nanakura/moeka-project
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

feat(speech): integrate Voice Pack model into speech catalog and update related tests

Browse Source
This commit is contained in:
RainbowBird
2026-07-01 20:18:13 +08:00
parent af7de96042
commit e1a2c51573
3 changed files with 104 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files
apps/server/src/routes/openai/v1/operations/speech-catalog/index.ts
+13 -3
View File
@@ -6,6 +6,8 @@ import { ofetch } from 'ofetch'
import { createBadGatewayError, createBadRequestError, createServiceUnavailableError } from '../../../../../utils/error'
const VOICE_PACK_MODEL_ID = 'voice-pack'
function voicePackCatalogVoice(pack: VoicePack) {
const cost = `Flux cost: ${pack.costMultiplier}x`
return {
@@ -55,14 +57,19 @@ export function createSpeechCatalogOperation(deps: V1RouteDeps): SpeechCatalogOp
? await deps.configKV.getOrThrow('DEFAULT_TTS_MODEL')
: requested
const voices = await deps.llmRouter.listTtsVoices(model)
const voicePacks = await deps.voicePackService.listEnabled()
if (model === VOICE_PACK_MODEL_ID) {
logger.withFields({ model, voiceCount: voicePacks.length, voicePackCount: voicePacks.length }).debug('list tts voices')
return Response.json({ voices: voicePacks.map(voicePackCatalogVoice), recommended: {} })
}
const voices = await deps.llmRouter.listTtsVoices(model)
const recommended = (await deps.configKV.getOptional('DEFAULT_TTS_VOICES'))?.[model] ?? {}
// Debug level: high-frequency catalog poll from UI selectors, no
// billing / user-facing side effect — useful only when debugging
// voice-picker drift, never as a permanent audit trail line.
logger.withFields({ model, voiceCount: voices.length, voicePackCount: voicePacks.length }).debug('list tts voices')
return Response.json({ voices: [...voicePacks.map(voicePackCatalogVoice), ...voices], recommended })
return Response.json({ voices, recommended })
}
/**
@@ -150,7 +157,10 @@ export function createSpeechCatalogOperation(deps: V1RouteDeps): SpeechCatalogOp
// a TS narrowing aid.
const modelIds = Object.keys(config?.tts?.models ?? {}).sort()
return Response.json({
models: modelIds.map(id => ({ id, name: id })),
models: [
{ id: VOICE_PACK_MODEL_ID, name: 'Voice Pack', description: 'Server-curated voices' },
...modelIds.map(id => ({ id, name: id })),
],
default: defaultModel,
})
}
apps/server/src/routes/openai/v1/route.test.ts
+89 -14
View File
@@ -687,7 +687,7 @@ describe('v1CompletionsRoutes', () => {
/**
* @example
* POST /api/v1/audio/speech { "model": "auto", "voice": "friendly-azure" }
* POST /api/v1/audio/speech { "model": "voice-pack", "voice": "friendly-azure" }
*/
it('resolves Voice Pack aliases to server-owned model, voice, and params', async () => {
const routeTts = vi.fn(async () => new Response(new Uint8Array([1]), {
@@ -728,7 +728,7 @@ describe('v1CompletionsRoutes', () => {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({
model: 'auto',
model: 'voice-pack',
input: 'test',
voice: 'friendly-azure',
}),
@@ -1196,7 +1196,7 @@ describe('v1CompletionsRoutes', () => {
})
describe('gET /api/v1/audio/models', () => {
it('exposes every configured tts model id', async () => {
it('exposes Voice Pack beside every configured tts model id', async () => {
const app = createTestApp(
createMockFluxService(),
createMockConfigKV({
@@ -1221,13 +1221,19 @@ describe('v1CompletionsRoutes', () => {
expect(res.status).toBe(200)
const data = await res.json() as { models: { id: string, name: string }[], default: string }
expect(data.models.map(m => m.id)).toEqual([
'voice-pack',
'alibaba/cosyvoice-v2',
'microsoft/v1',
])
expect(data.models[0]).toMatchObject({
id: 'voice-pack',
name: 'Voice Pack',
description: 'Server-curated voices',
})
expect(data.default).toBe('microsoft/v1')
})
it('returns an empty list when no tts models are configured', async () => {
it('keeps the Voice Pack model entry when no tts models are configured', async () => {
const app = createTestApp(
createMockFluxService(),
createMockConfigKV({
@@ -1242,7 +1248,11 @@ describe('v1CompletionsRoutes', () => {
expect(res.status).toBe(200)
const data = await res.json() as { models: { id: string, name: string }[] }
expect(data.models).toEqual([])
expect(data.models).toEqual([{
id: 'voice-pack',
name: 'Voice Pack',
description: 'Server-curated voices',
}])
})
it('should return 401 when unauthenticated', async () => {
@@ -1389,7 +1399,77 @@ describe('v1CompletionsRoutes', () => {
expect(llmRouter.listTtsVoices).toHaveBeenCalledWith('microsoft/v1')
})
it('includes enabled Voice Packs as official catalog voices without upstream details', async () => {
it('lists enabled Voice Packs from the Voice Pack model without upstream details', async () => {
const llmRouter = createMockLlmRouter({
listTtsVoices: vi.fn(async () => [
{ id: 'en-US-AvaMultilingualNeural', name: 'Ava', languages: [{ code: 'en-US', title: 'English' }] },
]) as any,
})
const voicePackService = createMockVoicePackService({
listEnabled: vi.fn(async () => [
{
id: 'vp-1',
name: 'Narrator',
description: 'Warm voice',
provider: 'azure',
model: 'microsoft/v1',
voiceId: 'narrator-alias',
upstreamVoiceId: 'en-US-AvaMultilingualNeural',
ttsModelId: 'microsoft/v1',
params: {},
costMultiplier: 2,
enabled: true,
createdAt: new Date(),
updatedAt: new Date(),
},
{
id: 'vp-other',
name: 'Other model pack',
description: null,
provider: 'alibaba',
model: 'cosyvoice-v1',
voiceId: 'other-model-alias',
upstreamVoiceId: 'longxiaochun',
ttsModelId: 'alibaba/cosyvoice-v1',
params: {},
costMultiplier: 1,
enabled: true,
createdAt: new Date(),
updatedAt: new Date(),
},
]),
})
const app = createTestApp(
createMockFluxService(),
createMockConfigKV({ DEFAULT_TTS_VOICES: { 'microsoft/v1': { 'en-US': 'en-US-AvaMultilingualNeural' } } }),
undefined,
undefined,
undefined,
llmRouter,
createMockLlmTracing(),
createMockProductEventService(),
voicePackService,
)
const res = await app.fetch(
new Request('http://localhost/api/v1/audio/voices?model=voice-pack', { method: 'GET' }),
{ user: testUser } as any,
)
expect(res.status).toBe(200)
const data = await res.json() as { voices: Array<Record<string, unknown>> }
expect(data.voices[0]).toMatchObject({
id: 'narrator-alias',
name: 'Narrator',
description: 'Warm voice · Flux cost: 2x',
})
expect(data.voices[0]).not.toHaveProperty('upstreamVoiceId')
expect(data.voices[0]).not.toHaveProperty('ttsModelId')
expect(data.voices[1]).toMatchObject({ id: 'other-model-alias' })
expect(llmRouter.listTtsVoices).not.toHaveBeenCalled()
})
it('does not mix Voice Packs into concrete model voice catalogs', async () => {
const llmRouter = createMockLlmRouter({
listTtsVoices: vi.fn(async () => [
{ id: 'en-US-AvaMultilingualNeural', name: 'Ava', languages: [{ code: 'en-US', title: 'English' }] },
@@ -1431,14 +1511,9 @@ describe('v1CompletionsRoutes', () => {
expect(res.status).toBe(200)
const data = await res.json() as { voices: Array<Record<string, unknown>> }
expect(data.voices[0]).toMatchObject({
id: 'narrator-alias',
name: 'Narrator',
description: 'Warm voice · Flux cost: 2x',
})
expect(data.voices[0]).not.toHaveProperty('upstreamVoiceId')
expect(data.voices[0]).not.toHaveProperty('ttsModelId')
expect(data.voices[1]).toMatchObject({ id: 'en-US-AvaMultilingualNeural' })
expect(data.voices).toEqual([
{ id: 'en-US-AvaMultilingualNeural', name: 'Ava', languages: [{ code: 'en-US', title: 'English' }] },
])
})
it('returns an empty recommended map when the resolved model has no bucket', async () => {
packages/stage-ui/src/libs/providers/providers/official/index.ts
+2 -1
View File
@@ -131,7 +131,7 @@ export const providerOfficialSpeech = defineProvider({
if (!res.ok)
throw new Error(`audio models upstream ${res.status}: ${await res.text().catch(() => '')}`.slice(0, 256))
const data = await res.json() as { models?: { id: string, name: string }[], default?: string | null }
const data = await res.json() as { models?: { id: string, name: string, description?: string }[], default?: string | null }
if (!Array.isArray(data.models))
throw new Error('audio models upstream returned malformed body')
@@ -140,6 +140,7 @@ export const providerOfficialSpeech = defineProvider({
return data.models.map(m => ({
id: m.id,
name: m.name,
description: m.description,
provider: OFFICIAL_SPEECH_PROVIDER_ID,
}))
},