Files
minstrel/internal/db/queries/albums.sql
T
bvandeusen 1126bfcf78
test-go / test (push) Failing after 46s
test-go / integration (push) Successful in 5m0s
feat(library): genre + year browse queries and endpoints — #367
Server half of #367. Web UI follows.

Genres are exposed AS-IS per the operator: split on the delimiter, trimmed,
but no case folding and no synonym mapping. So "Rock" and "rock" appear as
separate rows, as does "Rock/Pop" alongside "Rock" and "Pop". The raw spread
has to be visible before anyone can judge whether it needs normalising, and
the alternative is a mapping table to invent and then maintain.

Trimming is not an exception to that. Splitting "Rock; Pop" yields " Pop", and
showing that as a genre distinct from "Pop" would be a bug in OUR splitting,
not fidelity to the operator's tags.

## The correctness trap this had to avoid

ListAlbumsByGenre compared tracks.genre verbatim, while recommendation.sql and
discover.sql have always split it on [;,]. Building the browse index by
splitting while matching exactly would have listed genres whose pages are
empty — every multi-genre track unreachable from either of its genres.

So ListAlbumsByGenre now splits too. That also fixes Subsonic
getAlbumList?type=byGenre, its only caller, which silently missed every
multi-genre track. Its Genre param went *string → string as a result.

EXISTS rather than JOIN + DISTINCT ON throughout: the lateral split emits one
row per (track, fragment), so a join multiplies rows per album and needs
DISTINCT to undo itself. EXISTS asks the question directly, and the count
query then matches the list query by construction rather than by coincidence.

## Genre is a query parameter, not a path segment

Because "Rock/Pop" is a real ID3 tag — the one the task itself cites — and a
slash cannot survive a path segment: Go normalises %2F and the router would
split the value in two. So filtering rides GET /api/library/albums?genre=,
which also reuses the existing paged album surface instead of adding a
parallel one.

Endpoints:

  GET /api/library/genres                          unpaged index + track counts
  GET /api/library/years                           unpaged index + album counts
  GET /api/library/albums?genre=                   filtered page
  GET /api/library/albums?year_from=&year_to=      filtered page, either edge open

The indexes are unpaged deliberately: a client needs the whole set to render a
browsable picker, and paging would let it show only a prefix of an ordering
the user didn't choose.

Two refusals rather than guesses: genre+year together is a 400 (the UI browses
them as separate axes, and quietly dropping half a filter would report a
narrower result than it returned), and an inverted year range is a 400 rather
than being silently swapped.

Undated albums are absent from the year axis rather than bucketed under 0 —
"unknown" is not a year, and a 0 row would sort to one end of a chronological
list looking like data.

Tests: parseYearFilter is pure and runs in the fast lane. The integration
tests assert the thing that would otherwise be silently broken — that a
"Rock;Pop" track is reachable from BOTH genres, that "Rock/Pop" survives as a
filter value, that fragment whitespace is trimmed, and that undated albums
stay out of every year range. Reused the existing seedAlbum/seedTrackWithGenre
fixtures, which already took exactly the year and genre arguments needed.
2026-08-05 13:22:30 -04:00

207 lines
8.1 KiB
SQL

-- name: UpsertAlbum :one
INSERT INTO albums (title, sort_title, artist_id, release_date, mbid, cover_art_path)
VALUES ($1, $2, $3, $4, $5, $6)
ON CONFLICT (mbid) WHERE mbid IS NOT NULL
DO UPDATE SET
title = EXCLUDED.title,
sort_title = EXCLUDED.sort_title,
artist_id = EXCLUDED.artist_id,
release_date = EXCLUDED.release_date,
cover_art_path = EXCLUDED.cover_art_path,
updated_at = now()
RETURNING *;
-- name: GetAlbumByID :one
SELECT * FROM albums WHERE id = $1;
-- name: GetAlbumWithArtist :one
-- Combined fetch for /api/albums/{id}: returns the album row + the
-- joined artist name in a single round trip. Replaces a sequential
-- GetAlbumByID + GetArtistByID pair on the hot detail-page path.
SELECT sqlc.embed(albums), artists.name AS artist_name
FROM albums
JOIN artists ON artists.id = albums.artist_id
WHERE albums.id = $1;
-- name: ListAlbumsByArtistWithTrackCount :many
-- Replaces the N+1 pattern in handleGetArtist (1 + N CountTracksByAlbum
-- per album). Returns each album joined with its track count via a
-- correlated subquery — single round trip regardless of album count.
SELECT sqlc.embed(albums),
(SELECT count(*) FROM tracks t WHERE t.album_id = albums.id)::bigint
AS track_count
FROM albums
WHERE albums.artist_id = $1
ORDER BY release_date NULLS LAST, sort_title;
-- name: GetAlbumByArtistAndTitle :one
-- Scanner uses this for the no-mbid dedupe path: resolve-or-create.
SELECT * FROM albums WHERE artist_id = $1 AND title = $2 LIMIT 1;
-- name: ListAlbumsByArtist :many
SELECT * FROM albums WHERE artist_id = $1 ORDER BY release_date NULLS LAST, sort_title;
-- name: ListAlbumsAlphaByName :many
SELECT * FROM albums ORDER BY sort_title LIMIT $1 OFFSET $2;
-- name: ListAlbumsAlphaByArtist :many
-- Sorted by the owning artist's sort_name. Needed by Subsonic's
-- alphabeticalByArtist album list type.
SELECT sqlc.embed(albums), artists.sort_name AS artist_sort_name
FROM albums
JOIN artists ON artists.id = albums.artist_id
ORDER BY artists.sort_name, albums.sort_title
LIMIT $1 OFFSET $2;
-- name: ListAlbumsNewest :many
SELECT * FROM albums ORDER BY created_at DESC LIMIT $1 OFFSET $2;
-- name: ListAlbumsRandom :many
SELECT * FROM albums ORDER BY random() LIMIT $1;
-- name: ListAlbumsByGenre :many
-- Album "belongs to" a genre if any of its tracks carry that genre.
-- Serves Subsonic getAlbumList?type=byGenre.
--
-- Splits tracks.genre on [;,] as of #367. It previously compared the whole
-- column verbatim, so a track tagged "Rock;Pop" was unreachable from EITHER
-- "Rock" or "Pop" — a Subsonic client asking for a genre silently missed
-- every multi-genre track. This also aligns the endpoint with
-- recommendation.sql / discover.sql, which have always split, and with the
-- genre browse index that #367 adds.
--
-- EXISTS rather than JOIN + DISTINCT ON: the lateral split emits one row per
-- (track, genre-fragment), so a join would multiply rows per album and lean
-- on DISTINCT to undo it. EXISTS asks the question directly.
SELECT albums.*
FROM albums
WHERE EXISTS (
SELECT 1
FROM tracks
JOIN LATERAL regexp_split_to_table(coalesce(tracks.genre, ''), '[;,]') AS g(genre) ON true
WHERE tracks.album_id = albums.id
AND trim(g.genre) = trim(sqlc.arg(genre)::text)
)
ORDER BY albums.sort_title, albums.id
LIMIT sqlc.arg(lim) OFFSET sqlc.arg(off);
-- name: SearchAlbums :many
SELECT * FROM albums
WHERE title ILIKE '%' || $1 || '%'
ORDER BY sort_title
LIMIT $2 OFFSET $3;
-- name: CountAlbumsMatching :one
SELECT COUNT(*) FROM albums WHERE title ILIKE '%' || $1::text || '%';
-- name: ListRecentlyAddedAlbumsWithArtist :many
-- M6a: recently-added albums joined with artist_name + artist_id for the
-- home-page section. created_at is the row-insert timestamp from the
-- scanner — the closest proxy to "added to library". Stable secondary
-- ordering by id keeps pagination tie-break deterministic.
SELECT sqlc.embed(albums), artists.name AS artist_name
FROM albums
JOIN artists ON artists.id = albums.artist_id
ORDER BY albums.created_at DESC, albums.id
LIMIT $1;
-- name: ListAlbumsAlphaWithArtist :many
-- M6a: alpha-sorted album list joined with artist_name. Used by
-- /api/library/albums for the wrapping-grid page. Stable id-tiebreak.
SELECT sqlc.embed(albums), artists.name AS artist_name
FROM albums
JOIN artists ON artists.id = albums.artist_id
ORDER BY albums.sort_title, albums.id
LIMIT $1 OFFSET $2;
-- name: CountAlbums :one
-- M6a: total album count for the /api/library/albums envelope.
SELECT COUNT(*) FROM albums;
-- name: CountAlbumsByArtist :one
-- Used by request-progress reporting to count how many albums have been
-- ingested for a matched artist while a request is still in flight.
SELECT COUNT(*) FROM albums WHERE artist_id = $1;
-- name: DeleteAlbumIfEmpty :one
-- M7 #372: deletes the album row only when it has no remaining tracks.
-- Used after a track delete to tidy up orphaned albums. RETURNING gives
-- us the artist_id so the service can chain into the artist-empty check.
-- Returns no rows when the album still has tracks; the service treats
-- pgx.ErrNoRows as "album not orphaned, skip cascade".
DELETE FROM albums a
WHERE a.id = $1
AND NOT EXISTS (SELECT 1 FROM tracks t WHERE t.album_id = a.id)
RETURNING a.id, a.artist_id;
-- name: SetAlbumMbidIfNull :exec
-- M7 #379: heal MBID on existing album rows during rescan. Idempotent —
-- only updates when the existing mbid is NULL.
UPDATE albums
SET mbid = $2,
updated_at = now()
WHERE id = $1 AND mbid IS NULL;
-- name: ListAlbumsMissingMbidWithTrack :many
-- One-shot MBID backfill: returns each album where mbid IS NULL alongside
-- one of its tracks' file_path so the worker can re-read tags. LIMIT
-- supplied by caller for batching/progress purposes.
SELECT a.id AS album_id,
a.artist_id AS artist_id,
a.title AS title,
t.file_path AS track_file_path
FROM albums a
JOIN LATERAL (
SELECT file_path
FROM tracks
WHERE album_id = a.id
ORDER BY disc_number NULLS LAST, track_number NULLS LAST, id
LIMIT 1
) t ON true
WHERE a.mbid IS NULL
ORDER BY a.created_at ASC
LIMIT $1;
-- name: GetAlbumCoverageRollup :one
-- M7 coverage-gauge: library-wide cover-art coverage rollup for the
-- admin gauge. Single sequential scan with FILTER aggregates —
-- sub-millisecond on realistic libraries. pending_no_mbid is a SUBSET
-- of pending; the UI surfaces it separately so operators can see how
-- many "pending" albums are blocked on missing MBID and won't be
-- helped by another scan.
--
-- Invariant: with_art + pending + settled = total. (pending_no_mbid is
-- not part of the sum — it's a subset of pending, surfaced separately.)
--
-- The IN list below must stay in sync with the cover_art_source CHECK
-- constraint — currently relaxed by migration 0020 to include 'deezer'
-- and 'lastfm', and migration 0030 further relaxed it to any non-empty
-- string. If a new source value is added without updating this query,
-- with_art will silently undercount. Drift #556 caught the deezer +
-- lastfm omission introduced by migration 0020.
SELECT
COUNT(*) AS total,
COUNT(*) FILTER (WHERE cover_art_source IN
('sidecar','embedded','mbcaa','theaudiodb','deezer','lastfm')) AS with_art,
COUNT(*) FILTER (WHERE cover_art_source IS NULL) AS pending,
COUNT(*) FILTER (WHERE cover_art_source = 'none') AS settled,
COUNT(*) FILTER (WHERE cover_art_source IS NULL AND mbid IS NULL) AS pending_no_mbid
FROM albums;
-- name: SetAlbumCoverWithVersion :exec
-- M7 cover-sources: records a successful or settled enrichment with the
-- current sources_version stamp. cover_art_path is NULL when source is
-- 'none' (no file written). Atomic; called per album per pass.
UPDATE albums
SET cover_art_path = $2,
cover_art_source = $3,
cover_art_sources_version = $4,
updated_at = now()
WHERE id = $1;
-- name: GetAlbumsByIDs :many
-- Batched lookup used by /api/library/sync to hydrate upsert payloads
-- (#357). Mirror of GetArtistsByIDs.
SELECT * FROM albums WHERE id = ANY($1::uuid[]);