import streamlit as st
from pathlib import Path


# Load custom CSS
def load_css():
    css_file = Path(__file__).parent / ".streamlit" / "style.css"
    if css_file.exists():
        with open(css_file) as f:
            st.markdown(f"<style>{f.read()}</style>", unsafe_allow_html=True)


load_css()


home_page = [
    st.Page("pages/home.py", title="Welcome", icon=":material/home:", default=True),
]

bdld_pages = [
    st.Page(
        "pages/bayesian-deep-learning/00_function_space_manifold.py",
        title="Function Space Manifolds",
        icon=":material/animation:",
    ),
    st.Page(
        "pages/bayesian-deep-learning/01_deep_net_pathologies.py",
        title="Deep Net Prior Pathologies",
        icon=":material/visibility:",
    ),
    st.Page(
        "pages/bayesian-deep-learning/02_aleatoric_epistemic_uncertainty.py",
        title="Aleatoric vs Epistemic Uncertainty",
        icon=":material/insights:",
    ),
    st.Page(
        "pages/bayesian-deep-learning/03_bnn_prior_optimization.py",
        title="Prior Optimization",
        icon=":material/tune:",
    ),
]

gp_pages = [
    st.Page(
        "pages/gaussian-process/00_gp_prior_joint.py",
        title="GP Joint Distribution",
        icon=":material/scatter_plot:",
    ),
]

deep_learning_pages = [
    st.Page(
        "pages/deep-learning/00_grokking_modular_arithmetic.py",
        title="Grokking on Modular Arithmetic",
        icon=":material/trending_up:",
    ),
]


pg = st.navigation(
    {
        "Home": home_page,
        "Bayesian Deep Learning": bdld_pages,
        "Gaussian Processes": gp_pages,
        "Deep Learning": deep_learning_pages,
    },
    position="sidebar",
)
st.set_page_config(
    layout="wide",
    initial_sidebar_state="collapsed",
    page_icon=":material/menu_book:",
    page_title="Demo",
)
pg.run()