import gradio as gr
import os
import sys
import pandas as pd
import pdfplumber

# Add src to Python path
sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "..", "src")))
from txagent.txagent import TxAgent


def extract_all_text_from_csv(file_path):
    try:
        df = pd.read_csv(file_path, low_memory=False)
        return df.to_string(index=False)
    except Exception as e:
        return f"Error parsing CSV: {e}"


def extract_all_text_from_pdf(file_path):
    extracted = []
    try:
        with pdfplumber.open(file_path) as pdf:
            for page in pdf.pages:
                tables = page.extract_tables()
                for table in tables:
                    for row in table:
                        if any(row):
                            extracted.append("\t".join([cell or "" for cell in row]))
        return "\n".join(extracted)
    except Exception as e:
        return f"Error parsing PDF: {e}"


def create_ui(agent: TxAgent):
    with gr.Blocks(theme=gr.themes.Soft()) as demo:
        gr.Markdown("<h1 style='text-align: center;'>🧠 CPS: Clinical Processing System</h1>")
        chatbot = gr.Chatbot(label="CPS Assistant", height=600, type="messages")

        # Hidden file upload, attached to input bar
        with gr.Row():
            uploaded_files = gr.File(
                label="📎", file_types=[".pdf", ".txt", ".docx", ".jpg", ".png", ".csv"],
                file_count="multiple", visible=False
            )

            with gr.Column(scale=10):
                message_input = gr.Textbox(
                    placeholder="Type your medical question or upload files...", show_label=False, scale=10
                )

            with gr.Column(scale=1, min_width=60):
                file_icon = gr.UploadButton("📎", file_types=[".pdf", ".csv", ".docx", ".txt", ".jpg", ".png"], file_count="multiple")

            send_button = gr.Button("Send", variant="primary")

        conversation_state = gr.State([])

        def handle_chat(message, history, conversation, new_files):
            context = (
                "You are a clinical AI reviewing medical interview or form data. "
                "Analyze the extracted content and reason step-by-step about what the doctor could have missed. "
                "Don't answer yet — just reason."
            )

            if new_files:
                extracted_text = ""
                for file in new_files:
                    path = file.name
                    if path.endswith(".csv"):
                        extracted_text += extract_all_text_from_csv(path) + "\n"
                    elif path.endswith(".pdf"):
                        extracted_text += extract_all_text_from_pdf(path) + "\n"
                    else:
                        extracted_text += f"(Uploaded file: {os.path.basename(path)})\n"

                message = f"{context}\n\n---\n{extracted_text.strip()}\n---\n\nNow reason what the doctor might have missed."

            generator = agent.run_gradio_chat(
                message=message,
                history=history,
                temperature=0.3,
                max_new_tokens=1024,
                max_token=8192,
                call_agent=False,
                conversation=conversation,
                uploaded_files=new_files,
                max_round=30
            )
            for update in generator:
                yield update

        # Bind send logic
        file_icon.upload(fn=None, inputs=[], outputs=[uploaded_files])
        send_button.click(fn=handle_chat, inputs=[message_input, chatbot, conversation_state, uploaded_files], outputs=chatbot)
        message_input.submit(fn=handle_chat, inputs=[message_input, chatbot, conversation_state, uploaded_files], outputs=chatbot)

        gr.Examples([["Upload your medical form and ask what the doctor might’ve missed."]], inputs=message_input)

    return demo