dsh-audiogen 0.2.0 → 0.3.0

package/src/protocol.ts

@@ -8,7 +8,7 @@
 export const AUDIOGEN_SETTINGS_NAMESPACE = 'dsh-audiogen'
 
 /** Published package version shared by the host updater and the client UI. */
-export const PLUGIN_VERSION = '0.2.0'
+export const PLUGIN_VERSION = '0.3.0'
 
 /** Same-origin route family (loopback-only, mirroring dsh-imagegen). */
 export const SETTINGS_API = {
@@ -45,7 +45,7 @@ export const HISTORY_API = {
 export const HISTORY_MAX = 50
 
 /** Audio generation modes. */
-export type AudioMode = 'tts' | 'music' | 'sfx'
+export type AudioMode = 'tts' | 'music' | 'sfx' | 'voice_design'
 
 /** The capability category of an audio model/voice. */
 export type AudioModelCategory =
@@ -106,6 +106,8 @@ export interface GenerateAudioRequest {
   prompt: string
   /** Optional voice alias for TTS. */
   voice?: string
+  /** Optional preview text for voice-design APIs. */
+  previewText?: string
   /** Optional speaking rate / speed multiplier. */
   speed?: number
   /** Requested duration in seconds (music/sfx). */
@@ -134,6 +136,8 @@ export interface GeneratedAudio {
   url: string
   /** Stable audio id / file name. */
   id: string
+  /** Optional voice id returned by a voice-design API. */
+  voiceId?: string
 }
 
 /** Successful generate outcome. */
@@ -153,6 +157,8 @@ export interface HistoryAudioRef {
   mime: string
   /** Duration in seconds when known. */
   duration?: number
+  /** Optional generated voice id. */
+  voiceId?: string
 }
 
 /** A saved generation as the browser consumes it. */
@@ -163,6 +169,7 @@ export interface HistoryEntry {
   model: string
   prompt: string
   voice?: string
+  voiceId?: string
   speed?: number
   duration?: number
   format?: string
@@ -179,6 +186,7 @@ export interface HistoryEntryInput {
   model: string
   prompt: string
   voice?: string
+  voiceId?: string
   speed?: number
   duration?: number
   format?: string

package/src/routes.ts

@@ -90,7 +90,7 @@ function messageOf(error: unknown): string {
 }
 
 function parseGenerateRequest(body: Record<string, unknown>): GenerateAudioRequest | undefined {
-  const mode = body.mode === 'music' ? 'music' : body.mode === 'sfx' ? 'sfx' : 'tts'
+  const mode = body.mode === 'music' ? 'music' : body.mode === 'sfx' ? 'sfx' : body.mode === 'voice_design' ? 'voice_design' : 'tts'
   const prompt = typeof body.prompt === 'string' ? body.prompt.trim() : ''
   if (prompt === '') return undefined
   return {
@@ -98,6 +98,7 @@ function parseGenerateRequest(body: Record<string, unknown>): GenerateAudioReque
     model: typeof body.model === 'string' ? body.model.trim() : '',
     prompt,
     ...(typeof body.voice === 'string' && body.voice.trim() !== '' ? { voice: body.voice.trim() } : {}),
+    ...(typeof body.previewText === 'string' && body.previewText.trim() !== '' ? { previewText: body.previewText.trim() } : {}),
     ...(typeof body.speed === 'number' ? { speed: body.speed } : {}),
     ...(typeof body.duration === 'number' ? { duration: body.duration } : {}),
     ...(typeof body.format === 'string' && body.format.trim() !== '' ? { format: body.format.trim() } : {}),
@@ -141,6 +142,10 @@ function resolveChannelRequest(
   const target = explicit ?? defaults
   const asked = request.model.trim()
   if (asked === '') {
+    if (request.mode === 'voice_design') {
+      if (target === undefined) return { ok: false, code: 'no-channels', message: '尚未配置任何渠道' }
+      return { ok: true, request: { ...request, channelId: target.id, channel: target.name } }
+    }
     const alias = target?.models[0]?.alias ?? ''
     if (alias === '') {
       return { ok: false, code: 'no-models', message: `渠道「${target?.name ?? ''}」尚未配置模型/音色，请先在设置中添加` }
@@ -300,6 +305,7 @@ export function makeRoutes(deps: AudiogenRoutesDeps): WebRoute[] {
               mime: saved.mime,
               bytes: saved.bytes,
               url: `${AUDIO_API.file}/${encodeURIComponent(saved.file)}`,
+              ...(output.voiceId === undefined ? {} : { voiceId: output.voiceId }),
             })
           }
           let history