feat(player): CF AI preview/swap + fix PB token expiry + local build time
All checks were successful
Release / Test backend (push) Successful in 42s
Release / Check ui (push) Successful in 1m44s
Release / Docker / caddy (push) Successful in 41s
Release / Docker / backend (push) Successful in 2m42s
Release / Docker / runner (push) Successful in 2m56s
Release / Upload source maps (push) Successful in 1m30s
Release / Docker / ui (push) Successful in 2m42s
Release / Gitea Release (push) Successful in 1m14s

- Backend: add GET /api/audio-preview/{slug}/{n} — generates first ~1800-char
  chunk via CF AI so playback starts immediately; full chapter cached in MinIO
- Frontend: replace CF AI spinner with preview blob URL + background swap to
  full presigned URL when runner finishes, preserving currentTime
- AudioPlayer: isPreview state + 'preview' badge in mini-bar during swap
- pocketbase.ts: fix 403 on stale token — reduce TTL to 50 min + retry once
  on 401/403 with forced re-auth (was cached 12 h, PB tokens expire in 1 h)
- Footer build time now rendered in user's local timezone via toLocaleString()
This commit is contained in:
Admin
2026-04-05 22:12:22 +05:00
parent a308672317
commit 1abb4cd714
6 changed files with 236 additions and 29 deletions

View File

@@ -904,6 +904,115 @@ func (s *Server) handleAudioStream(w http.ResponseWriter, r *http.Request) {
// on its next poll as soon as the MinIO object is present.
}
// handleAudioPreview handles GET /api/audio-preview/{slug}/{n}.
//
// CF AI voices are batch-only and can take 1-2+ minutes to generate a full
// chapter. This endpoint generates only the FIRST chunk of text (~1 800 chars,
// roughly 1-2 minutes of audio) so the client can start playing immediately
// while the full audio is generated in the background by the runner.
//
// Fast path: if a preview object already exists in MinIO, redirects to its
// presigned URL (no regeneration).
//
// Slow path: generates the first chunk via CF AI, streams the MP3 bytes to the
// client, and simultaneously uploads to MinIO under a "_preview" key so future
// requests hit the fast path.
//
// Only CF AI voices are expected here. Calling this with a Kokoro/PocketTTS
// voice falls back to the normal audio-stream endpoint behaviour.
//
// Query params:
//
// voice (required — must be a cfai: voice)
func (s *Server) handleAudioPreview(w http.ResponseWriter, r *http.Request) {
slug := r.PathValue("slug")
n, err := strconv.Atoi(r.PathValue("n"))
if err != nil || n < 1 {
jsonError(w, http.StatusBadRequest, "invalid chapter")
return
}
voice := r.URL.Query().Get("voice")
if voice == "" {
voice = s.cfg.DefaultVoice
}
if s.deps.CFAI == nil {
jsonError(w, http.StatusServiceUnavailable, "cloudflare AI TTS not configured")
return
}
// Preview key: same as normal key with a "_preview" suffix before the extension.
// e.g. slug/1/cfai:luna_preview.mp3
previewKey := s.deps.AudioStore.AudioObjectKeyExt(slug, n, voice+"_preview", "mp3")
// ── Fast path: preview already in MinIO ──────────────────────────────────
if s.deps.AudioStore.AudioExists(r.Context(), previewKey) {
presignURL, err := s.deps.PresignStore.PresignAudio(r.Context(), previewKey, 1*time.Hour)
if err != nil {
s.deps.Log.Error("handleAudioPreview: PresignAudio failed", "slug", slug, "n", n, "err", err)
jsonError(w, http.StatusInternalServerError, "presign failed")
return
}
http.Redirect(w, r, presignURL, http.StatusFound)
return
}
// ── Slow path: generate first chunk + stream + save ──────────────────────
// Read the chapter text.
raw, err := s.deps.BookReader.ReadChapter(r.Context(), slug, n)
if err != nil {
s.deps.Log.Error("handleAudioPreview: ReadChapter failed", "slug", slug, "n", n, "err", err)
jsonError(w, http.StatusNotFound, "chapter not found")
return
}
text := stripMarkdown(raw)
if text == "" {
jsonError(w, http.StatusUnprocessableEntity, "chapter text is empty")
return
}
// Take only the first ~1 800 characters — one CF AI chunk, roughly 1-2 min.
const previewChars = 1800
firstChunk := text
if len([]rune(text)) > previewChars {
runes := []rune(text)
firstChunk = string(runes[:previewChars])
// Walk back to last sentence boundary (. ! ?) to avoid a mid-word cut.
for i := previewChars - 1; i > previewChars/2; i-- {
r := runes[i]
if r == '.' || r == '!' || r == '?' || r == '\n' {
firstChunk = string(runes[:i+1])
break
}
}
}
// Generate the preview chunk via CF AI.
mp3, err := s.deps.CFAI.GenerateAudio(r.Context(), firstChunk, voice)
if err != nil {
s.deps.Log.Error("handleAudioPreview: GenerateAudio failed", "slug", slug, "n", n, "voice", voice, "err", err)
jsonError(w, http.StatusInternalServerError, "tts generation failed")
return
}
// Upload to MinIO in the background so the next request hits the fast path.
go func() {
if uploadErr := s.deps.AudioStore.PutAudio(
context.Background(), previewKey, mp3,
); uploadErr != nil {
s.deps.Log.Error("handleAudioPreview: MinIO upload failed", "key", previewKey, "err", uploadErr)
}
}()
w.Header().Set("Content-Type", "audio/mpeg")
w.Header().Set("Content-Length", strconv.Itoa(len(mp3)))
w.Header().Set("Cache-Control", "no-store")
w.WriteHeader(http.StatusOK)
_, _ = w.Write(mp3)
}
// ── Translation ────────────────────────────────────────────────────────────────
// supportedTranslationLangs is the set of target locales the backend accepts.

View File

@@ -180,6 +180,9 @@ func (s *Server) ListenAndServe(ctx context.Context) error {
// Streaming audio: serves from MinIO if cached, else streams live TTS
// while simultaneously uploading to MinIO for future requests.
mux.HandleFunc("GET /api/audio-stream/{slug}/{n}", s.handleAudioStream)
// CF AI preview: generates only the first ~1 800-char chunk so the client
// can start playing immediately while the full audio is generated by the runner.
mux.HandleFunc("GET /api/audio-preview/{slug}/{n}", s.handleAudioPreview)
// Translation task creation (backend creates task; runner executes via LibreTranslate)
mux.HandleFunc("POST /api/translation/{slug}/{n}", s.handleTranslationGenerate)

View File

@@ -62,6 +62,13 @@ class AudioStore {
/** Pseudo-progress bar value 0100 during generation */
progress = $state(0);
/**
* True while playing a short CF AI preview clip (~1-2 min) and the full
* audio is still being generated in the background. Set to false once the
* full audio URL has been swapped in.
*/
isPreview = $state(false);
// ── Playback state (kept in sync with the <audio> element) ─────────────
currentTime = $state(0);
duration = $state(0);

View File

@@ -603,13 +603,17 @@
return;
}
// CF AI (batch-only) or already enqueued by presign: keep the traditional
// POST → poll → presign flow. For enqueued, we skip the POST and poll.
// CF AI voices: use preview/swap strategy.
// 1. Fetch a short ~1-2 min preview clip from the first text chunk
// so playback starts immediately — no more waiting behind a spinner.
// 2. Meanwhile keep polling the full audio job; when it finishes,
// swap the <audio> src to the full URL preserving currentTime.
audioStore.status = 'generating';
audioStore.isPreview = false;
startProgress();
// presignResult.enqueued=true means /api/presign/audio already POSTed on our
// behalf — skip the duplicate POST and go straight to polling.
// Kick off the full audio generation task in the background
// (presignResult.enqueued=true means the presign endpoint already did it).
if (!presignResult.enqueued) {
const res = await fetch(`/api/audio/${slug}/${chapter}`, {
method: 'POST',
@@ -618,7 +622,6 @@
});
if (res.status === 402) {
// Free daily limit reached — surface upgrade CTA
audioStore.status = 'idle';
stopProgress();
onProRequired?.();
@@ -628,37 +631,96 @@
if (!res.ok) throw new Error(`Generation failed: HTTP ${res.status}`);
if (res.status === 200) {
// Already cached — body is { status: 'done' }, no url needed.
// Already cached — fast path: presign and play directly.
await res.body?.cancel();
await finishProgress();
const doneUrl = await tryPresign(slug, chapter, voice);
if (!doneUrl.ready) throw new Error('Audio generated but presign returned 404');
audioStore.isPreview = false;
audioStore.audioUrl = doneUrl.url;
audioStore.status = 'ready';
maybeStartPrefetch();
return;
}
// 202: fall through to polling below.
// 202 accepted — fall through: start preview while runner generates
}
// Poll until the runner finishes generating.
const final = await pollAudioStatus(slug, chapter, voice);
if (final.status === 'failed') {
throw new Error(
`Generation failed: ${(final as { error?: string }).error ?? 'unknown error'}`
);
// Fetch the preview clip (first ~1-2 min chunk).
// Use an AbortController so we can cancel the background polling if the
// user navigates away or stops playback before the full audio is ready.
const previewAbort = new AbortController();
const qs = new URLSearchParams({ voice });
const previewUrl = `/api/audio-preview/${slug}/${chapter}?${qs}`;
try {
const previewRes = await fetch(previewUrl, { signal: previewAbort.signal });
if (previewRes.status === 402) {
audioStore.status = 'idle';
stopProgress();
onProRequired?.();
return;
}
if (!previewRes.ok) throw new Error(`Preview failed: HTTP ${previewRes.status}`);
// The backend responded with the MP3 bytes (or a redirect to MinIO).
// Build a blob URL so we can swap it out later without reloading the page.
const previewBlob = await previewRes.blob();
const previewBlobUrl = URL.createObjectURL(previewBlob);
audioStore.isPreview = true;
audioStore.audioUrl = previewBlobUrl;
audioStore.status = 'ready';
// Don't restore saved time here — preview is always from 0.
// Kick off prefetch of next chapter in the background.
maybeStartPrefetch();
} catch (previewErr: unknown) {
if (previewErr instanceof DOMException && previewErr.name === 'AbortError') return;
// Preview failed — fall through to the spinner (old behaviour).
// We'll wait for the full audio to finish instead.
audioStore.isPreview = false;
}
await finishProgress();
// Background: poll for full audio; when done, swap src preserving position.
try {
const final = await pollAudioStatus(slug, chapter, voice, 2000, previewAbort.signal);
if (final.status === 'failed') {
throw new Error(
`Generation failed: ${(final as { error?: string }).error ?? 'unknown error'}`
);
}
// Audio is ready in MinIO — always use a presigned URL for direct playback.
const doneUrl = await tryPresign(slug, chapter, voice);
if (!doneUrl.ready) throw new Error('Audio generated but presign returned 404');
audioStore.audioUrl = doneUrl.url;
audioStore.status = 'ready';
// Don't restore time for freshly generated audio — position is 0
// Immediately start pre-generating the next chapter in background.
maybeStartPrefetch();
await finishProgress();
const doneUrl = await tryPresign(slug, chapter, voice);
if (!doneUrl.ready) throw new Error('Audio generated but presign returned 404');
// Swap: save currentTime → update URL → seek to saved position.
const savedTime = audioStore.currentTime;
const blobUrlToRevoke = audioStore.audioUrl; // capture before overwrite
audioStore.isPreview = false;
audioStore.audioUrl = doneUrl.url;
// If we never started a preview (preview fetch failed), switch to ready now.
if (audioStore.status !== 'ready') audioStore.status = 'ready';
// The layout $effect will load the new src and auto-play from 0.
// We seek back to savedTime after a short delay to let the element
// attach the new source before accepting a seek.
if (savedTime > 0) {
setTimeout(() => {
audioStore.seekRequest = savedTime;
}, 300);
}
// Revoke the preview blob URL to free memory.
// (We need to wait until the new src is playing; 2 s is safe.)
setTimeout(() => {
if (blobUrlToRevoke.startsWith('blob:')) {
URL.revokeObjectURL(blobUrlToRevoke);
}
}, 2000);
maybeStartPrefetch();
} catch (pollErr: unknown) {
if (pollErr instanceof DOMException && pollErr.name === 'AbortError') return;
throw pollErr;
}
} catch (e) {
stopProgress();
audioStore.progress = 0;

View File

@@ -100,8 +100,8 @@ export interface User {
let _token = '';
let _tokenExp = 0;
async function getToken(): Promise<string> {
if (_token && Date.now() < _tokenExp) return _token;
async function getToken(forceRefresh = false): Promise<string> {
if (!forceRefresh && _token && Date.now() < _tokenExp) return _token;
log.debug('pocketbase', 'authenticating with admin credentials', { url: PB_URL, email: PB_EMAIL });
@@ -119,7 +119,9 @@ async function getToken(): Promise<string> {
const data = await res.json();
_token = data.token as string;
_tokenExp = Date.now() + 12 * 60 * 60 * 1000; // 12 hours
// PocketBase superuser tokens expire in ~1 hour by default.
// Cache for 50 minutes to stay safely within that window.
_tokenExp = Date.now() + 50 * 60 * 1000;
log.info('pocketbase', 'admin auth token refreshed', { url: PB_URL });
return _token;
}
@@ -132,6 +134,18 @@ async function pbGet<T>(path: string): Promise<T> {
headers: { Authorization: `Bearer ${token}` }
});
if (!res.ok) {
// On 401/403, the token may have expired on the PocketBase side even if
// our local TTL hasn't fired yet. Force a refresh and retry once.
if (res.status === 401 || res.status === 403) {
const freshToken = await getToken(true);
const retry = await fetch(`${PB_URL}${path}`, {
headers: { Authorization: `Bearer ${freshToken}` }
});
if (retry.ok) return retry.json() as Promise<T>;
const retryBody = await retry.text().catch(() => '');
log.error('pocketbase', 'GET failed', { path, status: retry.status, body: retryBody });
throw new Error(`PocketBase GET ${path} failed: ${retry.status}${retryBody}`);
}
const body = await res.text().catch(() => '');
log.error('pocketbase', 'GET failed', { path, status: res.status, body });
throw new Error(`PocketBase GET ${path} failed: ${res.status}${body}`);

View File

@@ -37,6 +37,10 @@
let activeChapterEl = $state<HTMLElement | null>(null);
let listeningModeOpen = $state(false);
// Build time formatted in the user's local timezone (populated on mount so
// SSR and CSR don't produce a mismatch — SSR renders nothing, hydration fills it in).
let buildTimeLocal = $state('');
function setIfActive(node: HTMLElement, isActive: boolean) {
if (isActive) activeChapterEl = node;
return {
@@ -51,6 +55,12 @@
}
});
$effect(() => {
if (env.PUBLIC_BUILD_TIME && env.PUBLIC_BUILD_TIME !== 'unknown') {
buildTimeLocal = new Date(env.PUBLIC_BUILD_TIME).toLocaleString();
}
});
// The single <audio> element that persists across navigations.
// AudioPlayer components in chapter pages control it via audioStore.
let audioEl = $state<HTMLAudioElement | null>(null);
@@ -750,10 +760,9 @@
</div>
<!-- Build version / commit SHA / build time -->
{#snippet buildTime()}
{#if env.PUBLIC_BUILD_TIME && env.PUBLIC_BUILD_TIME !== 'unknown'}
{@const d = new Date(env.PUBLIC_BUILD_TIME)}
{#if buildTimeLocal}
<span class="text-(--color-muted)" title="Build time">
· {d.toUTCString().replace(' GMT', ' UTC').replace(/:\d\d /, ' ')}
· {buildTimeLocal}
</span>
{/if}
{/snippet}
@@ -862,8 +871,11 @@
{m.player_generating({ percent: String(Math.round(audioStore.progress)) })}
</p>
{:else if audioStore.status === 'ready'}
<p class="text-xs text-(--color-muted) tabular-nums leading-tight">
<p class="text-xs text-(--color-muted) tabular-nums leading-tight flex items-center gap-1.5">
{formatTime(audioStore.currentTime)} / {formatTime(audioStore.duration)}
{#if audioStore.isPreview}
<span class="px-1 py-0.5 rounded text-[10px] font-medium bg-(--color-brand)/15 text-(--color-brand) leading-none">preview</span>
{/if}
</p>
{:else if audioStore.status === 'loading'}
<p class="text-xs text-(--color-muted) leading-tight">{m.player_loading()}</p>