import gradio as gr
# from huggingface_hub import InferenceClient

# # Initialize the client with your desired model
# client = InferenceClient("Bhaskar2611/Capstone")

# # Define the system prompt as an AI Dermatologist
# def format_prompt(message, history):
#     prompt = "<s>"
#     # Start the conversation with a system message
#     prompt += "[INST] You are an AI Dermatologist chatbot designed to assist users with skin by only providing text and if user information is not provided related to skin then ask what they want to know related to skin.[/INST]"
#     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

# # Function to generate responses with the AI Dermatologist context
# def generate(
#     prompt, history, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0
# ):
#     temperature = float(temperature)
#     if temperature < 1e-2:
#         temperature = 1e-2
#     top_p = float(top_p)

#     generate_kwargs = dict(
#         temperature=temperature,
#         max_new_tokens=max_new_tokens,
#         top_p=top_p,
#         repetition_penalty=repetition_penalty,
#         do_sample=True,
#         seed=42,
#     )

#     formatted_prompt = format_prompt(prompt, history)

#     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

# # Customizable input controls for the chatbot interface
# Settings = [
#     gr.Slider(
#         label="Temperature",
#         value=0.9,
#         minimum=0.0,
#         maximum=1.0,
#         step=0.05,
#         interactive=True,
#         info="Higher values produce more diverse outputs",
#     ),
#     gr.Slider(
#         label="Max new tokens",
#         value=256,
#         minimum=0,
#         maximum=1048,
#         step=64,
#         interactive=True,
#         info="The maximum numbers of new tokens",
#     ),
#     gr.Slider(
#         label="Top-p (nucleus sampling)",
#         value=0.90,
#         minimum=0.0,
#         maximum=1,
#         step=0.05,
#         interactive=True,
#         info="Higher values sample more low-probability tokens",
#     ),
#     gr.Slider(
#         label="Repetition penalty",
#         value=1.2,
#         minimum=1.0,
#         maximum=2.0,
#         step=0.05,
#         interactive=True,
#         info="Penalize repeated tokens",
#     )
# ]
# # Define the chatbot interface with the starting system message as AI Dermatologist
# gr.ChatInterface(
#     fn=generate,
#     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, layout="panel"),
#     additional_inputs = Settings,
#     title="Skin Bot"
# ).launch(show_api=False)

# # Load your model after launching the interface
# # gr.load("models/Bhaskar2611/Capstone").launch()

# import gradio as gr
# from huggingface_hub import InferenceClient

# # Initialize the client with your Hugging Face token
# client = InferenceClient(
#     model="HuggingFaceH4/zephyr-7b-beta",
#     hf_token = os.getenv("HF_TOKEN")
# )

# # Define the system prompt as an AI Dermatologist
# def format_prompt(message, history):
#     prompt = "<s>"
#     prompt += "[INST] You are an AI Dermatologist chatbot designed to assist users with skin by only providing text and if user information is not provided related to skin then ask what they want to know related to skin.[/INST]"
#     for user_prompt, bot_response in history:
#         prompt += f"[INST] {user_prompt} [/INST] {bot_response}</s> "
#     prompt += f"[INST] {message} [/INST]"
#     return prompt

# # Function to generate responses
# def generate(prompt, history, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0):
#     temperature = float(temperature)
#     if temperature < 1e-2:
#         temperature = 1e-2
#     top_p = float(top_p)

#     generate_kwargs = dict(
#         temperature=temperature,
#         max_new_tokens=max_new_tokens,
#         top_p=top_p,
#         repetition_penalty=repetition_penalty,
#         do_sample=True,
#         seed=42,
#     )

#     formatted_prompt = format_prompt(prompt, history)

#     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

# # Sliders for customization
# Settings = [
#     gr.Slider(label="Temperature", value=0.9, minimum=0.0, maximum=1.0, step=0.05, interactive=True, info="Higher values produce more diverse outputs"),
#     gr.Slider(label="Max new tokens", value=256, minimum=0, maximum=1048, step=64, interactive=True, info="The maximum number of new tokens"),
#     gr.Slider(label="Top-p (nucleus sampling)", value=0.90, minimum=0.0, maximum=1, step=0.05, interactive=True, info="Higher values sample more low-probability tokens"),
#     gr.Slider(label="Repetition penalty", value=1.2, minimum=1.0, maximum=2.0, step=0.05, interactive=True, info="Penalize repeated tokens")
# ]

# # Chat interface
# gr.ChatInterface(
#     fn=generate,
#     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, layout="panel"),
#     additional_inputs=Settings,
#     title="Skin Bot"
# ).launch(show_api=False)

# # Load any additional models if needed
# # gr.load("models/Bhaskar2611/Capstone").launch()

# import os
# from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
# import gradio as gr

# # Load your Hugging Face token (if needed for private models or API limit increases)
# hf_token = os.environ.get("HF_TOKEN")

# # Model ID for Mistral 7B Instruct
# model_id = "mistralai/Mistral-7B-Instruct-v0.1"

# # Load tokenizer
# tokenizer = AutoTokenizer.from_pretrained(model_id, token=hf_token)

# # BitsAndBytesConfig for 4-bit quantization to reduce memory usage
# bnb_config = BitsAndBytesConfig(load_in_4bit=True)

# # Load model with quantization and device mapping
# model = AutoModelForCausalLM.from_pretrained(
#     model_id,
#     quantization_config=bnb_config,
#     device_map="auto",
#     token=hf_token
# )

# # Skin assistant system prompt
# SKIN_ASSISTANT_PROMPT = (
#     "You are a helpful assistant specialized in skin diseases and dermatology. "
#     "Provide accurate, concise, and helpful advice about skin conditions, symptoms, "
#     "treatments, and care. Always respond in a clear and empathetic way.\n\n"
# )

# def generate_response(user_input):
#     prompt = SKIN_ASSISTANT_PROMPT + user_input
#     inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
#     outputs = model.generate(
#         **inputs,
#         max_new_tokens=1024,
#         do_sample=True,
#         temperature=0.7,
#         top_p=0.95,
#         repetition_penalty=1.1
#     )
#     response = tokenizer.decode(outputs[0], skip_special_tokens=True)
#     return response.replace(SKIN_ASSISTANT_PROMPT, "").strip()

# # Gradio interface
# iface = gr.Interface(
#     fn=generate_response,
#     inputs=gr.Textbox(lines=3, placeholder="Ask about skin diseases..."),
#     outputs="text",
#     title="Skin Disease Assistant",
#     description="Ask any questions related to skin diseases and get expert-like responses."
# )

# if __name__ == "__main__":
#     iface.launch()

import os
import gradio as gr
from huggingface_hub import InferenceClient
from dotenv import load_dotenv

# Load Hugging Face API token
load_dotenv()
HF_TOKEN = os.getenv("HF_TOKEN")

# Initialize Hugging Face client
client = InferenceClient(
    model="mistralai/Mistral-7B-Instruct-v0.3",
    token=HF_TOKEN
)

# System prompt about Indian monuments
system_message = (
    "You are an AI Dermatologist chatbot designed to assist users with skin by only providing text "
    "and if user information is not provided related to skin then ask what they want to know related to skin."
)

# Streaming chatbot logic
def respond(message, history):
    # Prepare messages with system prompt
    messages = [{"role": "system", "content": system_message}]
    for msg in history:
        messages.append(msg)
    messages.append({"role": "user", "content": message})

    # Stream response from the model
    response = ""
    for chunk in client.chat.completions.create(
        model="mistralai/Mistral-7B-Instruct-v0.3",
        messages=messages,
        max_tokens=1024,
        temperature=0.7,
        top_p=0.95,
        stream=True,
    ):
        token = chunk.choices[0].delta.get("content", "") or ""
        response += token
        yield response

# Create Gradio interface
with gr.Blocks() as demo:
    chatbot = gr.Chatbot(type='messages')  # Use modern message format
    gr.ChatInterface(fn=respond, chatbot=chatbot, type="messages")  # Match format

# Launch app
if __name__ == "__main__":
    demo.launch()