navidrome-music-sync/app/wave/collector.py

import os
import sqlite3
import re
from typing import Dict, List, Set, Any, Optional

def _normalize(s: Optional[str]) -> str:
    if not s:
        return ""
    s = s.lower()
    s = re.sub(r"[^\w\s]", "", s)
    return " ".join(s.split())

class WaveCollector:
    def __init__(self, db_path: str = "/navidrome-data/navidrome.db"):
        self.db_path = db_path
        # Fallback paths if running outside docker
        if not os.path.exists(self.db_path):
            candidates = [
                "/home/ai-projects/services/music.qpid.pro/data/navidrome.db",
                "../music.qpid.pro/data/navidrome.db"
            ]
            for c in candidates:
                if os.path.exists(c):
                    self.db_path = c
                    break

    def _get_connection(self) -> Optional[sqlite3.Connection]:
        if not os.path.exists(self.db_path):
            return None
        try:
            return sqlite3.connect(f"file:{self.db_path}?mode=ro", uri=True)
        except Exception as e:
            print(f"[WaveCollector] Error connecting to Navidrome DB: {e}")
            return None

    def get_anchor_profile(self) -> Dict[str, Any]:
        """Collect user preference anchors: starred tracks, played tracks, top artists."""
        profile: Dict[str, Any] = {
            "starred": [],
            "top_played": [],
            "top_artists": [],
            "recent_added": []
        }
        conn = self._get_connection()
        if not conn:
            return profile

        try:
            c = conn.cursor()

            # 1. Starred tracks (high signal)
            c.execute("""
                SELECT mf.id, mf.artist, mf.title, mf.album, mf.genre, a.starred_at
                FROM annotation a
                JOIN media_file mf ON a.item_id = mf.id
                WHERE a.starred = 1
                ORDER BY a.starred_at DESC
                LIMIT 50
            """)
            for row in c.fetchall():
                profile["starred"].append({
                    "id": row[0],
                    "artist": row[1],
                    "title": row[2],
                    "album": row[3],
                    "genre": row[4],
                    "starred_at": row[5]
                })

            # 2. Top played tracks (medium-high signal)
            c.execute("""
                SELECT mf.id, mf.artist, mf.title, a.play_count, a.play_date
                FROM annotation a
                JOIN media_file mf ON a.item_id = mf.id
                WHERE a.play_count > 0
                ORDER BY a.play_count DESC, a.play_date DESC
                LIMIT 30
            """)
            for row in c.fetchall():
                profile["top_played"].append({
                    "id": row[0],
                    "artist": row[1],
                    "title": row[2],
                    "play_count": row[3],
                    "play_date": row[4]
                })

            # 3. Top artists in library (general library profile)
            c.execute("""
                SELECT artist, count(*) as cnt
                FROM media_file
                WHERE artist IS NOT NULL AND artist != ''
                GROUP BY artist
                ORDER BY cnt DESC
                LIMIT 25
            """)
            for row in c.fetchall():
                profile["top_artists"].append({
                    "artist": row[0],
                    "count": row[1]
                })

            # 4. Recently added tracks
            c.execute("""
                SELECT id, artist, title, album, created_at
                FROM media_file
                ORDER BY created_at DESC
                LIMIT 20
            """)
            for row in c.fetchall():
                profile["recent_added"].append({
                    "id": row[0],
                    "artist": row[1],
                    "title": row[2],
                    "album": row[3],
                    "created_at": row[4]
                })

        except Exception as e:
            print(f"[WaveCollector] Profile extraction error: {e}")
        finally:
            conn.close()

        return profile

    def get_existing_track_keys(self) -> Set[str]:
        """Returns set of normalized 'artist|title' of all songs already in Navidrome library."""
        keys: Set[str] = set()
        conn = self._get_connection()
        if not conn:
            return keys

        try:
            c = conn.cursor()
            c.execute("SELECT artist, title FROM media_file")
            for r in c.fetchall():
                art = _normalize(r[0])
                tit = _normalize(r[1])
                if art and tit:
                    keys.add(f"{art}|{tit}")
        except Exception as e:
            print(f"[WaveCollector] Error reading existing track keys: {e}")
        finally:
            conn.close()

        return keys

    def get_starred_wave_tracks(self, wave_subpath: str = "AI Recommendations/Моя Волна") -> List[Dict[str, Any]]:
        """Find tracks in Wave folder that the user has starred (ready for permanent promotion)."""
        starred_wave: List[Dict[str, Any]] = []
        conn = self._get_connection()
        if not conn:
            return starred_wave

        try:
            c = conn.cursor()
            pattern = f"{wave_subpath}/%"
            c.execute("""
                SELECT mf.id, mf.path, mf.artist, mf.album, mf.title, a.starred_at
                FROM annotation a
                JOIN media_file mf ON a.item_id = mf.id
                WHERE a.starred = 1 AND mf.path LIKE ?
            """, (pattern,))
            for r in c.fetchall():
                starred_wave.append({
                    "id": r[0],
                    "path": r[1],
                    "artist": r[2],
                    "album": r[3],
                    "title": r[4],
                    "starred_at": r[5]
                })
        except Exception as e:
            print(f"[WaveCollector] Error finding starred wave tracks: {e}")
        finally:
            conn.close()

        return starred_wave