import gradio as gr
import spaces
from huggingface_hub import InferenceClient
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Svngoku/c4ai-command-r7b-12-2024-4bit")
model = AutoModelForCausalLM.from_pretrained("Svngoku/c4ai-command-r7b-12-2024-4bit")

"""
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
"""
client = InferenceClient("Svngoku/c4ai-command-r7b-12-2024-4bit")

def wrap_text_output(text):
    wrapped_text = f"<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>{text}<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>"
    return wrapped_text



@spaces.GPU
def generate_response(message, history, documents_text):
    conversation = history + [{"role": "user", "content": message}]
    documents = []
    for doc in documents_text.split('\n'):
        if doc.strip():
            try:
                heading, body = doc.split(':', 1)
                documents.append({"heading": heading.strip(), "body": body.strip()})
            except ValueError:
                print(f"Invalid document format: {doc}")

    input_prompt = tokenizer.apply_chat_template(
        conversation=conversation,
        documents=documents,
        tokenize=False,
        add_generation_prompt=True,
        return_tensors="pt",
    )
    input_ids = tokenizer(input_prompt, return_tensors="pt").input_ids
    input_ids = input_ids.to(model.device)
    gen_tokens = model.generate(
        input_ids, max_new_tokens=2048, do_sample=True, temperature=0.3
    )
    gen_text = tokenizer.decode(gen_tokens[0], skip_special_tokens=True)

    # Extract chatbot's response
    chatbot_response = gen_text.split("<|CHATBOT_TOKEN|>")[-1]

    return chatbot_response


"""
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
"""
demo = gr.ChatInterface(
    generate_response,
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
)


if __name__ == "__main__":
    demo.launch()