src/app.py

import sys
import os
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))

import streamlit as st
import requests
import subprocess
import pandas as pd

from load_graph import download_graph_if_needed
download_graph_if_needed()

st.set_page_config(
    page_title="드라마 Intelligence",
    page_icon="🎭",
    layout="wide",
    initial_sidebar_state="collapsed"
)

st.markdown("""
<style>
@import url('https://fonts.googleapis.com/css2?family=Noto+Serif+KR:wght@300;400;700&family=DM+Sans:wght@300;400;500&display=swap');
:root {
    --cherry: #C41E3A;
    --deep: #1a0a0f;
    --gold: #D4AF37;
    --cream: #FDF6EC;
}
html, body, [class*="css"] {
    font-family: 'DM Sans', sans-serif;
    background-color: var(--deep);
    color: var(--cream);
}
.stApp {
    background: linear-gradient(135deg, #1a0a0f 0%, #2d0f1a 50%, #1a0a0f 100%);
}
.hero {
    text-align: center;
    padding: 3rem 0 2rem;
    border-bottom: 1px solid rgba(212, 175, 55, 0.3);
    margin-bottom: 2rem;
}
.hero-title {
    font-family: 'Noto Serif KR', serif;
    font-size: 3.5rem;
    font-weight: 700;
    color: var(--cream);
    letter-spacing: -1px;
    line-height: 1.1;
    margin-bottom: 0.5rem;
}
.hero-korean {
    font-family: 'Noto Serif KR', serif;
    font-size: 1.1rem;
    color: var(--gold);
    letter-spacing: 4px;
    margin-bottom: 1rem;
}
.hero-sub {
    font-size: 1rem;
    color: rgba(253, 246, 236, 0.6);
    font-weight: 300;
}
.method-badge {
    display: inline-block;
    padding: 0.25rem 0.75rem;
    border-radius: 2px;
    font-size: 0.75rem;
    font-weight: 500;
    letter-spacing: 2px;
    text-transform: uppercase;
    margin-bottom: 1rem;
}
.global-badge {
    background: rgba(196, 30, 58, 0.2);
    border: 1px solid var(--cherry);
    color: var(--cherry);
}
.local-badge {
    background: rgba(212, 175, 55, 0.2);
    border: 1px solid var(--gold);
    color: var(--gold);
}
.answer-card {
    background: rgba(253, 246, 236, 0.04);
    border: 1px solid rgba(212, 175, 55, 0.2);
    border-left: 3px solid var(--gold);
    border-radius: 4px;
    padding: 2rem;
    margin-top: 1.5rem;
    font-size: 0.95rem;
    line-height: 1.8;
    color: rgba(253, 246, 236, 0.9);
}
.question-display {
    font-family: 'Noto Serif KR', serif;
    font-size: 1.3rem;
    color: var(--cream);
    margin-bottom: 0.5rem;
    font-style: italic;
}
.divider {
    border: none;
    border-top: 1px solid rgba(212, 175, 55, 0.2);
    margin: 2rem 0;
}
.stTextInput > div > div > input {
    background: rgba(253, 246, 236, 0.05) !important;
    border: 1px solid rgba(212, 175, 55, 0.3) !important;
    border-radius: 2px !important;
    color: var(--cream) !important;
    font-family: 'DM Sans', sans-serif !important;
    font-size: 1rem !important;
    padding: 0.75rem 1rem !important;
}
.stButton > button {
    background: var(--cherry) !important;
    color: var(--cream) !important;
    border: none !important;
    border-radius: 2px !important;
    font-family: 'DM Sans', sans-serif !important;
    font-weight: 500 !important;
    letter-spacing: 1px !important;
    padding: 0.6rem 2rem !important;
}
.stSelectbox > div > div {
    background: rgba(253, 246, 236, 0.05) !important;
    border: 1px solid rgba(212, 175, 55, 0.3) !important;
    color: var(--cream) !important;
    border-radius: 2px !important;
}
.stats-row {
    display: flex;
    gap: 2rem;
    justify-content: center;
    margin: 1.5rem 0;
}
.stat-item { text-align: center; }
.stat-num {
    font-family: 'Noto Serif KR', serif;
    font-size: 2rem;
    color: var(--gold);
    font-weight: 700;
}
.stat-label {
    font-size: 0.75rem;
    color: rgba(253, 246, 236, 0.5);
    letter-spacing: 2px;
    text-transform: uppercase;
}
footer {visibility: hidden;}
#MainMenu {visibility: hidden;}
header {visibility: hidden;}
</style>
""", unsafe_allow_html=True)

st.markdown("""
<div class="hero">
    <div class="hero-korean">드라마 인텔리전스</div>
    <div class="hero-title">K-Drama Intelligence</div>
    <div class="hero-sub">GraphRAG-powered reasoning across 1637 top Korean dramas</div>
    <div class="stats-row">
        <div class="stat-item">
            <div class="stat-num">1637</div>
            <div class="stat-label">Dramas</div>
        </div>
        <div class="stat-item">
            <div class="stat-num">∞</div>
            <div class="stat-label">Connections</div>
        </div>
        <div class="stat-item">
            <div class="stat-num">2</div>
            <div class="stat-label">Query Modes</div>
        </div>
    </div>
</div>
""", unsafe_allow_html=True)

@st.cache_resource(show_spinner=False)
def load_engine():
    project_root  = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
    graphrag_root = os.path.join(project_root, "graphrag_input")
    from query_engine import init_engine
    return init_engine(graphrag_root)


def run_graphrag_query(question, method):
    return load_engine().query_sync(question, method)


@st.cache_data(ttl=3600)
def cached_query(question, method):
    return run_graphrag_query(question, method)

tab1, tab2, tab3 = st.tabs(["🔍 Ask", "🎬 Recommend", "📊 Explore"])

with tab1:
    col1, col2 = st.columns([3, 1])
    with col1:
        question = st.text_input(
            "",
            placeholder="Ask anything about K-dramas — actors, tropes, cultural impact...",
            label_visibility="collapsed"
        )
    with col2:
        method = st.selectbox(
            "",
            ["global", "local"],
            label_visibility="collapsed",
            format_func=lambda x: "🌐 Global" if x == "global" else "🎯 Local"
        )

    examples = [
        "Which dramas became global hits and why?",
        "Which actors have the most diverse roles?",
        "What makes Squid Game different from other survival dramas?",
        "Which dramas share the enemies-to-lovers trope?",
        "Why did Crash Landing on You resonate globally?",
    ]

    st.markdown("**Try asking:**")
    cols = st.columns(len(examples))
    for i, ex in enumerate(examples):
        with cols[i]:
            if st.button(ex[:30] + "...", key=f"ex_{i}", use_container_width=True):
                question = ex

    if st.button("Ask", use_container_width=False):
        if not question.strip():
            st.warning("Please enter a question first.")
        else:
            badge_class = "global-badge" if method == "global" else "local-badge"
            badge_text = "GLOBAL SEARCH" if method == "global" else "LOCAL SEARCH"

            with st.spinner("Traversing the knowledge graph..."):
                try:
                    answer = cached_query(question, method)
                    st.markdown(f'<div class="method-badge {badge_class}">{badge_text}</div>', unsafe_allow_html=True)
                    st.markdown(f'<div class="question-display">"{question}"</div>', unsafe_allow_html=True)

                    from tmdb_helper import extract_dramas_from_text, extract_actors_from_text, get_poster_url, get_actor_photo_url, KNOWN_ACTORS
                    dramas = extract_dramas_from_text(answer)
                    show_actors = any(actor in question.lower() for actor in KNOWN_ACTORS if len(actor) > 4)
                    actors = extract_actors_from_text(answer) if show_actors else []

                    left_col, right_col = st.columns([2, 1])
                    with left_col:
                        st.markdown(f'<div class="answer-card">{answer}</div>', unsafe_allow_html=True)
                    with right_col:
                        for drama in dramas[:3]:
                            poster_url, drama_name = get_poster_url(drama)
                            if poster_url:
                                st.image(poster_url, caption=drama_name, use_container_width=True)
                        for actor in actors[:2]:
                            photo_url, actor_name = get_actor_photo_url(actor)
                            if photo_url:
                                st.image(photo_url, caption=actor_name, use_container_width=True)

                except Exception as e:
                    st.error(f"Error: {str(e)}")

with tab2:
    st.markdown("### Find Similar Dramas")
    neo4j_uri = os.getenv("NEO4J_URI", "")
    neo4j_pw = os.getenv("NEO4J_PASSWORD", "")
    neo4j_user = os.getenv("NEO4J_USERNAME", "61885e62")
    st.caption(f"Neo4j: URI=`{neo4j_uri[:20]}...` | USER=`{neo4j_user}` | PW=`{neo4j_pw[:4]}...{neo4j_pw[-4:]}`({len(neo4j_pw)} chars)" if neo4j_pw else "Neo4j: PASSWORD NOT SET")
    drama_input = st.text_input("Enter a drama name:", placeholder="e.g. Goblin, Crash Landing on You", key="drama_search")

    if st.button("Find Similar", key="recommend"):
        if drama_input.strip():
            from neo4j_queries import get_similar_dramas
            from tmdb_helper import get_poster_url
            try:
                results = get_similar_dramas(drama_input.strip())
                if results:
                    st.markdown(f"**Dramas similar to {drama_input}:**")
                    cols = st.columns(len(results))
                    for i, r in enumerate(results):
                        with cols[i]:
                            poster_url, _ = get_poster_url(r['title'].lower())
                            if poster_url:
                                st.image(poster_url, use_container_width=True)
                            st.markdown(f"**{r['title']}**")
                            st.markdown(f"{r['shared_tropes']} shared tropes")
                else:
                    st.warning("Drama not found. Try exact name like 'goblin'")
            except Exception as e:
                st.error(f"Neo4j error: {str(e)}")

with tab3:
    st.markdown("### K-Drama Knowledge Graph")
    
    col1, col2 = st.columns([2, 1])
    with col1:
        drama_name = st.text_input("Enter a drama to visualize:", 
                                    placeholder="e.g. goblin, crash landing on you",
                                    key="graph_input")
    with col2:
        st.markdown("<br>", unsafe_allow_html=True)
        show_graph = st.button("Visualize Graph", key="graph_btn")
    
    if show_graph and drama_name.strip():
        from neo4j_queries import get_similar_dramas, get_dramas_by_trope
        from neo4j import GraphDatabase
        from pyvis.network import Network
        import streamlit.components.v1 as components
        
        driver = GraphDatabase.driver(
            os.getenv('NEO4J_URI', '').strip(),
            auth=(os.getenv('NEO4J_USERNAME', '61885e62').strip(), os.getenv('NEO4J_PASSWORD', '').strip())
        )
        
        with driver.session() as session:
            result = session.run("""
                MATCH (d:Drama {name: $name})-[:HAS_TROPE]->(t:Trope)
                RETURN d.name as drama, t.name as trope
                LIMIT 20
            """, name=drama_name.lower())
            trope_rows = list(result)
            
            result = session.run("""
                MATCH (a:Actor)-[:ACTED_IN]->(d:Drama {name: $name})
                RETURN a.name as actor, d.name as drama
                LIMIT 10
            """, name=drama_name.lower())
            actor_rows = list(result)
            
            result = session.run("""
                MATCH (d:Drama {name: $name})-[:HAS_GENRE]->(g:Genre)
                RETURN d.name as drama, g.name as genre
            """, name=drama_name.lower())
            genre_rows = list(result)

        driver.close()

        if trope_rows or actor_rows or genre_rows:
            net = Network(
                height='500px', 
                width='100%', 
                bgcolor='#1a0a0f', 
                font_color='white'
            )
            net.barnes_hut()
            net.set_options("""
                    {
                    "nodes": {
                        "font": {
                        "size": 14,
                        "color": "white"
                        },
                        "borderWidth": 2
                    },
                    "edges": {
                        "color": {
                        "color": "rgba(255,255,255,0.3)"
                        },
                        "width": 1.5
                    },
                    "physics": {
                        "barnesHut": {
                        "gravitationalConstant": -8000,
                        "springLength": 150
                        }
                    }
                    }
                    """)
            
            # Add drama node
            net.add_node(
                drama_name, 
                label=drama_name.title(), 
                color='#C41E3A', 
                size=40,
                title=f"Drama: {drama_name.title()}"
            )
            
            # Add trope nodes
            for row in trope_rows:
                trope = row['trope']
                net.add_node(
                    f"t_{trope}", 
                    label=trope, 
                    color='#D4AF37', 
                    size=15,
                    title=f"Trope: {trope}"
                )
                net.add_edge(drama_name, f"t_{trope}")
            
            # Add actor nodes
            for row in actor_rows:
                actor = row['actor']
                net.add_node(
                    f"a_{actor}", 
                    label=actor, 
                    color='#4A90D9', 
                    size=20,
                    title=f"Actor: {actor}"
                )
                net.add_edge(f"a_{actor}", drama_name)
            
            # Add genre nodes
            for row in genre_rows:
                genre = row['genre']
                net.add_node(
                    f"g_{genre}", 
                    label=genre, 
                    color='#2ECC71', 
                    size=18,
                    title=f"Genre: {genre}"
                )
                net.add_edge(drama_name, f"g_{genre}")
            
            # Save and display
            graph_path = '/tmp/kdrama_graph.html'
            net.save_graph(graph_path)
            
            with open(graph_path, 'r') as f:
                html = f.read()
            
            st.markdown("**Legend:** 🔴 Drama · 🟡 Tropes · 🔵 Actors · 🟢 Genres")
            components.html(html, height=520)
        
        else:
            st.warning(f"Drama '{drama_name}' not found. Try lowercase like 'goblin'")
    
    st.markdown("---")
    st.markdown("### Top K-Drama Tropes")
    from neo4j_queries import get_top_tropes
    tropes = get_top_tropes(15)
    if tropes:
        df = pd.DataFrame(tropes)
        st.bar_chart(df.set_index('trope'))

st.markdown('<hr class="divider">', unsafe_allow_html=True)
st.markdown("""
<div style="text-align:center; color: rgba(253,246,236,0.3); font-size: 0.75rem; letter-spacing: 2px;">
    BUILT WITH MICROSOFT GRAPHRAG · OPENAI GPT-4O-MINI · PYTHON · NEO4J
</div>
""", unsafe_allow_html=True)