import gradio as gr
from huggingface_hub import InferenceClient
from transformers import AutoModelForCausalLM, pipeline

# Use a pipeline as a high-level helper
pipe = pipeline("visual-question-answering", model="dandelin/vilt-b32-finetuned-vqa", trust_remote_code=True)

# Load model directly
model = AutoModelForCausalLM.from_pretrained("microsoft/Florence-2-large", trust_remote_code=True)

"""
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("HuggingFaceH4/zephyr-7b-beta")

def respond(message, history, system_message, max_tokens, temperature, top_p):
    """
    Generates a response based on the user message and chat history.
    
    Args:
        message (str): The user message.
        history (list): A list of tuples containing user and assistant messages.
        system_message (str): The system message.
        max_tokens (int): Maximum number of tokens for the response.
        temperature (float): Temperature for the response generation.
        top_p (float): Top-p for nucleus sampling.
    
    Yields:
        str: The generated response.
    """
    messages = [{"role": "system", "content": system_message}]

    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

    messages.append({"role": "user", "content": message})

    response = ""

    for message in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token = message.choices[0].delta.content
        response += token
        yield response

def process_video(video):
    """
    Processes the uploaded video file.
    
    Args:
        video (gr.Video): The uploaded video file.
    
    Returns:
        str: Confirmation message for the uploaded video.
    """
    return f"Processing video: {video.name}"

def process_pdf(pdf):
    """
    Processes the uploaded PDF file.
    
    Args:
        pdf (gr.File): The uploaded PDF file.
    
    Returns:
        str: Confirmation message for the uploaded PDF.
    """
    return f"Processing PDF: {pdf.name}"

def process_image(image):
    """
    Processes the uploaded image file.
    
    Args:
        image (gr.Image): The uploaded image file.
    
    Returns:
        str: Confirmation message for the uploaded image.
    """
    return f"Processing image: {image.name}"

# Define upload interfaces
video_upload = gr.Interface(fn=process_video, inputs=gr.Video(), outputs="text", title="Upload a Video")
pdf_upload = gr.Interface(fn=process_pdf, inputs=gr.File(file_types=['.pdf']), outputs="text", title="Upload a PDF")
image_upload = gr.Interface(fn=process_image, inputs=gr.Image(), outputs="text", title="Upload an Image")

# Combine upload interfaces into tabs
tabbed_interface = gr.TabbedInterface([video_upload, pdf_upload, image_upload], ["Video", "PDF", "Image"])

# Main Gradio interface
demo = gr.Blocks()

with demo:
    with gr.Tab("Chat Interface"):
        gr.ChatInterface(
            respond,
            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)",
                ),
            ],
        )
    with gr.Tab("Upload Files"):
        tabbed_interface

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