app.py

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

sys.path.insert(0, os.path.join(os.path.dirname(__file__), "src"))
import txagent.txagent
importlib.reload(txagent.txagent)
from txagent.txagent import TxAgent

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

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")
}

question_examples = [
    ["Given a patient with WHIM syndrome on antibiotics, is Xolremdi + fluconazole advisable?"],
    ["What treatment options exist for HER2+ breast cancer resistant to trastuzumab?"]
]

def extract_sections(content):
    """
    Example extractor splitting into sections. You should improve it to parse actual keys.
    """
    return {
        "Summary": content[:1000],  # simulate
        "Clinical Studies": content[1000:2500],
        "Drug Interactions": "See CYP3A4 interactions...",
        "Pharmacokinetics": "- Absorption: Oral\n- Half-life: ~24h\n- Metabolized by CYP3A4"
    }

def create_ui(agent):
    with gr.Blocks() as demo:
        gr.Markdown("<h1 style='text-align: center;'>TxAgent: Therapeutic Reasoning</h1>")
        gr.Markdown("Ask therapeutic or biomedical questions. Results are categorized for readability.")

        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.Tabs()
        summary_box = gr.Markdown(label="Summary")
        studies_box = gr.Markdown(label="Clinical Studies")
        interactions_box = gr.Markdown(label="Drug Interactions")
        kinetics_box = gr.Markdown(label="Pharmacokinetics")

        with chatbot:
            with gr.TabItem("Summary"):
                summary_display = summary_box
            with gr.TabItem("Clinical Studies"):
                studies_display = studies_box
            with gr.TabItem("Drug Interactions"):
                interactions_display = interactions_box
            with gr.TabItem("Pharmacokinetics"):
                kinetics_display = kinetics_box

        message_input = gr.Textbox(placeholder="Ask your biomedical question...", show_label=False)
        send_button = gr.Button("Send", variant="primary")

        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
            )

            final_output = ""
            for update in generator:
                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", "")
                    if role == "assistant":
                        final_output += content + "\n"

            sections = extract_sections(final_output)
            return sections["Summary"], sections["Clinical Studies"], sections["Drug Interactions"], sections["Pharmacokinetics"]

        send_button.click(
            fn=handle_chat,
            inputs=[message_input, [], temperature, max_new_tokens, max_tokens, multi_agent, conversation_state, max_round],
            outputs=[summary_box, studies_box, interactions_box, kinetics_box]
        )

        message_input.submit(
            fn=handle_chat,
            inputs=[message_input, [], temperature, max_new_tokens, max_tokens, multi_agent, conversation_state, max_round],
            outputs=[summary_box, studies_box, interactions_box, kinetics_box]
        )

        gr.Examples(examples=question_examples, inputs=message_input)
        gr.Markdown("**DISCLAIMER**: For research only. Not medical advice.")

    return demo

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.launch(show_error=True)

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