# import gradio as gr
# from groq import Groq

# client = Groq(
#     api_key=("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"),
# )

# def generate_response(input_text):
#     chat_completion = client.chat.completions.create(
#         messages=[
#             {
#                 "role": "user",
#                 "content": input_text,
#             }
#         ],
#         model="llama3-8b-8192",
#     )
#     return chat_completion.choices[0].message.content


# iface = gr.Interface(
#     fn=generate_response,
#     inputs=gr.Textbox(label="ورودی" , lines=2, placeholder="اینجا یه چی بپرس... "),
#     outputs=gr.Textbox(label="جواب"),
#     title="💬 Parviz GPT",
#     description="زنده باد",
#     theme="dark",
#     allow_flagging="never"

# )
# iface.launch()

import gradio as gr
from groq import Groq
import time

client = Groq(api_key="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")

def generate_response(message, chat_history):
    chat_completion = client.chat.completions.create(
        messages=[{"role": "user", "content": message}],
        model="llama3-8b-8192",
    )
    bot_message = chat_completion.choices[0].message.content
    
    for i in range(0, len(bot_message), 10):
        yield chat_history + [(message, bot_message[:i + 10])]
        time.sleep(0.1)
    
    yield chat_history + [(message, bot_message)]


with gr.Blocks() as demo:
    gr.Markdown("<h1 style='text-align: center;'>💬 Parviz GPT</h1><p style='text-align: center; color: #e0e0e0;'>زنده باد</p>")
    
    chatbot = gr.Chatbot(label="جواب")
    msg = gr.Textbox(label="ورودی", placeholder="اینجا یه چی بپرس... ", lines=1)
    
    msg.submit(generate_response, [msg, chatbot], chatbot)
    
    clear = gr.ClearButton([msg, chatbot])
demo.launch()













# import gradio as gr
# import torch
# from transformers import AutoTokenizer, AutoModelForCausalLM

# tokenizer = AutoTokenizer.from_pretrained("universitytehran/PersianMind-v1.0", use_fast=True)
# model = AutoModelForCausalLM.from_pretrained(
#     "universitytehran/PersianMind-v1.0",
#     torch_dtype=torch.bfloat16
# ).to("cpu")

# CONTEXT = (
#     "This is a conversation with ParvizGPT. It is an artificial intelligence model designed by Amir Mahdi Parviz, "
#     "an NLP expert, to help you with various tasks such as answering questions, "
#     "providing recommendations, and assisting with decision-making. Ask it anything!"
# )
# pretokenized_context = tokenizer(CONTEXT, return_tensors="pt").input_ids.to("cpu")

# def generate_response(message, chat_history):
#     prompt = torch.cat(
#         [pretokenized_context, tokenizer("\nYou: " + message + "\nParvizGPT: ", return_tensors="pt").input_ids.to("cpu")],
#         dim=1
#     )

#     with torch.no_grad():
#         outputs = model.generate(
#             prompt,
#             max_new_tokens=32,
#             temperature=0.6,
#             top_k=20,
#             top_p=0.8,
#             do_sample=True
#         )

#     result = tokenizer.decode(outputs[0], skip_special_tokens=True)
#     response = result.split("ParvizGPT:")[-1].strip()
#     return chat_history + [(message, response)]

# with gr.Blocks() as demo:
#     gr.Markdown("<h1 style='text-align: center;'>💬 Parviz GPT</h1>")
#     chatbot = gr.Chatbot(label="Response")
#     msg = gr.Textbox(label="Input", placeholder="Ask your question...", lines=1)
#     msg.submit(generate_response, [msg, chatbot], chatbot)
#     gr.ClearButton([msg, chatbot])

# demo.launch()