app.py

import os
import sys
import gradio as gr
from multiprocessing import freeze_support
import importlib
import inspect
import json

# Fix path to include src
sys.path.insert(0, os.path.join(os.path.dirname(__file__), "src"))

# Reload TxAgent from txagent.py
import txagent.txagent
importlib.reload(txagent.txagent)
from txagent.txagent import TxAgent

# Debug info
print(">>> TxAgent loaded from:", inspect.getfile(TxAgent))
print(">>> TxAgent has run_gradio_chat:", hasattr(TxAgent, "run_gradio_chat"))

# Env vars
current_dir = os.path.abspath(os.path.dirname(__file__))
os.environ["MKL_THREADING_LAYER"] = "GNU"
os.environ["TOKENIZERS_PARALLELISM"] = "false"

# Model config
model_name = "mims-harvard/TxAgent-T1-Llama-3.1-8B"
rag_model_name = "mims-harvard/ToolRAG-T1-GTE-Qwen2-1.5B"
new_tool_files = {
    "new_tool": os.path.join(current_dir, "data", "new_tool.json")
}

# Sample questions
question_examples = [
    ["Given a patient with WHIM syndrome on prophylactic antibiotics, is it advisable to co-administer Xolremdi with fluconazole?"],
    ["What treatment options exist for HER2+ breast cancer resistant to trastuzumab?"]
]

# Helper: collapsible format with tool name
def format_collapsible(content, tool_name=None):
    if isinstance(content, (dict, list)):
        try:
            formatted = json.dumps(content, indent=2)
        except Exception:
            formatted = str(content)
    else:
        formatted = str(content)

    title = f"Answer from: {tool_name}" if tool_name else "Answer"

    return f"""
    <details style='border: 1px solid #444; background-color: #1e1e1e; border-radius: 8px; padding: 10px; margin: 10px 0;'>
        <summary style='color: #37B6E9; font-weight: 600; font-size: 16px;'>{title}</summary>
        <pre style='color: #eee; font-family: Consolas, monospace; padding-top: 10px;'>{formatted}</pre>
    </details>
    """

# UI setup
def create_ui(agent):
    custom_css = """
        body {
            font-family: Inter, sans-serif;
            background-color: #121212;
            color: #ffffff;
        }
        .gradio-container {
            max-width: 900px;
            margin: auto;
        }
        textarea, input, .gr-button {
            font-size: 16px;
        }
        .gr-button {
            background: linear-gradient(to right, #37B6E9, #4B4CED);
            color: white;
            border-radius: 8px;
            font-weight: bold;
        }
        .gr-button:hover {
            background: linear-gradient(to right, #4B4CED, #37B6E9);
        }
        .gr-chatbot {
            background-color: #1e1e1e;
            border-radius: 10px;
        }
    """

    with gr.Blocks(css=custom_css) as demo:
        gr.Markdown("<h1 style='text-align: center; color: #4B4CED;'>TxAgent: Therapeutic Reasoning</h1>")
        gr.Markdown("Ask biomedical or therapeutic questions. Powered by step-by-step reasoning and intelligent tool use.")

        temperature = gr.Slider(0, 1, value=0.3, label="Temperature")
        max_new_tokens = gr.Slider(128, 4096, value=1024, label="Max New Tokens")
        max_tokens = gr.Slider(128, 32000, value=8192, label="Max Total Tokens")
        max_round = gr.Slider(1, 50, value=30, label="Max Rounds")
        multi_agent = gr.Checkbox(label="Enable Multi-agent Reasoning", value=False)
        conversation_state = gr.State([])

        chatbot = gr.Chatbot(label="TxAgent", height=600, type="messages")
        message_input = gr.Textbox(placeholder="Ask your biomedical question...", show_label=False)
        send_button = gr.Button("Send")

        # Chat logic
        def handle_chat(message, history, temperature, max_new_tokens, max_tokens, multi_agent, conversation, max_round):
            generator = agent.run_gradio_chat(
                message=message,
                history=history,
                temperature=temperature,
                max_new_tokens=max_new_tokens,
                max_token=max_tokens,
                call_agent=multi_agent,
                conversation=conversation,
                max_round=max_round
            )

            for update in generator:
                formatted = []
                for m in update:
                    role = m["role"] if isinstance(m, dict) else getattr(m, "role", "assistant")
                    content = m["content"] if isinstance(m, dict) else getattr(m, "content", "")
                    tool_name = m.get("metadata", {}).get("tool") if isinstance(m, dict) else getattr(m, "metadata", {}).get("tool", None)

                    if role == "assistant":
                        content = format_collapsible(content, tool_name)

                    formatted.append({"role": role, "content": content})
                yield formatted

        # Events
        inputs = [message_input, chatbot, temperature, max_new_tokens, max_tokens, multi_agent, conversation_state, max_round]
        send_button.click(fn=handle_chat, inputs=inputs, outputs=chatbot)
        message_input.submit(fn=handle_chat, inputs=inputs, outputs=chatbot)

        gr.Examples(examples=question_examples, inputs=message_input)
        gr.Markdown("<div style='text-align: center; font-size: 0.9em; color: #999;'>This demo is for research purposes only and does not provide medical advice.</div>")

    return demo

# Main
if __name__ == "__main__":
    freeze_support()
    try:
        agent = TxAgent(
            model_name=model_name,
            rag_model_name=rag_model_name,
            tool_files_dict=new_tool_files,
            force_finish=True,
            enable_checker=True,
            step_rag_num=10,
            seed=100,
            additional_default_tools=[]
        )
        agent.init_model()

        if not hasattr(agent, "run_gradio_chat"):
            raise AttributeError("TxAgent missing run_gradio_chat")

        demo = create_ui(agent)
        demo.queue().launch(server_name="0.0.0.0", server_port=7860, share=True, show_error=True)

    except Exception as e:
        print(f"❌ App failed to start: {e}")
        raise