import base64
from pathlib import Path

import streamlit as st

from src.ui_style import apply_global_style

st.set_page_config(page_title="Home", layout="wide")
apply_global_style()

logo_path = Path(__file__).resolve().parent / "icons" / "logo.png"
logo_data_uri = ""
if logo_path.exists():
    logo_data_uri = "data:image/png;base64," + base64.b64encode(logo_path.read_bytes()).decode("ascii")

logo_html = (
    f'<img src="{logo_data_uri}" style="width:120px; height:120px; object-fit:contain; border-radius:10px;" />'
    if logo_data_uri
    else ""
)

st.markdown(
    f"""
<div style="background:#123a69; border-radius:14px; padding:16px 22px; margin: 4px 0 12px 0;">
  <div style="display:flex; align-items:center; gap:18px;">
    {logo_html}
    <div style="color:#ffffff; font-size:2.8rem; line-height:1.05; font-weight:800;">
      Polymer Discovery Platform
    </div>
  </div>
</div>
""",
    unsafe_allow_html=True,
)

st.markdown(
    """
This platform provides an end-to-end workflow for polymer screening and selection: quick single-polymer checks,
bulk property prediction, 2D/3D molecular visualization, and multi-objective discovery with Pareto analysis,
trust scoring, and diversity-aware candidate selection. You can run the process with manual controls or
AI-assisted setup to accelerate exploration from requirements to shortlisted candidates.
"""
)

st.divider()
st.markdown("### Platform Modules")
st.caption(
    "Use the modules below to probe, predict, visualize, and discover polymers with manual control or AI support."
)

cards = [
    (
        "Property Probe",
        "Input a single SMILES or polymer name and retrieve predicted or available values for one target property. "
        "Best for quick validation before larger screening.",
        "pages/1_Property_Probe.py",
    ),
    (
        "Batch Prediction",
        "Upload or paste many SMILES and run bulk property prediction in one job. "
        "Useful when you want ranked outputs and exportable tables for downstream analysis.",
        "pages/2_Batch_Prediction.py",
    ),
    (
        "Molecular View",
        "Render 2D and 3D molecular structures, inspect composition, and download visual assets "
        "or MOL files for documentation and simulation setup.",
        "pages/3_Molecular_View.py",
    ),
    (
        "Discovery (Manual)",
        "Set hard constraints, objectives, trust/selection weights, and diversity settings directly. "
        "Designed for controlled multi-objective exploration with transparent parameter tuning.",
        "pages/4_Discovery_(Manual).py",
    ),
    (
        "Discovery (AI)",
        "Describe target behavior in natural language and let the LLM build discovery settings. "
        "You can run directly or inspect/edit the generated JSON in advanced mode.",
        "pages/5_Discovery_(AI).py",
    ),
    (
        "Novel SMILES Generation",
        "Sample new polymer SMILES with the pretrained RNN and filter out molecules already present "
        "in local datasets (EXP/MD/DFT/GC/POLYINFO/PI1M).",
        "pages/6_Novel_SMILES_Generation.py",
    ),
    (
        "Feedback",
        "Send bug reports, feature requests, and usage feedback.",
        "pages/7_Feedback.py",
    ),
]

for i, (title, desc, page_path) in enumerate(cards, start=1):
    box = st.container(border=True)
    with box:
        c1, c2 = st.columns([5, 1.2])
        with c1:
            st.markdown(f"**{title}**")
            st.caption(desc)
        with c2:
            if st.button("Go", type="primary", key=f"home_go_{i}"):
                st.switch_page(page_path)

st.divider()
st.markdown("**Developed by**")
st.markdown("### MONSTER Lab")
st.markdown("**Molecular/Nano-Scale Transport & Energy Research Laboratory**")
st.markdown("**College of Engineering, University of Notre Dame**")
st.markdown(
    "The MÖNSTER Lab (MOlecular/Nano-Scale Transport & Energy Research Laboratory) "
    "studies the fundamental physics of energy and mass transport from the molecular and "
    "nano-scale using theories, simulations, data-driven approaches and experiments, and "
    "apply the knowledge toward engineering materials with tailored thermal properties, "
    "thermal management of electronics, improving efficiency of energy devices, designing "
    "molecules and system for water desalination, high-sensitivity bio-sensing and additive "
    "manufacturing."
)
st.markdown("[Visit MONSTER Lab Website](https://monsterlab.nd.edu/)")