Album picker #730: add word-boundary full-phrase bonus (from PR #731 review)

Compared my #730 fix against contributor PR #731 (same independent design).
Grafted their good idea — a confidence bonus when the album's full core phrase
appears intact in the release title (rescues long multi-word names whose token
coverage gets diluted) — and kept my accent-folding, which #731 lacks (their
normalize drops accented chars: Bjork -> 'bj rk').

IMPORTANT: implemented the phrase bonus WORD-BOUNDARY anchored, not as a raw
substring. My first cut used 'phrase in norm_title' (matching #731) and it
immediately reintroduced the substring bug #730 exists to fix — 'heroes'
matched 'superheroes' and the wrong album scored 0.9/passed. PR #731 has this
latent flaw. The regex anchors the phrase to word boundaries so the bonus
fires for real matches only.

Verified: substring trap (Superheroes/Scary Monsters) rejected; edition
suffixes + intact-phrase albums kept. +1 phrase-bonus test (incl. the
word-boundary guard). 126 plugin tests pass; ruff clean.

Co-authored-by: Tyler Richardson-LaPlume <170156756+IamGroot60@users.noreply.github.com>

core/download_plugins/album_bundle.py

@@ -167,7 +167,17 @@ def album_title_relevance(candidate_title: str, album_name: str) -> float:
     if not album_words:
         return 1.0
     matched = sum(1 for w in album_words if w in title_words)
-    return matched / len(album_words)
+    coverage = matched / len(album_words)
+    # Full-phrase bonus (idea from contributor PR #731): when the album's core
+    # phrase appears intact in the title, we're highly confident it's the right
+    # release even if token-coverage is dragged down by a long multi-word name.
+    # MUST be word-boundary anchored, NOT a raw substring — a naive
+    # `phrase in norm_title` lets "heroes" match "superheroes" and reintroduces
+    # the exact wrong-album bug #730 fixes (PR #731's version has this flaw).
+    core_phrase = " ".join(album_words)
+    if core_phrase and re.search(rf"(?:^| ){re.escape(core_phrase)}(?: |$)", norm_title):
+        coverage = max(coverage, 0.9)
+    return coverage
 
 
 def pick_best_album_release(candidates, quality_guess,

tests/test_album_bundle.py

@@ -163,6 +163,30 @@ def test_relevance_ignores_edition_suffix_on_album_name():
     assert album_title_relevance("Daft Punk - Discovery [FLAC]", "Discovery (Remastered Edition)") >= floor
 
 
+def test_relevance_full_phrase_bonus():
+    """Full core phrase present intact -> high confidence (>=0.9) even when a
+    long multi-word album name would otherwise drag token-coverage down.
+    (Idea grafted from contributor PR #731.)"""
+    # All core words present AND contiguous -> already 1.0; the bonus matters
+    # most when extra album words aren't all matched. Construct that case:
+    # album core = [in, rainbows, the, basement] but title has the phrase
+    # "in rainbows" intact while missing 'basement'.
+    # album core = [in, rainbows, from, basement] ('the' is noise); title has
+    # 'in','rainbows' but not 'from'/'basement', and the full core phrase is
+    # NOT intact -> pure coverage 2/4 = 0.5, no bonus.
+    score = album_title_relevance(
+        "Radiohead - In Rainbows [FLAC]", "In Rainbows from the Basement")
+    assert score == 0.5
+    # Full core phrase intact in the title -> bonus floors at 0.9+ (here it's
+    # already 1.0 since both core words match).
+    score2 = album_title_relevance(
+        "Radiohead - In Rainbows Disc 2 [FLAC]", "In Rainbows")
+    assert score2 >= 0.9
+    # The bonus is word-boundary anchored: a phrase that's only a SUBSTRING of
+    # a title word must NOT get it.
+    assert album_title_relevance("Various - Superheroes", "Heroes") == 0.0
+
+
 def test_relevance_album_named_only_with_noise_or_number():
     # If the album name is JUST a noise/number word, don't strip it to nothing
     # and match everything — keep the literal word.