Remove Russian Songs

"""
Move 'Russian' songs from your Liked Songs into a separate playlist.
Detection heuristics:
  1) Cyrillic characters in track name or any artist name.
  2) Artist genres contain keywords like 'russian', 'russia', 'cccp', 'cis', 'soviet', 'рус', 'ру', etc.

No likes are removed. If you want to also unlike after moving, set REMOVE_FROM_LIKED=True.
"""

import json
import re
import time
import unicodedata
from pathlib import Path
from typing import List, Dict, Iterable

from dotenv import load_dotenv
from spotipy import Spotify
from spotipy.exceptions import SpotifyException
from spotipy.oauth2 import SpotifyOAuth

# Load environment variables from .env file
load_dotenv()

# -------------------- CONFIG --------------------
PLAYLIST_NAME = "Russian Songs (from Liked)"
PLAYLIST_PUBLIC = False
REMOVE_FROM_LIKED = False  # set True if you want to unlike after moving
BATCH_SIZE = 100  # Spotify allows 100 per add/remove call
CACHE_FILE = Path(".spotify_cache.json")

# Keywords to flag in artist genres (lowercased)
GENRE_KEYWORDS = {
    "russian", "russia", "ru", "рус", "ру", "soviet", "русский", "cccp", "russkiy", "russky",
    "russian hip hop", "russian rap", "russian pop", "russian rock"
}

# Optional: add specific artist IDs you want to always treat as Russian (override)
ALWAYS_RUSSIAN_ARTIST_IDS = set()  # e.g., {"0Xb4bTQ9C7k2rgnAru8v9A"}

# Optional: add specific artist IDs you want to exclude (never Russian)
NEVER_RUSSIAN_ARTIST_IDS = set()

# ------------------------------------------------

def has_cyrillic(text: str) -> bool:
    if not text:
        return False
    for ch in text:
        if "CYRILLIC" in unicodedata.name(ch, ""):
            return True
    return False

def chunked(iterable: Iterable, n: int) -> Iterable[List]:
    buf = []
    for item in iterable:
        buf.append(item)
        if len(buf) == n:
            yield buf
            buf = []
    if buf:
        yield buf

def safe_call(fn, *args, **kwargs):
    """Retry on rate limits (HTTP 429)."""
    while True:
        try:
            return fn(*args, **kwargs)
        except SpotifyException as e:
            if e.http_status == 429:
                retry_after = int(e.headers.get("Retry-After", "2"))
                time.sleep(retry_after + 1)
                continue
            raise

def get_or_create_playlist(sp: Spotify, user_id: str, name: str, public: bool) -> str:
    # Try to find existing playlist by exact name (first page or two is enough for most users)
    results = safe_call(sp.current_user_playlists, limit=50)
    while results:
        for pl in results["items"]:
            if pl["name"] == name and pl["owner"]["id"] == user_id:
                return pl["id"]
        if results.get("next"):
            results = safe_call(sp.next, results)
        else:
            break
    # Create if not found
    created = safe_call(sp.user_playlist_create, user=user_id, name=name, public=public, description="Auto-collected Russian songs from Liked")
    return created["id"]

def classify_track(sp: Spotify, track_obj: Dict, artist_cache: Dict[str, Dict]) -> tuple[bool, str]:
    """
    Returns (True, reason) if the track is considered 'Russian' by heuristics.
    Returns (False, "") otherwise.
    """
    track_name = (track_obj.get("name") or "").strip()

    # Heuristic 1: Cyrillic in track title
    if has_cyrillic(track_name):
        return True, f"Cyrillic in track title: '{track_name}'"

    artists = track_obj.get("artists") or []
    artist_ids = [a["id"] for a in artists if a.get("id")]
    artist_names = [a.get("name") or "" for a in artists]

    # Heuristic 2: Cyrillic in any artist name
    for artist_name in artist_names:
        if has_cyrillic(artist_name):
            return True, f"Cyrillic in artist name: '{artist_name}'"

    # Respect explicit overrides
    if any(aid in ALWAYS_RUSSIAN_ARTIST_IDS for aid in artist_ids):
        return True, "Artist in ALWAYS_RUSSIAN_ARTIST_IDS"
    if any(aid in NEVER_RUSSIAN_ARTIST_IDS for aid in artist_ids):
        return False, ""

    # Heuristic 3: genre-based
    # Fetch artist details (cached)
    to_fetch = [aid for aid in artist_ids if aid and aid not in artist_cache]
    for chunk in chunked(to_fetch, 50):
        if chunk:
            arts = safe_call(sp.artists, chunk)
            for a in arts["artists"]:
                artist_cache[a["id"]] = a

    for aid in artist_ids:
        a = artist_cache.get(aid)
        if not a:
            continue
        artist_name = a.get("name", "Unknown")
        genres = [g.lower() for g in (a.get("genres") or [])]
        for genre in genres:
            for keyword in GENRE_KEYWORDS:
                # Use word boundary matching to avoid false positives
                # e.g., "ru" shouldn't match "ruby" or "grunge"
                if re.search(r'\b' + re.escape(keyword) + r'\b', genre):
                    return True, f"Genre match: '{genre}' (artist: {artist_name}, keyword: '{keyword}')"

    return False, ""

def load_cache() -> Dict:
    if not CACHE_FILE.is_file():
        return {}
    with CACHE_FILE.open("r", encoding="utf-8") as f:
        return json.load(f)

def save_cache(data: Dict):
    with CACHE_FILE.open("w", encoding="utf-8") as f:
        json.dump(data, f, ensure_ascii=False, indent=2)

def get_playlist_tracks(sp: Spotify, playlist_id: str) -> set:
    """Get all track URIs currently in the playlist to avoid duplicates."""
    track_uris = set()
    results = safe_call(sp.playlist_items, playlist_id, limit=100)
    while results:
        for item in results.get("items", []):
            track = item.get("track")
            if track and track.get("uri"):
                track_uris.add(track["uri"])
        if results.get("next"):
            results = safe_call(sp.next, results)
        else:
            break
    return track_uris

def main():
    # Auth
    scope = "user-library-read playlist-modify-private playlist-modify-public user-library-modify"
    sp = Spotify(auth_manager=SpotifyOAuth(scope=scope))
    me = safe_call(sp.current_user)
    user_id = me["id"]

    # Load cached artist data
    cache_data = load_cache()
    artist_cache = cache_data.get("artists", {})

    playlist_id = get_or_create_playlist(sp, user_id, PLAYLIST_NAME, PLAYLIST_PUBLIC)

    print(f"Collecting Russian tracks from Liked Songs into playlist: {PLAYLIST_NAME}")
    print(f"Playlist ID: {playlist_id}")

    # Get existing playlist tracks to avoid duplicates
    print("Checking existing playlist contents...")
    existing_tracks = get_playlist_tracks(sp, playlist_id)
    print(f"Found {len(existing_tracks)} tracks already in playlist")

    # Iterate through liked tracks
    russian_track_uris = []
    results = safe_call(sp.current_user_saved_tracks, limit=50)
    total = results.get("total", 0)
    processed = 0

    while True:
        items = results.get("items", [])
        for item in items:
            track = item.get("track")
            if not track:
                continue
            processed += 1

            uri = track.get("uri")

            # Skip if already in playlist
            if uri in existing_tracks:
                continue

            try:
                is_russian, reason = classify_track(sp, track, artist_cache)
                if is_russian:
                    track_name = track.get("name", "Unknown")
                    artists_str = ", ".join([a.get("name", "") for a in track.get("artists", [])])
                    print(f"✓ [{track_name}] by [{artists_str}] - {reason}")
                    if uri:
                        russian_track_uris.append(uri)
            except Exception as e:
                # Log and keep going
                tname = (track.get("name") or "Unknown").encode("utf-8", "ignore")
                print(f"[WARN] Failed to classify track: {tname} | {e}")

        if processed % 100 == 0:
            print(f"Processed {processed}/{total} liked tracks...")
            time.sleep(0.5)  # Small delay every 100 tracks to avoid rate limits

        if results.get("next"):
            results = safe_call(sp.next, results)
        else:
            break

    print(f"Processed {processed}/{total} liked tracks total")

    # Add to playlist in batches
    if russian_track_uris:
        print(f"Found {len(russian_track_uris)} new Russian tracks to add")
        added = 0
        for chunk in chunked(russian_track_uris, BATCH_SIZE):
            safe_call(sp.playlist_add_items, playlist_id, chunk)
            added += len(chunk)
            print(f"Added {added}/{len(russian_track_uris)} to playlist")
            time.sleep(0.3)  # Small delay between batches

        # Optional: unlike after moving
        if REMOVE_FROM_LIKED:
            removed = 0
            # Convert URIs to IDs
            ids = [uri.split(":")[-1] for uri in russian_track_uris]
            for chunk in chunked(ids, BATCH_SIZE):
                safe_call(sp.current_user_saved_tracks_delete, chunk)
                removed += len(chunk)
                print(f"Removed {removed}/{len(ids)} from Liked Songs")
                time.sleep(0.3)  # Small delay between batches

        print(f"Done. Moved {len(russian_track_uris)} new tracks to '{PLAYLIST_NAME}'.")
    else:
        print("No new Russian tracks to add.")

    # Save updated cache
    save_cache({"artists": artist_cache})

if __name__ == "__main__":
    main()