app.py

import os
import gradio as gr
from openai import OpenAI
from typing import List, Tuple

CLIENTS = {
    "perplexity":{"key":os.getenv('PX_KEY'),"endpoint":"https://api.perplexity.ai"},
    "hyperbolic":{"key":os.getenv('HYPERBOLIC_XYZ_KEY'),"endpoint":"https://api.hyperbolic.xyz/v1"},
    "huggingface":{"key":os.getenv('HF_KEY'),"endpoint":"https://huggingface.co/api/inference-proxy/together"},
}
for client_type in CLIENTS:
    CLIENTS[client_type]["client"] = OpenAI(base_url=CLIENTS[client_type]["endpoint"], api_key=CLIENTS[client_type]["key"])

PASSWORD = os.getenv("PASSWD")

# Define available models
AVAILABLE_MODELS = {
    "DeepSeek V3 (Hyperbolic.xyz)": {"model_name":"deepseek-ai/DeepSeek-V3","type":"hyperbolic"},
    "DeepSeek V3 (HuggingFace.co)": {"model_name":"deepseek-ai/DeepSeek-V3","type":"huggingface"},
    "Llama3.3-70b-Instruct": {"model_name":"meta-llama/Llama-3.3-70B-Instruct","type":"hyperbolic"},
    "Llama3.1-8b-Instruct": {"model_name":"meta-llama/Meta-Llama-3.1-8B-Instruct","type":"hyperbolic"},
    "Sonar Pro": {"model_name":"sonar-pro","type":"perplexity"},
    "Sonar": {"model_name":"sonar","type":"perplexity"},
}

def respond(
    message: str,
    history: List[Tuple[str, str]],
    system_message: str,
    model_choice: str,
    max_tokens: int,
    temperature: float,
    top_p: float,
):
    """Handles chatbot responses with Perplexity AI."""
    
    if model_choice not in AVAILABLE_MODELS:
        return "Error: Invalid model selection."

    messages = [{"role": "system", "content": system_message}]
    for user_msg, assistant_msg in history:
        if user_msg:
            messages.append({"role": "user", "content": user_msg})
        if assistant_msg:
            messages.append({"role": "assistant", "content": assistant_msg})
    messages.append({"role": "user", "content": message})

    response = ""
    citations = []

    selected_client = CLIENTS[AVAILABLE_MODELS[model_choice]["type"]]["client"]
    
    try:
        stream = selected_client.chat.completions.create(
            model=AVAILABLE_MODELS[model_choice]["model_name"],
            messages=messages,
            max_tokens=max_tokens,
            temperature=temperature,
            top_p=top_p,
            stream=True,
        )

        for chunk in stream:
            if hasattr(chunk, "choices") and chunk.choices:
                token = chunk.choices[0].delta.content or ""
                response += token
                yield response  # Stream response as it arrives
            if hasattr(chunk, "citations") and chunk.citations:
                citations = chunk.citations

        # Append citations as clickable links
        if citations:
            citation_text = "\n\nSources:\n" + "\n".join(
                [f"[{i+1}] [{url}]({url})" for i, url in enumerate(citations)]
            )
            response += citation_text
            yield response

    except Exception as e:
        yield f"Error: {str(e)}"

def check_password(input_password):
    """Validates the password before showing the chat interface."""
    if input_password == PASSWORD:
        return gr.update(visible=False), gr.update(visible=True)
    else:
        return gr.update(value="", interactive=True), gr.update(visible=False)

with gr.Blocks() as demo:
    with gr.Column():
        password_input = gr.Textbox(
            type="password", label="Enter Password", interactive=True
        )
        submit_button = gr.Button("Submit")
        error_message = gr.Textbox(
            label="Error", visible=False, interactive=False
        )

    with gr.Column(visible=False) as chat_interface:
        system_prompt = gr.Textbox(
            value="You are a helpful assistant.", label="System message"
        )
        model_choice = gr.Dropdown(
            choices=list(AVAILABLE_MODELS.keys()),
            value=list(AVAILABLE_MODELS.keys())[0],
            label="Select Model"
        )
        max_tokens = gr.Slider(
            minimum=1, maximum=30000, value=2048, step=100, label="Max new tokens"
        )
        temperature = gr.Slider(
            minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"
        )
        top_p = gr.Slider(
            minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"
        )

        chat = gr.ChatInterface(
            respond,
            api_name=False,
            chatbot=gr.Chatbot(height=400),  # Set the desired height here
            additional_inputs=[system_prompt, model_choice, max_tokens, temperature, top_p]  # Pass extra parameters
        )

    submit_button.click(
        check_password, inputs=password_input, outputs=[password_input, chat_interface]
    )

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