import json
import os
import re
from difflib import get_close_matches

# --- Paste hier die Model-Dateinamen rein (oder lies sie aus dem Ordner) ---
MODEL_FILES = [
    "Ebisu.glb", "Enenra.glb", "Enenra2.glb", "Oboroguruma.glb", "Oiwa.glb", "Okiku.glb", "Okomeki.001.glb",
    "Okuninushi.glb", "Oni.glb", "Onryo.glb", "Oyamatsumi.001.glb", "Raijin.glb", "Rokurokubi.glb", "Ryujin.glb",
    "Sarutahiko_Okami.glb", "Shinigami.001.glb", "Shuten_Doji.glb", "Sojobo.glb", "Sojobo2.glb", "Susanoo.glb",
    "Takeminakata.glb", "Takeminakata2.001.glb", "Tanuki.glb", "Tengu.glb", "Tenjin.glb", "Tsukumogami.glb",
    "Tsukuyomi_No_Mikoto.glb", "Tsurube_Otoshi.glb", "Tsurube_Otoshi2.glb", "Tsurube_Otoshi3.glb", "Tsurube_Otoshi4.glb",
    "Ubume.glb", "Yama_Uba.glb", "Yama_Uba2.glb", "Yamata_No_Orichi.glb", "Yamawaro.glb", "Yatagarasu2.glb",
    "Yuki_Onna.glb", "Yurei.glb", "Abe_No_Seimei.glb", "Abura_Akago.glb", "Abura_Sumashi.glb", "Abura_Sumashi2.glb",
    "Aka_Manto.glb", "Akaname.glb", "Akateko2.glb", "Akkorokamui.glb", "Akuchu.glb", "Amabie2.glb", "Amanojaku.glb",
    "Amaterasu.glb", "Ame_No_Uzume.001.glb", "Amenominakanushi.glb", "Aoandon.001.glb", "Aoandon2.001.glb",
    "Ashiari_Yashiki.glb", "Ashinaga_Tenaga2.glb", "Azukiarai.glb", "Azukibabaa.glb", "Azukihakari.glb",
    "Bake_Kujira.glb", "Bake_Kujira2.glb", "Bake_Kujira3.glb", "Bakezori.glb", "Baku.glb", "Basan.glb",
    "Benzaiten.glb", "Betobeto_San.glb", "Bishamonten.glb", "Biwa_Bokuboku.glb", "Chochin_Obake.glb", "Daidarabotchi.glb",
    "Daikokuten+Text.glb", "Daikokuten.glb", "Fujin.glb", "Funayurei.glb", "Furaribi.glb", "Futakuchi_Onna.glb",
    "Gaki.glb", "Gashadokuro.glb", "Hachiman.glb", "Hiderigami.001.glb", "Hitotsume_Kozo.glb", "Hoko.glb",
    "Inari_Okami.glb", "Ittan_Momen2.glb", "Izanagi_No_Mikoto.glb", "Izanami_No_Mikoto.glb", "Jikininki.glb",
    "Jorogumo3.glb", "Kamaitachi.glb", "Kamikiri.glb", "Kappa.glb", "Karakasa_Obake.glb", "Karakasa_Obake2.glb",
    "Kitsune.glb", "Kodama.glb", "Kudan.glb", "Mizushi.glb", "Mokumokuren.glb", "Mujina.glb", "Nekomata.glb",
    "Noppera_Bo.glb", "Nue.glb", "Nuppeppo2.glb", "Nurarihyon.glb", "Nure_Onna.glb", "Nurikabe.glb", "Nurikabe2.glb"
]

def normalize(s):
    s = s.lower()
    s = re.sub(r"[^a-z0-9]", "", s)
    s = s.replace("ou", "o") # YamatanoOrOchi vs Yamata_No_Orichi etc.
    return s

def generate_candidates(spirit_name):
    # z.B. "Tsukuyomi (月読命)" => ["Tsukuyomi", "TsukuyomiNoMikoto", ...]
    # Extrahiere lateinische Namen und alle Wörter, splitte auf Sonderzeichen
    base = spirit_name.split()[0]
    latin = re.split(r"\s|\(|（", spirit_name)[0]
    # Alternativen, z.B. alles ohne Klammern, nur erstes Wort etc.
    candidates = [latin]
    candidates += [latin.replace("-", "_"), latin.replace("-", ""), latin.replace("_", ""), latin.title(), latin.upper()]
    # Häufig bei Kami: _No_Mikoto-Suffix
    if not latin.endswith("NoMikoto"):
        candidates.append(latin + "NoMikoto")
        candidates.append(latin + "_No_Mikoto")
    # Auch mal alles Klein, Snake, Camel
    return list(set([normalize(c) for c in candidates]))

def find_best_model(spirit_name):
    candidates = generate_candidates(spirit_name)
    model_names = [f[:-4] for f in MODEL_FILES] # .glb weg
    normalized_models = [normalize(n) for n in model_names]
    # Alle Matches (ab Distanz <= 2 oder exaktes Substring-Match)
    results = []
    for c in candidates:
        for i, n in enumerate(normalized_models):
            dist = levenshtein(c, n)
            if dist <= 2 or c in n or n in c:
                results.append(MODEL_FILES[i])
    # Unique!
    results = sorted(list(set(results)))
    if not results:
        # Fuzzy best 3
        matches = get_close_matches(candidates[0], normalized_models, n=3, cutoff=0.6)
        models = [MODEL_FILES[normalized_models.index(m)] for m in matches]
        return models
    return results

# Levenshtein-Distanz
def levenshtein(a, b):
    if a == b: return 0
    if not a: return len(b)
    if not b: return len(a)
    v0 = list(range(len(b) + 1))
    v1 = [0] * (len(b) + 1)
    for i in range(len(a)):
        v1[0] = i + 1
        for j in range(len(b)):
            cost = 0 if a[i] == b[j] else 1
            v1[j + 1] = min(v1[j] + 1, v0[j + 1] + 1, v0[j] + cost)
        v0, v1 = v1, v0
    return v0[len(b)]

# ---- Hauptlogik ----
def main():
    with open("spirit_list.json", encoding="utf-8") as f:
        spirits = json.load(f)
    for spirit in spirits:
        if "Model URL" in spirit and spirit["Model URL"]:
            continue # already done
        name = spirit.get("Name", "")
        matches = find_best_model(name)
        if len(matches) == 1:
            spirit["Model URL"] = "/assets/models/spirits/" + matches[0]
        elif len(matches) > 1:
            print(f"\n[?] Mehrere mögliche Modelle für '{name}': {matches}")
            spirit["Model URL"] = "/assets/models/spirits/" + matches[0] # Default das erste, Handcheck empfohlen!
        else:
            print(f"[!] Kein Modell gefunden für '{name}'!")
            spirit["Model URL"] = ""
    with open("spirit_list_out.json", "w", encoding="utf-8") as f:
        json.dump(spirits, f, ensure_ascii=False, indent=2)
    print("\nFERTIG. Ergebnis: spirit_list_out.json")

if __name__ == "__main__":
    main()