feat: async generation queue with serial execution

Generations now return immediately with a 'generating' status and appear
in history right away. TTS runs in a serial background queue to avoid
GPU contention. Users can kick off multiple generations without blocking.

- Async POST /generate creates DB record immediately, queues TTS work
- Serial generation queue prevents concurrent GPU access (Metal/CUDA/CPU)
- SSE endpoint GET /generate/{id}/status for real-time completion tracking
- Retry endpoint POST /generate/{id}/retry for failed generations
- Store engine and model_size on generation records for retry support
- History cards show animated loader (react-loaders) for generating/playing
- Failed generations show retry button instead of actions menu
- Model downloads happen inline in the queue instead of rejecting with 202
- Stale 'generating' records marked as failed on server startup
- Autoplay on generate setting (default: on)
- Show engine name on generation cards
- Remove sidebar generation spinner
- Checkbox alignment fix in settings
This commit is contained in:
Jamie Pine
2026-03-13 10:02:41 -07:00
parent 52285362ce
commit 655a60ca81
22 changed files with 755 additions and 315 deletions
+6 -13
View File
@@ -8,7 +8,6 @@ import { LANGUAGE_CODES, type LanguageCode } from '@/lib/constants/languages';
import { useGeneration } from '@/lib/hooks/useGeneration';
import { useModelDownloadToast } from '@/lib/hooks/useModelDownloadToast';
import { useGenerationStore } from '@/stores/generationStore';
import { usePlayerStore } from '@/stores/playerStore';
import { useServerStore } from '@/stores/serverStore';
const generationSchema = z.object({
@@ -30,8 +29,7 @@ interface UseGenerationFormOptions {
export function useGenerationForm(options: UseGenerationFormOptions = {}) {
const { toast } = useToast();
const generation = useGeneration();
const setAudioWithAutoPlay = usePlayerStore((state) => state.setAudioWithAutoPlay);
const setIsGenerating = useGenerationStore((state) => state.setIsGenerating);
const addPendingGeneration = useGenerationStore((state) => state.addPendingGeneration);
const maxChunkChars = useServerStore((state) => state.maxChunkChars);
const crossfadeMs = useServerStore((state) => state.crossfadeMs);
const normalizeAudio = useServerStore((state) => state.normalizeAudio);
@@ -71,8 +69,6 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
}
try {
setIsGenerating(true);
const engine = data.engine || 'qwen';
const modelName =
engine === 'luxtts'
@@ -93,6 +89,7 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
? 'Qwen TTS 1.7B'
: 'Qwen TTS 0.6B';
// Check if model needs downloading
try {
const modelStatus = await apiClient.getModelStatus();
const model = modelStatus.models.find((m) => m.model_name === modelName);
@@ -106,6 +103,7 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
}
const isQwen = engine === 'qwen';
// This now returns immediately with status="generating"
const result = await generation.mutateAsync({
profile_id: selectedProfileId,
text: data.text,
@@ -119,14 +117,10 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
normalize: normalizeAudio,
});
toast({
title: 'Generation complete!',
description: `Audio generated (${result.duration.toFixed(2)}s)`,
});
const audioUrl = apiClient.getAudioUrl(result.id);
setAudioWithAutoPlay(audioUrl, result.id, selectedProfileId, data.text.substring(0, 50));
// Track this generation for SSE status updates
addPendingGeneration(result.id);
// Reset form immediately — user can start typing again
form.reset({
text: '',
language: data.language,
@@ -143,7 +137,6 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
variant: 'destructive',
});
} finally {
setIsGenerating(false);
setDownloadingModelName(null);
setDownloadingDisplayName(null);
}