import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline

# Load the model and tokenizer from Hugging Face
model_id = "seedboxai/KafkaLM-8x7B-German-V0.1-DPO"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, load_in_4bit=True, trust_remote_code=True)

# Define the text generation function
def generate_text(user_input, system_prompt):
    # Combine the system prompt and the user input to form the full prompt
    full_prompt = f"{system_prompt.strip()}\n\n{user_input.strip()}"
    
    # Initialize the pipeline for text generation
    text_generator = pipeline('text-generation', model=model, tokenizer=tokenizer,
                              return_full_text=True, temperature=0.5,
                              max_new_tokens=512, top_p=0.95, top_k=50, do_sample=True)
    
    # Generate text based on the full prompt
    results = text_generator(full_prompt)
    generated_text = results[0]['generated_text']
    
    return generated_text

# Setup the Gradio interface
iface = gr.Interface(fn=generate_text,
                     inputs=[gr.Textbox(lines=2, label="User Prompt"), gr.Textbox(lines=5, label="System Prompt")],
                     outputs=gr.Textbox(label="Generated Text"),
                     title="Text Generation with KafkaLM",
                     description="Enter a user prompt and a system prompt to generate text using the KafkaLM model.")

# Launch the Gradio app
if __name__ == "__main__":
    iface.launch()