Translation reassembles translated text onto original timings (§15), so
cues need a structured form. Parsing is tolerant of real files (CRLF,
BOM, missing indices, dot milliseconds); rendering is strict and
renumbers from 1.
#184 and #185 were built in parallel and their candidate types did not
meet. Ranking returned no identity for a candidate, so the winner of a
rank() could not be handed back to Provider::download -- rank() reorders,
so the caller could not recover it by position either.
RankedSubtitle now carries the index of the candidate in the slice it was
given, and Candidate::to_core is the one place the two shapes are mapped:
hash_match against a compared hash, sdh against hearing_impaired, group
against release_group, and the two optional tiebreakers defaulted to sort
last rather than block a candidate.
Recording a subtitle is idempotent on both keys the schema carries: the
sidecar path, and the language an embedded track satisfies, so a second
probe or a re-import converges instead of duplicating. Attempts are
upserted per (media file, language); the work list is every language that
is not satisfied, newest import first, which is the order §15 wants the
daily allowance spent in.
DESIGN.md §15 needs two shapes: what subtitles exist for a media file,
and what arr has tried per wanted language. An embedded track carries no
path — it is recorded because it satisfies a language, not because there
is a file — and the CHECK constraints tie provider, engine and path to
the origin so an impossible row cannot be written.
The skeleton the subtitles milestone hangs off: provider domain types, an
object-safe Provider trait, the crate's own error type, and the cargo
features the translation backends will sit behind. No provider, no
translation, no ranking — DESIGN.md §15 keeps ranking pure in arr-core.
An empty season deck was three truths wearing one message, and the one
it chose to blame was wrong: a season on the per-episode lane sat on
"sweeping indexers…" for the full wait and then blamed a backoff for a
pack search that was never going to run.
`GET /api/series/{id}/seasons/{n}/pack-state` says which lane the
season takes and why, from `season_grab_reason` in arr-core, plus the
failed-pack tally and when #181's window reopens. Seasons gain
`last_pack_search_at`, written only by a season-scoped sweep, so a
pack search that ran and found nothing is a settled answer rather than
a pending one.
The deck then says the true thing in each case, and a season held off
the pack lane by a failure offers the retry that waives its window.
Refs #182
A size band describes one episode (DESIGN.md §5.5), so both the target
penalty and the floor now compare a release's size divided by the number
of episodes it covers. The caller supplies the count — arr-core has no
IO. Movies and unknown counts divide by one, so movie scoring is
unchanged and pinned by test.
Closes#180
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A failed season-pack grab held the season off the pack lane forever:
pack_hard_failed was a bare EXISTS over failed grabs, so one bad
torrent disabled pack search for good, against §6.2's "it never gives
up entirely, it goes quiet".
The guard now rides the shared backoff curve (backoff_elapsed, 1h → 6h
→ 1d → 3d, capped 7d), counting failed pack grabs as attempts and
anchoring on the latest one's grabbed_at. Both the targeted and RSS
lanes agree. A manual season search waives the window outright — the
season deck is §6.2's escape hatch.
Closes#181.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Issue #155 removed the sqlx 0.8 rebuild workarounds by editing migrations
0007, 0014 and 0021 in place. Editing an applied migration changes its
checksum, and `migrate()` refuses to run when one no longer matches what
`_sqlx_migrations` recorded, so the daemon exited on startup against any
database that had already applied them — production included.
The sqlx 0.9 bump is the fix and survives: a new migration can carry
`-- no-transaction` so `PRAGMA foreign_keys = OFF` holds and a table
rebuild stops cascade-deleting its children. Only the retroactive cleanup
of migrations that already ran is reverted, along with #153's
`CHECK (title <> '')`, which rode on the 0021 edit and needs a migration
of its own rather than a rewrite of history.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
With sqlx 0.9 honouring `-- no-transaction`, 0007 loses its
movie_releases_backup stash like 0014 did, and 0021 gains
the CHECK (title <> '') that #153 abandoned because the 0.8
migrator could not run a rebuild with foreign keys off. The
rebuild test now enters at migration 6 so the child links
ride through all three rebuilds, and a new test proves the
'' -> 'TBA' backfill and the rejection of new empty titles.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The metadata lane runs daily, so a series added a moment ago showed no
seasons for up to 24 hours and a movie had no digital release date —
the field §6.2 gates targeted search on.
AppState now carries a MetadataCommand channel alongside the movie,
episode and season ones. Both create handlers send on it after the row
is committed, and a new daemon lane drains it. Its own task rather than
an arm of manual::run: a refresh against TMDB can take a while and must
not sit in front of an operator's manual search.
The add never waits on TMDB and never fails because of it. A refresh
that fails leaves metadata_refreshed_at NULL, which is what the daily
sweep already treats as due, so the title is retried rather than lost.
A command naming a title deleted in between finds no row and does
nothing. METADATA_INTERVAL is unchanged.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
sqlx-sqlite 0.8 ignored Migration.no_tx and wrapped every
migration in a transaction, where PRAGMA foreign_keys = OFF
is a no-op — so any table rebuild cascade-deleted children.
0.9 honours `-- no-transaction`, so 0014 drops its
movie_releases_backup / episode_releases_backup workaround
and runs the plain rebuild recipe with foreign keys off.
A migration test rebuilds both parents from a pre-0014
database and asserts the child link rows survive.
Closes#155
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
DELETE /api/series/{id}/seasons/{n}/files and
DELETE /api/episodes/{id}/files unlink what the scope covers, drop the
matching media_files rows and clear wanted, in one action. 204 on
success, 404 for an unknown season or episode, and a scope with nothing
on disk still clears intent.
Season and episode rows stay: TMDB owns that metadata and the next
refresh would recreate them.
The three scopes share one unlink path. A whole series still resolves to
its title folder (§7.4, atomic); a season or episode resolves to the
recorded file and nothing else, so a narrow call cannot reach a sibling.
The intent clear goes through arr_core::tracking::apply_tracked(false),
the same §4.1 rule #171 landed.
An episode whose file just went is set back to 'missing' when it was
'available', matching what a failed import already does. Closes#174.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Search returns titles only (§9.2, amended): the episode branch of the
library query, its json_each token machinery, and the SPA's episode
row rendering are removed. Deep links to episode releases stay.
Fixes#172
DESIGN.md §9.6 no longer carries cast (#163). Removes the cast
sections, the MetadataCastMember schema and field from both metadata
responses, and every .cast-* rule. arr-meta keeps CastMember and the
credits append; the TMDB call is unchanged.
DESIGN.md §9.6 now puts the current season at the top: seasons
descend by number within a series (season 0 lands last under plain
numeric descending) and episodes descend within each season. Changed
the two queries in load_seasons; the SPA renders this order as given.
arr-compat checked and untouched: it reads seasons with its own
ORDER BY number query straight from the database, never through
arr-api, and SeasonResource carries season_number, which Jellyseerr
matches on rather than position. Sonarr's real API also returns
seasons ascending, so the shim keeps the contract it emulates.
The existing vanished-flag test indexed seasons positionally; it now
looks them up by number so it tests the flag, not the order. New test
asserts [2, 1, 0] for seasons and descending episodes within each.
.sqlx regenerated via just db-prepare. just ci green locally.
Per amended DESIGN.md 9.2, an episode row now also requires at
least one query token to match the episode title on its own, so
a series-title-only query lists the series and no episodes.
A series' first metadata refresh reveals its whole back catalogue, and
apply_auto_track flagged every season as new — adding Rick and Morty
tracked S01-S09 and wanted 91 episodes (#160).
Per DESIGN.md 4.1 the rule applies from the second refresh onward.
The caller passes whether metadata_refreshed_at is set; arr-core stays
IO-free and decides. Wired through the daemon's daily refresh and the
API's create-season endpoint.
Every search row — library and TMDB alike — gains a w92 poster
thumbnail with a same-size blank fallback, a TMDB rating chip
(hidden when there are no votes), and a trailer chip that resolves
GET /api/trailer on click only, opening the tab inside the handler.
Episode hits carry the series' TMDB id so their chip resolves too.
Widen the unified search so both §9.2 halves can render a thumbnail
and a rating: TMDB results gain vote_average/vote_count from the
search bodies, library results read the #145 artwork columns, and
episodes take their series' art. GET /api/movies and /api/series gain
the same two fields for the #151 grid. Mirrors in web updated.
TMDB reports vote_average: 0 where no rating exists (#156). Normalise
it to None at the arr-meta edge, like non_empty does for "", and let
the Option flow through the daemon refresh and API add paths so the
nullable columns from #145 do their job.
§9.6 keeps rich detail out of the database except for the three fields
pure-SQL views need. Adds poster_path, backdrop_path and vote_average to
movies and series, written by the daily metadata refresh in both lanes
and filled at add time from the TMDB response the create flows already
fetch.
movie_detail/series_detail fetch credits, videos and external ids in
one upstream request via append_to_response; cast is truncated to the
top 10 billed in the crate and a trailer is chosen by rule (official
YouTube trailer, any YouTube trailer, YouTube teaser, none).
movie_videos/series_videos serve #144's search-row chip from the
videos endpoint alone. Everything rides the existing 24h cache.
Miguel Palhas