From 027549be59c1b84655d2ad2bb4ad30a74f943ffc Mon Sep 17 00:00:00 2001
From: Broque Thomas <26755000+Nezreka@users.noreply.github.com>
Date: Sun, 12 Apr 2026 10:48:24 -0700
Subject: [PATCH] Fix album matching using full similarity instead of word
 subset check
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The old subset check treated "Paradise" as matching "Club Paradise"
because {'paradise'} is a subset of {'club', 'paradise'}. Both got
the same +0.10 bonus, so the wrong album could be selected.

Now uses SequenceMatcher for full-string similarity between the wanted
album name and each path segment. Exact matches (>= 0.85) get +0.10,
partial matches (>= 0.60) get +0.03, no match gets +0.00. No penalty
applied — purely adjusts bonus sizing so the correct album ranks higher.
---
 core/matching_engine.py | 26 +++++++++++++++++---------
 1 file changed, 17 insertions(+), 9 deletions(-)

diff --git a/core/matching_engine.py b/core/matching_engine.py
index 83ed1c09..586914c5 100644
--- a/core/matching_engine.py
+++ b/core/matching_engine.py
@@ -657,22 +657,30 @@ class MusicMatchingEngine:
         # --- Source Type ---
         is_youtube = slskd_track.username == 'youtube'
 
-        # 4b. Album Bonus: Prefer results from the correct album folder
-        # This ensures "Dark Side of the Moon" sources rank above "Greatest Hits"
-        # when both have the same song. Binary check — album words in path or not.
+        # 4b. Album Bonus/Penalty: Prefer results from the correct album folder.
+        # Uses full-string similarity to prevent "Paradise" matching "Club Paradise".
+        # The old subset check said "paradise" ⊂ {"club", "paradise"} = True, which was wrong.
         album_bonus = 0.0
         album_name = getattr(spotify_track, 'album', None)
         if album_name and not is_youtube:
-            album_words = set(self.clean_album_name(album_name).split())
-            if album_words:
+            album_cleaned = self.clean_album_name(album_name)
+            if album_cleaned:
+                best_album_sim = 0.0
                 path_segments = re.split(r'[/\\]', slskd_track.filename)
                 for segment in path_segments:
                     if not segment:
                         continue
-                    seg_words = set(self.normalize_string(segment).split())
-                    if album_words.issubset(seg_words):
-                        album_bonus = 0.10
-                        break
+                    seg_cleaned = self.normalize_string(segment)
+                    if not seg_cleaned:
+                        continue
+                    sim = SequenceMatcher(None, album_cleaned, seg_cleaned).ratio()
+                    best_album_sim = max(best_album_sim, sim)
+
+                if best_album_sim >= 0.85:
+                    album_bonus = 0.10  # Strong album match (e.g. "Paradise" vs "Paradise")
+                elif best_album_sim >= 0.60:
+                    album_bonus = 0.03  # Partial match — small bonus
+                # No penalty for low similarity — the file might just not have album folders
 
         # 5. Special handling for short titles (high false positive risk)
         # Titles like "Run", "Love", "Girls", "Stay" need stricter artist matching