import streamlit as st
import pickle
from scipy import spatial
import random
import os

# --- 1. SAYFA YAPISI VE CSS (Titremeyi Engellemek İçin En Önemli Kısım) ---
st.set_page_config(page_title="CastMatch AI", layout="wide")

# Resim kutusunun zıplamasını engelleyen özel CSS
st.markdown("""
    <style>
    .stImage > img {
        border-radius: 10px;
        max-height: 450px;
        object-fit: cover;
    }
    div[data-testid="stVerticalBlock"] > div:has(div.stImage) {
        min-height: 450px; 
    }
    </style>
    """, unsafe_allow_html=True)

# --- 2. VERİ VE MODEL (Cache/Önbellek) ---
@st.cache_resource
def load_model():
    try:
        with open('movie_model.pkl', 'rb') as f:
            return pickle.load(f)
    except:
        return None

model_data = load_model()
movie_dict = model_data['movie_dict'] if model_data else {}

# Veritabanı (Sabit tutuldu)
tum_unlu_verileri = {
    "Sylvester Stallone": {"c": "Male / Erkek", "v": [0.9, 0.2, 0.1, 0.1, 0.1, 0.8]},
    "Robert De Niro": {"c": "Male / Erkek", "v": [0.2, 0.1, 0.1, 0.9, 0.1, 0.8]},
    "Johnny Depp": {"c": "Male / Erkek", "v": [0.1, 0.9, 0.8, 0.4, 0.1, 0.1]},
    "Jason Statham": {"c": "Male / Erkek", "v": [0.95, 0.3, 0.1, 0.1, 0.1, 0.7]},
    "Keanu Reeves": {"c": "Male / Erkek", "v": [0.9, 0.4, 0.1, 0.2, 0.9, 0.6]},
    "Brad Pitt": {"c": "Male / Erkek", "v": [0.7, 0.5, 0.3, 0.8, 0.2, 0.4]},
    "Tom Cruise": {"c": "Male / Erkek", "v": [0.9, 0.8, 0.2, 0.1, 0.4, 0.5]},
    "Anthony Hopkins": {"c": "Male / Erkek", "v": [0.1, 0.1, 0.1, 0.95, 0.1, 0.9]},
    "Arnold Schwarzenegger": {"c": "Male / Erkek", "v": [0.95, 0.6, 0.2, 0.1, 0.8, 0.1]},
    "Bruce Willis": {"c": "Male / Erkek", "v": [0.9, 0.2, 0.1, 0.3, 0.4, 0.7]},
    "Christian Bale": {"c": "Male / Erkek", "v": [0.8, 0.2, 0.1, 0.9, 0.1, 0.7]},
    "Tom Hanks": {"c": "Male / Erkek", "v": [0.2, 0.6, 0.5, 0.9, 0.1, 0.1]},
    "Morgan Freeman": {"c": "Male / Erkek", "v": [0.1, 0.3, 0.1, 0.9, 0.1, 0.6]},
    "Al Pacino": {"c": "Male / Erkek", "v": [0.2, 0.1, 0.1, 0.95, 0.1, 0.8]},
    "Cillian Murphy": {"c": "Male / Erkek", "v": [0.4, 0.1, 0.1, 0.9, 0.4, 0.8]},
    "Jackie Chan": {"c": "Male / Erkek", "v": [0.9, 0.7, 0.9, 0.1, 0.1, 0.1]},
    "Denzel Washington": {"c": "Male / Erkek", "v": [0.8, 0.2, 0.1, 0.9, 0.1, 0.7]},
    "George Clooney": {"c": "Male / Erkek", "v": [0.3, 0.4, 0.5, 0.8, 0.1, 0.6]},
    "Jake Gyllenhaal": {"c": "Male / Erkek", "v": [0.6, 0.1, 0.1, 0.9, 0.2, 0.8]},
    "Pierce Brosnan": {"c": "Male / Erkek", "v": [0.8, 0.7, 0.3, 0.4, 0.1, 0.6]},
    "Angelina Jolie": {"c": "Female / Kadın", "v": [0.9, 0.7, 0.1, 0.8, 0.1, 0.6]},
    "Natalie Portman": {"c": "Female / Kadın", "v": [0.3, 0.2, 0.1, 0.9, 0.6, 0.7]},
    "Scarlett Johansson": {"c": "Female / Kadın", "v": [0.9, 0.6, 0.2, 0.7, 0.8, 0.5]},
    "Charlize Theron": {"c": "Female / Kadın", "v": [0.8, 0.3, 0.1, 0.9, 0.2, 0.7]},
    "Milla Jovovich": {"c": "Female / Kadın", "v": [0.9, 0.4, 0.1, 0.1, 0.9, 0.3]},
    "Sandra Bullock": {"c": "Female / Kadın", "v": [0.2, 0.1, 0.9, 0.8, 0.1, 0.6]},
    "Nicole Kidman": {"c": "Female / Kadın", "v": [0.1, 0.2, 0.1, 0.95, 0.1, 0.7]},
    "Jennifer Lawrence": {"c": "Female / Kadın", "v": [0.7, 0.8, 0.3, 0.9, 0.4, 0.2]},
    "Meryl Streep": {"c": "Female / Kadın", "v": [0.1, 0.1, 0.4, 0.95, 0.1, 0.1]},
    "Salma Hayek": {"c": "Female / Kadın", "v": [0.6, 0.3, 0.5, 0.8, 0.1, 0.1]},
    "Anne Hathaway": {"c": "Female / Kadın", "v": [0.2, 0.7, 0.8, 0.8, 0.1, 0.1]},
    "Keira Knightley": {"c": "Female / Kadın", "v": [0.1, 0.8, 0.1, 0.95, 0.1, 0.1]},
    "Julia Roberts": {"c": "Female / Kadın", "v": [0.1, 0.2, 0.9, 0.9, 0.1, 0.1]},
    "Penelope Cruz": {"c": "Female / Kadın", "v": [0.2, 0.3, 0.6, 0.9, 0.1, 0.6]},
    "Monica Bellucci": {"c": "Female / Kadın", "v": [0.6, 0.1, 0.1, 0.8, 0.5, 0.3]},
    "Emily Blunt": {"c": "Female / Kadın", "v": [0.7, 0.6, 0.1, 0.8, 0.5, 0.6]},
    "Cameron Diaz": {"c": "Female / Kadın", "v": [0.1, 0.2, 0.95, 0.7, 0.1, 0.1]}
}

# --- 3. DOSYA TARAMA (Cache/Önbellek) ---
@st.cache_data
def get_images():
    return [f for f in os.listdir(os.getcwd()) if f.lower().endswith(('.png', '.jpg', '.jpeg'))]

all_images = get_images()

# Mevcut olanları filtrele
unlu_verileri = {}
for n, v in tum_unlu_verileri.items():
    if any(n.lower().replace(" ","_") in f.lower() or n.lower() in f.lower() for f in all_images):
        unlu_verileri[n] = v

# --- 4. ARAYÜZ ---
st.title("🎬 CastMatch AI")
st.markdown("---")

col_actor, col_match = st.columns([1, 2])

# Sidebar Kategorisi
gender = st.sidebar.radio("Category / Kategori:", ["Male / Erkek", "Female / Kadın"], key="g_radio")
names = [n for n, d in unlu_verileri.items() if d['c'] == gender]

with col_actor:
    st.subheader("👤 Cast / Oyuncu")
    # INDEX kullanarak seçimi sabitlemek titremeyi azaltır
    actor = st.selectbox("Select / Seç:", names if names else ["-"], key="act_select")
    
    # Resim alanı için sabit bir placeholder
    placeholder = st.empty()
    
    if actor != "-":
        target = actor.lower().replace(" ","_")
        target_alt = actor.lower()
        img_file = next((f for f in all_images if f.lower().startswith(target) or f.lower().startswith(target_alt)), None)
        
        if img_file:
            placeholder.image(img_file, use_container_width=True)

with col_match:
    st.subheader("🎯 Matches / Eşleşmeler")
    if st.button("🚀 Match / Eşleştir", key="m_btn") and actor != "-":
        if movie_dict:
            vec = unlu_verileri[actor]['v']
            res = []
            for i in movie_dict:
                d = spatial.distance.cosine(vec, movie_dict[i][1][:6]) + random.uniform(0, 0.00001)
                res.append((movie_dict[i][0], d))
            
            top = sorted(res, key=lambda x: x[1])[:5]
            for i, (f, s) in enumerate(top, 1):
                p = round((1-s)*100, 1)
                st.success(f"**{i}. {f}** (%{p})")
                st.progress(p/100)