feat(enrichment): loose MusicBrainz search fallback (find aliased/romanized artists) (#771)

* feat(enrichment): loose MusicBrainz search fallback (find aliased artists)

The MB text search used a strict field-phrase query
(`recording:"<title>" AND artist:"<artist>"`). A field phrase only matches
MusicBrainz's *primary* artist/title — it never searches ALIASES — so a
recording stored under a non-Latin primary name (大橋純子) whose romanized
form ("Junko Ohashi") is only an alias returns ZERO results, even though MB
has it. Whole swaths of a community library (e.g. romanized J-pop / city-pop
charts) were unsearchable.

- `build_recording_query(..., loose=True)` drops the field scoping + phrases
  for plain AND-ed term groups (`(telephone number) AND (junko ohashi)`),
  which searches the whole document incl. aliases.
- `_mb_search_recordings` runs the strict query first (unchanged, high
  precision) and only on an EMPTY result retries once with the loose query —
  so mainstream matches are untouched and the extra throttled request is spent
  only on a miss. Results are re-scored by rank_candidates, so recall goes up
  without lowering match quality (auto-accept still needs the per-field floors).

Verified live: "Junko Ohashi / Telephone Number" and "Anri / Windy Summer"
(both 0 under the strict query) now surface the real records; "AC/DC /
Highway to Hell" still hits strict at score 1.0 with no loose retry.

Follow-up (separate): alias-aware SCORING so these can auto-confirm, not just
appear as manual candidates.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(enrichment): keep live exclusion in the loose search fallback

The loose fallback dropped the strict path's -secondarytype:Live filter, so a
studio chart whose strict query missed could fall back to — and, since
score_candidate doesn't penalize live takes, auto-confirm — a live-only
recording. Apply the same live gate to the loose query (skipped only when the
source title is itself a live take).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: byrongamatos <xasiklas@gmail.com>

tests/test_mb_enrichment.py

@@ -97,6 +97,43 @@ def mb_doc(rid="rec-1", title="Thunderstruck", artist="AC/DC", artist_id="art-1"
     }
 
 
+# ── strict-then-loose search fallback ────────────────────────────────────────
+
+def test_search_falls_back_to_loose_when_strict_is_empty(server, monkeypatch):
+    """The strict field-phrase query misses a non-Latin-primary artist; the
+    loose retry (no field scoping) searches aliases and finds it."""
+    calls = []
+
+    def _routed(path, params):
+        q = params.get("query", "")
+        calls.append(q)
+        if q.startswith("recording:"):        # strict phrase → nothing
+            return {"recordings": []}
+        return {"recordings": [mb_doc(rid="rec-x", title="Telephone Number")]}
+
+    monkeypatch.setattr(server, "_mb_http_get", _routed)
+    cands = server._mb_search_recordings("Junko Ohashi", "Telephone Number")
+    assert len(cands) == 1
+    assert len(calls) == 2                     # strict first, then the loose retry
+    assert calls[0].startswith("recording:")   # strict is the field-phrase form
+    assert "artist:" not in calls[1] and '"' not in calls[1]   # loose retry
+
+
+def test_search_does_not_retry_when_strict_hits(server, monkeypatch):
+    """A strict hit must not spend a second (throttled) request on the loose
+    query."""
+    calls = []
+
+    def _routed(path, params):
+        calls.append(params.get("query", ""))
+        return {"recordings": [mb_doc()]}
+
+    monkeypatch.setattr(server, "_mb_http_get", _routed)
+    cands = server._mb_search_recordings("AC/DC", "Thunderstruck")
+    assert len(cands) == 1
+    assert len(calls) == 1
+
+
 # ── offline safety (the pytest-never-hits-network contract) ──────────────────
 
 def test_offline_default_skips_matching(server, monkeypatch):

tests/test_mb_match.py

@@ -189,6 +189,33 @@ def test_build_recording_query_escapes_and_handles_missing_artist():
     assert "artist:" not in q
 
 
+def test_build_recording_query_loose_drops_field_phrases():
+    # The strict form locks to the *primary* artist/title phrase (and drops
+    # live-only recordings — the chart isn't a live take).
+    assert m.build_recording_query("Junko Ohashi", "Telephone Number") == \
+        'recording:"telephone number" AND artist:"junko ohashi" AND -secondarytype:Live'
+    # The loose form has no field scoping and no phrases, so MusicBrainz also
+    # searches artist ALIASES — rescues non-Latin-primary artists (大橋純子) —
+    # but keeps the same live exclusion (a studio chart must not fall back to a
+    # live-only recording).
+    loose = m.build_recording_query("Junko Ohashi", "Telephone Number", loose=True)
+    assert loose == "(telephone number) AND (junko ohashi) AND -secondarytype:Live"
+    assert "artist:" not in loose and '"' not in loose
+
+
+def test_build_recording_query_loose_missing_artist():
+    assert m.build_recording_query("", "Fantasy", loose=True) == \
+        "(fantasy) AND -secondarytype:Live"
+
+
+def test_build_recording_query_loose_keeps_live_for_live_charts():
+    # A live chart's loose fallback must NOT exclude live recordings (same gate
+    # as the strict path) — else its only correct recording is filtered out.
+    loose = m.build_recording_query("AC/DC", "Highway to Hell (Live at Donington)", loose=True)
+    assert "-secondarytype:Live" not in loose
+    assert loose == "(highway to hell) AND (ac dc)"
+
+
 # ── MusicBrainz response parsing ──────────────────────────────────────────────
 
 MB_DOC = {