summaryrefslogtreecommitdiffstats
path: root/packages/meshbay-node/src/meshbay_node/musicbrainz.py
diff options
context:
space:
mode:
Diffstat (limited to 'packages/meshbay-node/src/meshbay_node/musicbrainz.py')
-rw-r--r--packages/meshbay-node/src/meshbay_node/musicbrainz.py6
1 files changed, 3 insertions, 3 deletions
diff --git a/packages/meshbay-node/src/meshbay_node/musicbrainz.py b/packages/meshbay-node/src/meshbay_node/musicbrainz.py
index d5ca1e9..ce2d233 100644
--- a/packages/meshbay-node/src/meshbay_node/musicbrainz.py
+++ b/packages/meshbay-node/src/meshbay_node/musicbrainz.py
@@ -2,7 +2,7 @@
MusicBrainz (musicbrainz.org) + Cover Art Archive (coverartarchive.org)
client for the Music group app.
-Called only by the node, never by a client (docs/musicbay.md §3): the node
+Called only by the node, never by a client (docs/MESHBAY_DESIGN.md §9.8): the node
makes the one lookup per unique release, shared by every member, and self-
paces against MusicBrainz's shared rate limit rather than letting several
members' tile requests multiply it.
@@ -70,8 +70,8 @@ def _similarity(query: str, val: str | None) -> float:
def _best_match_release(artist: str, album: str, results: list[dict]) -> tuple[dict | None, float]:
"""
Same "trust the search's own ranking" shape as tmdb.py's `_best_match`
- (§3.3 of mediacenter.md found a locally-recomputed re-rank pick a
- coincidentally closer-looking wrong result once — no reason to expect
+ (docs/MESHBAY_DESIGN.md §9.7: a locally-recomputed re-rank was found to
+ pick a coincidentally closer-looking wrong result once — no reason to expect
MusicBrainz's own scored search to fare differently under the same
treatment). MusicBrainz already returns results ordered by its own
`score`; only the top one is considered.