feat(db): archive lifecycle schema and queries for missing media

Add the storage behind the archive-instead-of-delete lifecycle:

- media_items.missing_scan_count (INT, default 0) and archived_at
  (TIMESTAMPTZ, partial index), both as idempotent ADD COLUMN IF NOT
  EXISTS backfills for existing installs.
- MarkMediaItemMissing / ArchiveMediaItem / ClearMediaItemArchive plus
  PurgeExpiredArchivedMediaItems (retention cutoff) and
  PurgeAllArchivedMediaItems (manual bulk), with CountArchivedMediaItems
  for the admin UI.
- Archived items are hidden from every user-facing listing:
  ListMediaItems, ListMediaItemsByLibrary, ListMediaItemsSorted,
  SearchMediaItems, SearchMediaItemsUnified, the next_books CTE, the
  library media counts, and the search autocomplete value lists. Detail
  lookups by id/path/sha are intentionally unfiltered, and a dedicated
  ListMediaItemsByLibraryIncludingArchived feeds the scanner so the
  lifecycle pass can see and restore archived rows.
This commit is contained in:
John O'Keefe
2026-09-12 17:50:01 -04:00
parent 44b98f3fc3
commit e6aceae0da
5 changed files with 482 additions and 46 deletions
+54 -3
View File
@@ -141,7 +141,7 @@ ORDER BY l.created_at ASC;
SELECT l.id, COUNT(mi.id) as media_count
FROM libraries l
LEFT JOIN library_visibility lv ON l.id = lv.library_id AND lv.user_id = $1
LEFT JOIN media_items mi ON mi.library_id = l.id
LEFT JOIN media_items mi ON mi.library_id = l.id AND mi.archived_at IS NULL
WHERE COALESCE(lv.is_visible, true) = true
GROUP BY l.id;
@@ -160,6 +160,7 @@ SELECT mi.*, l.name as library_name, lt.name as library_type_name
FROM media_items mi
JOIN libraries l ON mi.library_id = l.id
JOIN library_types lt ON l.library_type_id = lt.id
WHERE mi.archived_at IS NULL
ORDER BY mi.created_at DESC LIMIT $1 OFFSET $2;
-- name: ListMediaItemsByLibrary :many
@@ -167,6 +168,15 @@ SELECT mi.*, l.name as library_name, lt.name as library_type_name
FROM media_items mi
JOIN libraries l ON mi.library_id = l.id
JOIN library_types lt ON l.library_type_id = lt.id
WHERE mi.library_id = $1
AND mi.archived_at IS NULL
ORDER BY mi.created_at DESC;
-- name: ListMediaItemsByLibraryIncludingArchived :many
SELECT mi.*, l.name as library_name, lt.name as library_type_name
FROM media_items mi
JOIN libraries l ON mi.library_id = l.id
JOIN library_types lt ON l.library_type_id = lt.id
WHERE mi.library_id = $1
ORDER BY mi.created_at DESC;
@@ -175,7 +185,8 @@ SELECT mi.*, l.name as library_name, lt.name as library_type_name
FROM media_items mi
JOIN libraries l ON mi.library_id = l.id
JOIN library_types lt ON l.library_type_id = lt.id
WHERE mi.library_id = sqlc.narg('library_id')
WHERE mi.archived_at IS NULL
AND mi.library_id = sqlc.narg('library_id')
ORDER BY
CASE
WHEN sqlc.narg('sort') = 'title ASC' THEN mi.title
@@ -293,6 +304,39 @@ RETURNING *;
UPDATE media_items SET metadata_overrides = '{}', updated_at = NOW()
WHERE id = $1;
-- name: MarkMediaItemMissing :exec
-- First consecutive scan that cannot find the file on disk.
UPDATE media_items SET missing_scan_count = missing_scan_count + 1, updated_at = NOW()
WHERE id = $1;
-- name: ArchiveMediaItem :exec
-- Second consecutive missing scan: hide the item from all browsing while
-- preserving reading history in case the file returns.
UPDATE media_items SET archived_at = NOW(), updated_at = NOW()
WHERE id = $1;
-- name: ClearMediaItemArchive :exec
-- File is back on disk (by path or content hash): restore visibility and
-- reset the missing-scan counter. No-op for items that were never archived.
UPDATE media_items SET archived_at = NULL, missing_scan_count = 0, updated_at = NOW()
WHERE id = $1 AND (archived_at IS NOT NULL OR missing_scan_count > 0);
-- name: PurgeExpiredArchivedMediaItems :many
-- Retention sweep at library-scan time: hard-delete archived items older
-- than the cutoff. Cascades remove reading history with the row.
DELETE FROM media_items
WHERE archived_at IS NOT NULL AND archived_at < $1
RETURNING id, title;
-- name: PurgeAllArchivedMediaItems :many
-- Manual bulk purge from the library admin page.
DELETE FROM media_items
WHERE archived_at IS NOT NULL
RETURNING id, title;
-- name: CountArchivedMediaItems :one
SELECT COUNT(*) FROM media_items WHERE archived_at IS NOT NULL;
-- name: DeleteMediaItem :exec
DELETE FROM media_items WHERE id = $1;
@@ -441,6 +485,7 @@ JOIN libraries l ON mi.library_id = l.id
JOIN library_types lt ON l.library_type_id = lt.id
LEFT JOIN library_visibility lv ON l.id = lv.library_id AND lv.user_id = sqlc.narg('user_id')
WHERE COALESCE(lv.is_visible, true) = true
AND mi.archived_at IS NULL
AND (sqlc.narg('library_id')::uuid IS NULL OR mi.library_id = sqlc.narg('library_id')::uuid)
AND (
mi.title ILIKE sqlc.narg('search_pattern') OR
@@ -513,6 +558,7 @@ JOIN libraries l ON mi.library_id = l.id
JOIN library_types lt ON l.library_type_id = lt.id
LEFT JOIN library_visibility lv ON l.id = lv.library_id AND lv.user_id = sqlc.narg('user_id')
WHERE COALESCE(lv.is_visible, true) = true
AND mi.archived_at IS NULL
AND (sqlc.narg('library_id')::uuid IS NULL
OR mi.library_id = sqlc.narg('library_id')::uuid)
-- Fuzzy author filter
@@ -646,6 +692,7 @@ FROM media_items mi
JOIN libraries l ON mi.library_id = l.id
LEFT JOIN library_visibility lv ON l.id = lv.library_id AND lv.user_id = sqlc.narg('user_id')
WHERE COALESCE(lv.is_visible, true) = true
AND mi.archived_at IS NULL
AND mi.library_id = sqlc.narg('library_id')
AND word_similarity(sqlc.narg('search_query'), COALESCE(mi.author, '')) > 0.3
AND mi.author IS NOT NULL
@@ -663,6 +710,7 @@ FROM media_items mi
JOIN libraries l ON mi.library_id = l.id
LEFT JOIN library_visibility lv ON l.id = lv.library_id AND lv.user_id = sqlc.narg('user_id')
WHERE COALESCE(lv.is_visible, true) = true
AND mi.archived_at IS NULL
AND mi.library_id = sqlc.narg('library_id')
AND word_similarity(sqlc.narg('search_query'), COALESCE(mi.genre, '')) > 0.3
AND mi.genre IS NOT NULL
@@ -697,6 +745,7 @@ FROM media_items mi
JOIN libraries l ON mi.library_id = l.id
LEFT JOIN library_visibility lv ON l.id = lv.library_id AND lv.user_id = sqlc.narg('user_id')
WHERE COALESCE(lv.is_visible, true) = true
AND mi.archived_at IS NULL
AND mi.library_id = sqlc.narg('library_id')
AND word_similarity(sqlc.narg('search_query'), COALESCE(mi.series, '')) > 0.3
AND mi.series IS NOT NULL
@@ -714,6 +763,7 @@ FROM media_items mi
JOIN libraries l ON mi.library_id = l.id
LEFT JOIN library_visibility lv ON l.id = lv.library_id AND lv.user_id = sqlc.narg('user_id')
WHERE COALESCE(lv.is_visible, true) = true
AND mi.archived_at IS NULL
AND mi.library_id = sqlc.narg('library_id')
AND word_similarity(sqlc.narg('search_query'), COALESCE(mi.language, '')) > 0.3
AND mi.language IS NOT NULL
@@ -2461,7 +2511,8 @@ next_books AS (
usp.last_read_at
FROM media_items mi
JOIN user_series_progress usp ON mi.series = usp.series
WHERE (sqlc.narg('library_id')::uuid IS NULL OR mi.library_id = sqlc.narg('library_id')::uuid)
WHERE mi.archived_at IS NULL
AND (sqlc.narg('library_id')::uuid IS NULL OR mi.library_id = sqlc.narg('library_id')::uuid)
AND (mi.series_number > usp.max_read_number OR usp.max_read_number IS NULL)
ORDER BY mi.series, mi.series_number ASC NULLS LAST
)