feat: artist aliases in FTS5 index + BM25 blended scoring
Migration 14: add aliases TEXT column to explore_index, rebuild FTS5 with 3 columns (title, artist_name, aliases), recreate sync triggers. Clears index build timestamps to force alias population on next build. Artist image provider fetches inc=url-rels+aliases (single call, no extra cost). GetAliases() extracts alias names from cached MB rels. indexOneArtist stores aliases as space-separated text after image resolution populates the cache. Search query now uses BM25 blended scoring: ORDER BY bm25(fts, 3.0, 1.0, 0.5) - (ln(popularity+1) * 0.5) Column weights: title=3.0, artist_name=1.0, aliases=0.5 - Title matches score 3x higher than artist name matches - Alias matches are helpful but don't dominate - Popularity is a log-scaled boost, not an override - Exact title match on niche entity beats weak match on mega-popular Enables: 'rhcp' → Red Hot Chili Peppers, 'gnr' → Guns N' Roses, 'sabbath' → Black Sabbath (once index build runs with aliases).
This commit is contained in:
@@ -168,7 +168,7 @@ func (p *ArtistImageProvider) fetchMBRels(artistMBID string) []mbRelation {
|
||||
}
|
||||
|
||||
url := fmt.Sprintf(
|
||||
"https://musicbrainz.org/ws/2/artist/%s?fmt=json&inc=url-rels",
|
||||
"https://musicbrainz.org/ws/2/artist/%s?fmt=json&inc=url-rels+aliases",
|
||||
artistMBID,
|
||||
)
|
||||
|
||||
@@ -368,6 +368,38 @@ func wikimediaThumbURL(filename string) string {
|
||||
)
|
||||
}
|
||||
|
||||
// GetAliases returns the artist's aliases as a space-separated
|
||||
// string, extracted from the cached MB rels response. Returns ""
|
||||
// if no aliases are cached.
|
||||
func (p *ArtistImageProvider) GetAliases(artistMBID string) string {
|
||||
cacheKey := "mb:artist-rels:" + artistMBID
|
||||
|
||||
data, ok := p.cache.Get(cacheKey)
|
||||
if !ok {
|
||||
return ""
|
||||
}
|
||||
|
||||
var envelope struct {
|
||||
Aliases []struct {
|
||||
Name string `json:"name"`
|
||||
} `json:"aliases"`
|
||||
}
|
||||
|
||||
if err := json.Unmarshal(data, &envelope); err != nil || len(envelope.Aliases) == 0 {
|
||||
return ""
|
||||
}
|
||||
|
||||
names := make([]string, 0, len(envelope.Aliases))
|
||||
|
||||
for _, a := range envelope.Aliases {
|
||||
if a.Name != "" {
|
||||
names = append(names, a.Name)
|
||||
}
|
||||
}
|
||||
|
||||
return strings.Join(names, " ")
|
||||
}
|
||||
|
||||
func (p *ArtistImageProvider) fetchURL(url string) ([]byte, error) {
|
||||
ctx, cancel := context.WithTimeout(context.Background(), artistImageTimeout)
|
||||
defer cancel()
|
||||
|
||||
@@ -82,6 +82,7 @@ type SearchIndexResult struct {
|
||||
ArtistMBID string `json:"artistMbid"`
|
||||
Popularity int `json:"popularity"`
|
||||
ExtraJSON string `json:"extraJson,omitempty"`
|
||||
Aliases string `json:"aliases,omitempty"`
|
||||
}
|
||||
|
||||
// lbSitewideArtist is the response shape from the LB sitewide
|
||||
@@ -253,7 +254,7 @@ func (si *SearchIndex) Search(query string, limit int) []SearchIndexResult {
|
||||
FROM explore_index i
|
||||
JOIN explore_index_fts f ON f.rowid = i.id
|
||||
WHERE explore_index_fts MATCH ?
|
||||
ORDER BY i.popularity DESC
|
||||
ORDER BY bm25(explore_index_fts, 3.0, 1.0, 0.5) - (ln(i.popularity + 1) * 0.5)
|
||||
LIMIT ?
|
||||
`, ftsQuery, limit)
|
||||
if err != nil {
|
||||
@@ -1014,6 +1015,23 @@ func (si *SearchIndex) indexOneArtist(
|
||||
|
||||
wg.Wait()
|
||||
|
||||
// Extract aliases from the now-cached MB rels (populated by
|
||||
// the image resolution above) and update the artist's index entry.
|
||||
if si.artistImg != nil {
|
||||
aliases := si.artistImg.GetAliases(artist.ArtistMBID)
|
||||
if aliases != "" {
|
||||
si.writeBatch([]SearchIndexResult{{
|
||||
EntityType: "artist",
|
||||
MBID: artist.ArtistMBID,
|
||||
Title: artist.ArtistName,
|
||||
ArtistName: artist.ArtistName,
|
||||
ArtistMBID: artist.ArtistMBID,
|
||||
Popularity: artist.ListenCount,
|
||||
Aliases: aliases,
|
||||
}})
|
||||
}
|
||||
}
|
||||
|
||||
// Batch write discography results.
|
||||
all := make([]SearchIndexResult, 0, len(rgs)+len(recs))
|
||||
all = append(all, rgs...)
|
||||
@@ -1216,9 +1234,9 @@ func (si *SearchIndex) writeBatch(entries []SearchIndexResult) {
|
||||
for _, e := range entries {
|
||||
if _, err := tx.Exec(`
|
||||
INSERT OR REPLACE INTO explore_index
|
||||
(entity_type, mbid, title, artist_name, artist_mbid, popularity, extra_json)
|
||||
VALUES (?, ?, ?, ?, ?, ?, NULLIF(?, ''))
|
||||
`, e.EntityType, e.MBID, e.Title, e.ArtistName, e.ArtistMBID, e.Popularity, e.ExtraJSON,
|
||||
(entity_type, mbid, title, artist_name, artist_mbid, popularity, extra_json, aliases)
|
||||
VALUES (?, ?, ?, ?, ?, ?, NULLIF(?, ''), NULLIF(?, ''))
|
||||
`, e.EntityType, e.MBID, e.Title, e.ArtistName, e.ArtistMBID, e.Popularity, e.ExtraJSON, e.Aliases,
|
||||
); err != nil {
|
||||
si.logger.Warn("search index: insert error",
|
||||
"mbid", e.MBID,
|
||||
|
||||
Reference in New Issue
Block a user