import streamlit as st
import numpy as np
from pathlib import Path
import sys

# Ensure src directory is on Python path for inference import
sys.path.append(str((Path(__file__).resolve().parent)))

from inference.model import run_inference

# Configure the Streamlit page
st.set_page_config(page_title="DeepECG PDF Demo", layout="centered")

st.title("\U0001F4AC DeepECG - PDF Analyse (Demo)")
st.markdown(
    """
    **Hinweis:**
    Diese Demo analysiert EKG-PDFs zu Evaluations- und Forschungszwecken.
    Kein Medizinprodukt. Keine Patientendaten speichern.
    """
)

# File uploader for EKG PDFs
uploaded_file = st.file_uploader(
    "EKG-PDF hochladen",
    type=["pdf"]
)

# When a file is uploaded
if uploaded_file is not None:
    st.success("PDF erfolgreich hochgeladen")

    if st.button("\U0001F9E0 Analyse starten"):
        with st.spinner("Analyse läuft ... bitte warten"):
            # TODO: Replace this dummy signal with real PDF-to-signal extraction
            dummy_signal = np.random.rand(12, 5000)

            # Run inference (real model)
            results = run_inference(dummy_signal)

            # Display results
            st.subheader("Ergebnis")

            st.metric(
                label="Gesamtrisiko",
                value=f"{results['risk_score']} %"
            )

            st.subheader("Top-Diagnosen")
            for diag in results["top_diagnoses"]:
                st.write(f"- {diag['label']} ({diag['probability']} %)")

            st.subheader("Einschätzung")
            st.success(results["interpretation"])

            st.subheader("Beispiel-EKG (Lead I)")
            st.line_chart(dummy_signal[0])