Replaces the PlaceholderScreen at /app/collections with a real screen that
covers manual and smart collections, matching the SPA's extracted-screen
pattern (screens/CollectionsScreen.tsx; App.tsx gets only the route branch +
import).
Manual collections: list, create, rename, delete, and a per-row toggle for
UseCustomPlaybackOrder (PUT). Item management: an add-items picker that searches
the library (library/browse) and buckets results into the typed
AddItemsToCollectionRequest, plus per-item remove.
Smart collections: list, create, edit (name + query), delete, with a live
result preview that runs the query through library/browse.
API layer: new api/collections.ts owns the typed CRUD (getCollections/
getSmartCollections moved here from schedules.ts; schedules imports them). Unit
tests cover the client, the add-items bucket mapping, and the screen.
Honest gap: no API endpoint lists a manual collection's items by id. The items
view uses a best-effort collection:"name" Lucene search (movies/shows/seasons/
artists only) behind a prominent note. Multi/rerun/playlist collections are out
of scope (no API) and pointed to the Classic UI. Follow-ups #151/#152/#153.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Add a full ChannelEditScreen so the SPA can edit an existing channel
(the list/builder could only create). Mirrors the Blazor ChannelEditor
field set via hydrate-then-PUT against GET/PUT /api/channels/{id}.
- api/channels.ts: getChannelById + updateChannel (typed from generated
ChannelViewModel / UpdateChannelRequest).
- screens/ChannelEditScreen.tsx: section-rail form (General, Playout,
Streaming, Stream selection, Music video, Branding) following the
SettingsScreen draft/saved dirtiness + floating save bar pattern. Logo
upload reuses uploadArtwork; option data via existing pickers clients.
- App.tsx: /app/edit-channel/{id} route (allowSubPaths, screen parses the
id); repoint the channel-table edit pencil at it.
- Tests: API-client GET/PUT + screen hydrate/edit-save/404 coverage.
Deferred per scope: ExternalLogoUrl (backend contract gap), and the
/channels/numbers bulk renumber (existing Renumber toolbar = parity).
Language codes / stream selector / credits template are free-text inputs
(no enumeration endpoints exist); mirror-source list can't filter to
Generated-only (ChannelResponseModel omits playoutSource).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Reference data (profiles, watermarks, fillers, media sources, version,
health) now loads via allSettled with per-resource failure notes; only
the settings groups themselves gate the screen. Error state renders
before the loading guard (was unreachable). +2 tests (136 total).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
SettingsScreen with linkable sub-sections (/app/settings/<section>),
settings API module + useSettingsScreenQuery, floating save bar that
PUTs only changed groups, custom resolution add/delete, legacy-UI
callouts for read-only resources. mockDashboardApi extended; 23 new
tests (suite 112 -> 134).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Adversarial review found the previous multi-flood design non-viable:
PlayoutModeSchedulerFlood never yields to a following Dynamic-start
schedule item, so only the first item ever played, and grouping
media items into a Collection silently dropped the requested order.
Redesign:
- Single-item lineup: one ProgramScheduleItemFlood referencing the
target directly (media item / collection / smart / multi / rerun /
playlist); no generated collection or playlist. Response PlaylistId
is null.
- Multi-item lineup (>= 2): one generated IsSystem Playlist in a
get-or-created IsSystem PlaylistGroup ("Channel Lineups"), one
PlaylistItem per entry in lineup order with PlayAll=true, referenced
by a single Flood schedule item. Rerun collections and playlists are
rejected (422) in multi lineups (PlaylistItem/CollectionKey lack the
fields to enumerate them).
Review fixes:
- Normalize + strict-validate MediaType<->CollectionType pairs once up
front (422 on mismatch / wrong id / not exactly one id).
- Reject MultiCollection with non-Shuffle order (mirrors
PlayoutModeMustBeValid), Mirror playout source, all via 422.
- OnDemand parity: queue TimeShiftOnDemandPlayout post-commit.
- De-collide generated ProgramSchedule and Playlist names against their
unique indexes instead of leaking a UNIQUE-constraint DbUpdateException.
- Generic 422 on save failure + ILogger; AnyAsync existence checks;
Either/Validation unwrap via Match; XML doc on request DTO + endpoint.
- Response model: ChannelId, PlaylistId (nullable), ProgramScheduleId,
PlayoutId (CollectionId removed). Regenerated OpenAPI v1.json + v1.d.ts.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Normalize scan-status percent at the API boundary: the wire value is a
0-1 fraction despite the field name; the bar previously showed ~0% for
the whole scan (fixtures now use fractional wire values incl. 0/1)
- Fix the poll-start race: a triggered library enters a pending set that
survives status pruning until seen active or a 3-tick grace window
expires; the poll runs while pending ∪ active is nonempty (previously
the queue-to-start lag could mean polling never started and the button
re-enabled mid-scan); grace ticks also drain on persistent status-fetch
errors so the UI can't stick
- Keep setState updaters pure: pending/grace bookkeeping hoisted into
pruneGraceExpiredPending() outside the updater (StrictMode-safe)
- Wire the prototype's per-source Scan button (one POST per library,
disabled while any of its libraries is pending/active)
- Honest bare-500 media-sources failure test (backend has no exception
middleware); loadScanStatuses returns its promise; poll effect depends
on derived hasActiveScans; synchronous ref double-submit guard;
--radius-xs fallback 3px; StatusDot label on source cards (75 tests)
Review: PR #128 findings
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Filler unit: treat fillerKind 'None'/null as non-filler (real items no
longer render the Filler badge); add the prototype's Show filler toggle
wired to ?showFiller=true (the client previously never requested filler,
making the filler UI dead code); fixtures now use real wire values
- Hoist selection fetch out of the setState updater (StrictMode-safe)
via a baseRef snapshot; merge polled channel states functionally
instead of clobbering with the captured snapshot
- Fix undefined --ctv-surface-1 token on the hero (→ --surface-card per
prototype); correct --text-lg fallback; collapse the new playout grids
in the 980px media block (no page-level horizontal scroll)
- Visible deferrals: disabled Reset / Schedule reset buttons with note
(no per-playout reset endpoint yet)
- itemsLoading spinners on selection switch; time-window now-matching
with title fallback; rail disabled while reset-all mutates
- Tests: mount-count delta isolation from the Dashboard prefetch;
reset-all reframed as transport-failure (mock honors fixture status;
backend is unconditionally 202); rail filter test; paging honesty
('shown of total' label); listPlayout() fixture matches the list DTO
(63 tests)
Review: PR #127 findings
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Add a schedule selector (previously only the first schedule was ever
shown); switching loads items only, with a pane-level spinner
- Fix duplicated mount effect that fetched the whole screen twice
- Add a collection picker to the add-item flow (was silently using the
alphabetically-first collection with no way to correct it)
- Reorder now reconciles from the PUT response; add/delete refetch items
only — pickers are no longer refetched on every mutation
- Guard reorder against double-submit via a ref; remove the inert Save
button; document the widened item type (OpenAPI gap → #126)
- Inspector fidelity: Tail filler row + Keep multi-part together
- a11y: aria-current for the active item, draggable role description
- CSS: correct --text-md fallbacks; replace nonexistent --text-faint
with --text-disabled in new blocks
- Tests: single-fetch-on-mount assertion, chosen-collection POST body,
picker-failure path, collection() fixture with all required fields;
drop unreachable mock branch (52 tests)
Review: PR #125 findings; refs #126
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Match the backend's exact health status contract ('pass'|'fail'|'warn'|'info')
instead of fictional strings; 'info' now renders as a neutral/idle presentation
and is excluded from failing/warning counts in summarizeHealth.
- Drop the redundant /api/version fetch from getDashboardData/DashboardData;
SidebarVersion's useDashboardVersionQuery remains the single source.
- Disable "Refresh health" (via Button's loading prop) while a health request
is in flight to prevent concurrent double-click requests.
- Add the active-flag unmount guard to useDashboardHealthQuery for consistency
with useDashboardQuery/useChannelsQuery.
- Remove dead .ctv-activity-*/.ctv-release-* CSS left over from the removed
activity feed/release notes UI.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
ChannelResponseModel and BulkChannelRequests DTOs lacked file-scoped
#nullable enable, so the OpenAPI generator emitted every string
property as a ["null","string"] union (project Nullable default is
disable), polluting SPA typegen with needless `| null` unions.
ChannelStateResponseModel already complied with this convention
(precedent: PlayoutResponseModel). All properties in both files are
always populated (never null in practice), so no property needed to
become explicitly optional/nullable - just adding #nullable enable
was sufficient to clean up the generated schema.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Adversarial review fixes for the #97 endpoint:
- nowPlaying now resolves like the XMLTV guide: when the wall-clock item
is pre/mid-roll filler, surface the guide group's program item (title +
full program bounds) instead of the filler; a guide entry with no
program item (e.g. fallback loop) reports null. Previously an ad break
would retitle the channel and reset progress every 30 seconds.
- covering-item lookup now runs one cheap projected query per distinct
playout offset instead of hydrating the full metadata include tree for
a window widened by the largest offset across ALL channels (up to 24h);
metadata is fetched in a second query scoped to the covering guide
groups, time-bounded to stay safe against GuideGroup recycling (mod
10000). Regression caught while testing: covering items are keyed by
(source channel, lookup time) so a source channel no longer inherits
its offset mirror's lookup result.
- GetChannelStatesForApi now carries Now explicitly (controller passes
DateTime.UtcNow), making boundary semantics testable: new tests pin
now == Start (playing) and now == Finish (not playing), plus mid-roll,
filler-only, and mirror-source-row cases.
- ChannelNowPlayingResponseModel title is non-nullable in the schema
(#nullable enable) — GetDisplayTitle always returns a string.
- GetState uses the typed-return controller convention; Playouts.Mapper
reverted to internal (handler lives in the same assembly).
- Spec doc notes the OnDemand drift caveat and filler semantics.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
An off-air channel has no current playout item, so the handler returns
null for NowPlaying; the DTO now declares that (file-scoped #nullable —
ErsatzTV.Core has NRT disabled project-wide). Regenerated OpenAPI and
web types: nowPlaying keeps its nullable oneOf schema (matching the
schema contract test) and channelNumber loses a spurious nullable
annotation left over from an earlier draft — the field is required.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
timothy
